aiLL 6/14/23 - Discover How AI Will Change Everything Forever

Description

Chapters

Transcript

0:01 foreign
0:30 with the AI learning lab welcome
0:33 everybody
0:36 here we talk about all things AI
0:40 um
0:40 looks like there's a couple of people
0:42 trickling in and they got rid of my dual
0:46 camera feature so that's a pain in the
0:48 butt
0:49 uh if you're new here welcome
0:53 and let me
0:57 flip the camera so if you're new to AI
1:02 go to these sites well go to the top one
1:05 anyway the top site there is chatgpt's
1:07 official website if you've not played
1:09 with catch EPT go there this is the AI
1:13 learning lab my name is Kyle Shannon I'm
1:15 an entrepreneur
1:16 I'm the co-founder and CEO of a company
1:18 called storyvine I'm fairly deep in the
1:22 waters on this generative AI stuff
1:24 and my channel is to um
1:28 demystify talk about this stuff answer
1:30 questions educate
1:33 thank you Kevin appreciate that
1:36 hey Sonia I'm gonna try something new
1:38 tonight normally I try to answer every
1:40 question I'm going to just keep the
1:41 questions rolling live uh I also don't
1:44 have my dual camera thing they Tick Tock
1:47 turned off that feature for me for some
1:49 reason so I can't show you my screen but
1:52 I'll flip back and forth if people have
1:53 questions if you have questions about uh
1:56 Hey Amelia's wife if you have questions
1:58 about AI drop them in the comments I
2:01 will do my best to answer them I'll show
2:03 what I can
2:05 let me
2:08 flip my camera this is a pain they
2:10 change they changed everything on me I'm
2:12 an old man set in my ways what am I
2:14 supposed to do here got an iced tea
2:19 or is it something stronger I don't know
2:21 it's 32 ounces whatever it is so I hope
2:23 it's not stronger
2:26 um welcome everybody if you're new to AI
2:29 this channel is about that I like
2:31 talking about that I answer questions
2:32 about that
2:34 um these sites the top one there is the
2:38 official chat gbt site if you've not
2:40 played with chat gbt I strongly
2:43 encourage you to go do that right below
2:45 it bing.com that's Microsoft search
2:47 engine but if you click the chat button
2:49 that's also chat GPT it's gpt4 it's free
2:53 and it's connected to the internet
2:55 and I can take the comments on Discord
2:58 or another platform and answer them here
3:00 oh that's interesting
3:02 yeah that's not a bad idea so I've got
3:04 let me turn on something else
3:07 so if you go to the I I started a thing
3:10 called the AI Salon we meet every other
3:12 week we have an online meeting and we
3:15 meet in person here in Denver and if you
3:17 go to that URL there's a link tree there
3:20 and in that link tree is the
3:23 um is the link to the Discord server so
3:26 we have a Discord Community there so you
3:27 can ask questions there that's not a bad
3:30 idea Kevin
3:31 and what else
3:35 um yeah if you have questions pop them
3:36 into the comments below hang on I gotta
3:40 flip my camera again because they took
3:43 away my dual cameras what am I gonna do
3:46 um does tick tock have an API I'm pretty
3:49 sure they do
3:51 um I have not and I don't think I have
3:53 any interest in developing anything for
3:55 it but maybe I will I could probably do
3:56 it on zapier
3:58 um I used Bing today with gpt4 feature
4:01 it's amazing thanks for telling us about
4:02 it absolutely yeah it's I mean what's
4:04 nice about it is it's free right and
4:06 it's if you go to normal chatgpt which
4:09 oh dang it I just keep flipping my
4:11 camera here
4:13 um if you go to normal chat GPT
4:16 down
4:20 the gpt4 is for pay
4:24 so you have to pay for GPT plus to get
4:27 access to it bing is free
4:30 um it their open AI is making some some
4:33 big movements they they just added some
4:36 things
4:39 I'm not answering that last comment
4:41 there
4:43 scuba Brett
4:46 you trying to get me in trouble I'm not
4:49 gonna do it I am not going to fall for
4:51 your Wiley way scuba Brett never
4:54 I won't do it
4:56 um do I think there will be virtual
4:58 companionship yes am I going to answer
5:01 your question no
5:06 uh if you're new here my name is Kyle
5:08 Shannon this is the AI learning lab if
5:09 you have questions about AI drop them in
5:11 the comments below hair growth geek are
5:15 you here for a reason is it is it that
5:17 my forehead is ever expanding is that
5:20 why you came
5:21 you're supposed to be asking me
5:23 questions about AI not talking about my
5:26 follicles
5:27 nuggets fan I live in Denver we won the
5:30 championship last night so I'm a fair
5:32 weather fan I don't I mean basketball's
5:34 cool I mean watching jokic is like
5:36 watching that I saw a really funny meme
5:38 where someone someone was acting like
5:40 Napoleon Dynamite Just awkward and just
5:42 like throwing the ball up and say I'll
5:44 just swish he's totally like
5:48 oh Comcast oh Xfinity
5:52 Denver wins and suddenly Skiles and
5:54 Nuggets fan
5:56 um is this an easier way to make an
5:58 avatar and sync it with your voice
6:00 is what what I'm doing here best free
6:03 phone app to use
6:06 um for chat GPT I would just hold
6:12 for chat GPT if I were you I would just
6:15 go to on my phone
6:18 um just point your browser at that top
6:20 URL or it you could get the Bing app
6:24 um or you could just point your Chrome
6:25 browser or your Safari browser to
6:28 chat.openai.com I'd use that rather than
6:31 a phone app because it's more it's got
6:32 more features on it there is an official
6:36 chat CPT app but it's really hard to
6:39 find in the App Store so if you want the
6:40 official chat GPT app go to openai.com
6:47 and they change the home page it's not
6:50 on the home page anymore let's see
6:58 um
7:02 let's see where it is I gotta I gotta
7:04 find it now
7:05 product
7:08 chat GPT
7:10 you too can make money
7:14 okay
7:16 so if you go to
7:19 open AI
7:20 go to their product
7:22 menu go to chat GPT not gpt4 sorry chat
7:27 GPT and then at the bottom left hand
7:29 corner it says
7:32 um download in the App Store
7:34 so if you click on that it'll take you
7:36 over to the that's the official app any
7:39 other app that says chatgpt in the app
7:41 store or anything in Android right now
7:43 is fake [ __ ] stuff yeah they've
7:46 completely changed the home page yeah
7:49 um
7:51 is there an easy way sorry wait is there
7:53 an easy way
7:55 to make an avatar and sync it with your
7:57 voice I think you might have reached out
7:58 to me Kevin
8:00 um
8:02 I know there's a way to do it I I
8:05 haven't done it myself yet so I don't
8:08 actually know what's the easiest way to
8:11 do it the thing that I've made is one
8:12 where I'm just taking a still picture
8:15 and you can record your voice and it'll
8:17 animate to that but it's not in real
8:18 time I think what you are asking about
8:20 was doing it in real time I know it can
8:25 be done because there's apps out there
8:26 that are doing it um
8:28 best option for transcribing right now
8:30 is open ai's whisper and you can if you
8:34 go to open ai's developer portal which
8:39 is platform
8:41 dot openai.com hang on that's play to oh
8:45 wait hang on let me flip my camera since
8:47 they took away dual camera for me if you
8:50 go to platform.openai.com
8:53 that takes you to this thing and then
8:56 boy you can't see it at all hold please
8:59 zooming in
9:01 in the upper right hand corner of the
9:04 chat window is this little microphone
9:07 and that's whisper
9:09 so if you click on that it allows you to
9:12 upload an audio or video file or even
9:14 just start recording and that will
9:16 transcribe
9:17 your stuff and then there's versions of
9:20 this you can download and run on a
9:22 machine there's versions of this you can
9:24 you know call from
9:26 from openai but uh yeah that's that's
9:31 what I that's what I'm using in my
9:33 current company is
9:35 um
9:36 storyline and we do transcriptions in
9:38 there and we're using whisper for that
9:40 and it's it's really good the the the
9:42 transcriptions are not horrible
9:47 um hope it's not too expensive the the
9:49 version that we have because we're
9:51 running it locally we just installed it
9:54 locally it's not costing us anything
9:58 well that's not true it costs our
10:00 developers time to
10:02 do it
10:04 to build it to build it welcome
10:06 everybody this I I am thrown for a loop
10:09 because I don't have normally right here
10:11 I've got a little window that's got my
10:13 back camera on it and they took that
10:14 feature away from me
10:16 so it's called whisper open AI is the
10:19 company whisper is the transcription
10:21 service and if you go to
10:24 platform.openai.com playground in that
10:27 chat window the little microphone in the
10:29 upper right hand corner that's where you
10:30 can upload things to transcribe them
10:32 it's pretty slick it's pretty slick and
10:35 it's really good
10:37 uh and it could do different languages
10:39 as well
10:43 um the other thing you can do is we
10:45 haven't installed this yet but we're
10:47 going to be able to translate English
10:49 into about 120 different languages it
10:51 does subtitles does all sorts of stuff
10:54 it's pretty it's pretty darn good pretty
10:56 darn good
10:58 um if you're new here my name is Kyle
11:00 Shannon this is the AI learning lab I'm
11:02 an entrepreneur CEO of a company called
11:04 storybine which is a video technology
11:06 company
11:07 um in the mid 90s I was uh and and uh I
11:12 was there early for the early days of
11:13 the world wide web and I started one of
11:15 the first digital agencies a company
11:16 called agency.com
11:18 and
11:20 um
11:21 the early days of the web and the early
11:23 days of this AI I'm counting that
11:26 starting like six months ago the launch
11:28 of chat GPT kind of marks the beginning
11:30 of a new era and it's very very
11:33 reminiscent of the early days of the web
11:35 so it's really exciting times right now
11:37 so that's why that's why I have this
11:39 channel that's why I'm talking about
11:40 this stuff and if you have questions
11:42 about AI pop them in the in the comments
11:45 below I'll do my best to answer them
11:47 uh trying to think what else what else
11:50 might be good you know just sitting
11:53 around talking
11:54 talking what are some things that
11:56 happened uh open AI just updated all of
11:59 their apis so if you're a developer the
12:04 um the open AI
12:06 um interfaces for GPT you can now call
12:09 functions which for the rest of us
12:12 doesn't really mean crap but it's kind
12:15 of like it's kind of like you can have
12:17 chat GPT with plugins but instead of
12:19 calling a plug-in you're actually
12:21 calling some service so you could
12:22 basically point it at you know apis that
12:25 you use for your business and build that
12:27 right into a chat GPT application so
12:29 it's pretty pretty powerful stuff is
12:31 coming if you have questions pop them in
12:34 the comments below do my best to answer
12:35 them
12:36 I heard Bard was blocked by the European
12:39 Union over privacy concerns interesting
12:41 I hadn't heard that let's go let me go
12:43 take a look at that and see if that's
12:46 is that news that's out there let's see
12:50 EU bard
12:55 Google bard blocks All European
12:57 countries
12:59 when was that
13:01 yesterday
13:04 it was announced that Google would be
13:06 undergoing
13:07 hold please
13:09 let's go to learn together
13:14 it would be undergoing a massive
13:16 expansion bringing AI chatbot experiment
13:18 to 180 countries however what Google
13:21 didn't mention is that Bard still isn't
13:23 available in the European Union
13:28 last month Italy ban Chach EVT we knew
13:31 that they unbanned it after Sam Altman
13:33 called them
13:34 said hey Italy
13:36 and whatever he told them they were fine
13:38 with so they turned it back on
13:43 uh
13:45 yeah interesting
13:47 interesting I I mean
13:50 you know I think this is going to
13:51 continue to happen for a while but
13:54 um
13:56 but I like I honestly don't know like
13:59 like like when when
14:02 Google rolls barred into all of their
14:04 office suite and when Microsoft rolls
14:07 open AI into all of office 365.
14:11 is that feature just not going to be
14:12 available in certain countries like I
14:14 don't I don't quite know how that's
14:15 going to work and then with these log
14:17 with the open source large language
14:18 models I again I'm I'm not quite sure
14:21 how this is all going to work
14:24 um hair growth geek was just talking
14:27 about this at work wait oh wait I'm
14:30 going back
14:31 um any AI apps for HIPAA compliant data
14:34 sharing
14:37 um we do a lot of work with Pharma so
14:39 we're not HIPAA compliant but we deal
14:41 with uh Phi all the time so
14:46 um
14:48 I don't know well
14:51 you can you can run your own instance so
14:53 so open AI in conjunction with Microsoft
14:56 you can do Standalone virtual machines
14:58 for your company and then you can you
15:01 can manage all of your own HIPAA
15:03 compliance that way I assume you could
15:05 get a HIPAA compliant virtual machine
15:07 through Microsoft azure
15:10 um
15:11 but if you're doing any sort of calls to
15:15 um to open ai's API according to our
15:18 lawyer their terms are kind of squishy
15:21 now now they're coming out with chat GPT
15:24 business where I up
15:27 champ is protecting us from the
15:29 neighbors
15:30 moms and strollers very dangerous
15:33 apparently
15:36 um
15:37 champ
15:40 it doesn't help it just makes me feel
15:43 better yelling at him okay if you wanted
15:45 to invest
15:47 um what would some companies to buy
15:48 stock I don't know I can't listen I'm
15:50 horrible with money so don't ask me
15:52 about investment
15:54 I you know I would pay attention to to
15:57 who's making Who's Making Moves
16:00 um Microsoft is acting like a startup
16:03 and they're they're a thorn in Google
16:05 side right now Google's struggling and
16:07 playing catch-up
16:10 um and I think the
16:12 the
16:13 search advertising model is is at threat
16:17 and I think they know that and I think
16:18 that's why they're scrambling to get
16:20 their AI [ __ ] together
16:21 Nvidia just hit a trillion dollar market
16:23 cap they're not going away anytime soon
16:26 um open AI is interesting but not
16:28 investable
16:29 um I really like the company anthropic
16:31 but I don't think they're investable
16:34 um
16:37 I don't know I wouldn't bet against
16:39 Apple I think Apple Apple's got the
16:41 hardware to be able to do a lot of this
16:43 machine learning and AI stuff locally on
16:45 devices so from a privacy perspective
16:48 like when when it when it comes time for
16:51 actual privacy and actual
16:54 um authentication of content I think a
16:57 company like Apple's gonna they have the
16:59 ecosystem to be able to nail it
17:02 so when Apple gets into the AI game
17:04 they're going to probably get in right
17:06 around the time when the 2024
17:08 presidential election in the U.S has hit
17:11 fevered pitch and there's misinformation
17:13 everywhere and deep fakes everywhere and
17:15 people are like we need privacy and we
17:17 need authentication Apple will just step
17:19 up and go here you go put it on a silver
17:22 platter so that'll probably happen there
17:24 uh um
17:27 all right
17:29 oh you're an online pharmacy oh that's
17:31 cool yeah I don't like right now like
17:34 what I would what I would work on until
17:37 it there might be
17:39 here take a look
17:43 this is for you hair growth person
17:45 what's your what's your handle hair
17:47 growth geek
17:48 um go to go to that bottom URL
17:50 futurepedia.io that's a directory of
17:53 just like an infinite number of uh AI
17:57 tools see if anyone has done the the
17:59 HIPAA thing in there or hi just just
18:01 search for privacy or hip to see if
18:03 anything shows up there my guess my
18:06 strong guess is that we're kind of too
18:09 early right now
18:10 for actual professional applications
18:13 everything right now like think think of
18:16 the generative AI industry right now is
18:19 just we're in the spit and duct tape
18:20 phase like shit's launching you know two
18:23 or three major projects a day are
18:26 launching people are like here's the
18:28 next best shiny thing and then you
18:29 realize oh it's like a 19 year old in
18:31 this mom's basement you know putting out
18:34 that thing that looked all professional
18:37 um so there's a lot of that going on
18:38 right now um so but I would go look at
18:40 futurepedia that might help despite why
18:42 spread reporting that the market
18:44 expected
18:45 wait that's a market expected
18:48 fed not to raise rates Bard kept saying
18:51 otherwise
18:53 oh so so listen do not expect any of
18:57 these large language models to to do
19:00 anything resembling the truth
19:02 um they hallucinate like mad
19:05 um and I don't I think I'm not sure if
19:08 Bard's connected to the internet or not
19:11 um but Bard is Bard's a mess Bard's not
19:14 very good right now I would use chat gbt
19:16 so if you
19:19 um want to use chat GPT go to that top
19:22 URL chat.openai.com if you haven't used
19:24 it go use that
19:26 and then Bing is gpt4 for free and Poe
19:30 lets you see a bunch of uh different
19:31 models all on the same interface and
19:34 someone someone described it yesterday
19:35 Poe is like the play school of of chat
19:39 interfaces which I thought was funny and
19:41 pretty true
19:42 all right
19:44 um Jasper won't retrain on your business
19:46 data okay that's cool
19:49 um you can also turn off
19:51 um chat gpt's training on your data I'll
19:54 show you how to do that they changed
19:55 that interface as well everything's
19:57 changing all the time right now but let
19:59 me show you that so if you're in chat
20:01 GPT
20:06 and you go to your settings
20:10 down here
20:12 and go settings
20:14 you have data controls and then in here
20:18 you've got your chat history
20:22 which you can turn off but they just
20:24 they they just move something hang on a
20:26 sec
20:27 clear all chats
20:29 huh
20:32 that's very strange okay there used to
20:34 be a section in here that said download
20:36 my data so if you turn off your chat
20:40 history and training so they can't train
20:42 on their data it deletes all of your
20:44 chat history and there used to be a
20:46 thing here that said you could download
20:48 your data they don't have that anymore
20:50 so this is the uh the joy and wonder of
20:53 cloud-based software is that the thing
20:55 that you took you know a month to learn
20:57 how to use you log in the next time and
21:00 it's like no that's different now like
21:02 like Tick Tock I used to have a dual
21:04 camera feature where I could put my
21:05 laptop screen right here that's gone
21:09 you know what are you gonna do
21:12 is
21:13 the world we live in hey I thought that
21:15 thing worked that way it did where's the
21:17 manual we don't have manuals just figure
21:19 it out yourself you loser
21:27 all right
21:29 thank you for the kind words I
21:31 appreciate that
21:33 um
21:35 I have great hair and a lovely forehead
21:38 oh not a five head thank you thank you I
21:41 I appreciate the forehead comment it is
21:43 getting larger my forehead so you know
21:45 I'm not saying it's a hair problem but
21:48 you know I got to get some new toupee
21:50 glue or something
21:55 spilling crap all over myself
22:00 uh welcome if you're new here my name is
22:03 Kyle Shannon this is the AI learning lab
22:04 I'm an entrepreneur
22:06 I'm not an AI or machine learning guy
22:10 um I'm a business guy that likes to sort
22:12 of take it advantage of Technology when
22:15 it's kind of there for the masses I did
22:17 it in the early days of the World Wide
22:19 Web
22:20 um and started one of the early digital
22:22 agencies company called agency.com and
22:25 uh the minute I saw this uh generative
22:28 AI stuff I started on the image side of
22:30 the house and then when chat EPT came
22:32 out I was like oh this is something
22:35 different special profound uh game
22:39 changing and uh yeah had a big Epiphany
22:42 about it and so just I just went crazy I
22:45 started this Tick Tock Channel I started
22:47 a an a newsletter so my newsletter I'll
22:51 give you the info for that if you're
22:53 curious because it's it is valuable we
22:57 we go through stuff so let me turn off
22:59 this stuff and on this stuff so I
23:03 started uh the AI salon so that's the
23:05 the link tree to the AI Salon is up at
23:07 the top there you can join our Discord
23:09 you can come to we have meetups every
23:11 other
23:12 um every other week and then if you go
23:15 to to tick tock the everyday AI news
23:18 um is our Channel there and there's
23:19 links there to subscribe to the
23:22 newsletter so if you want to keep up
23:23 with this stuff what I can tell you
23:25 running a newsletter is it is impossible
23:28 to keep up with this stuff it's
23:30 impossible it's like
23:32 I thought stuff was moving fast back in
23:35 the mid 90s with the web stuff
23:37 that was sleepy compared to what's
23:39 happening right now it's crazy it's
23:42 crazy I tell ya so if you have questions
23:45 about AI drop it in there great webinar
23:47 today thank you very much I appreciate
23:49 that that was the uh the one Club thing
23:51 the uh yes the the fireside chat
23:55 it was a Fireside well it was a chat
23:58 there was no fire unless you consider
24:00 the burning phosphorus well we don't
24:02 even have burning phosphorus anymore
24:03 that's a CRT reference for you old
24:06 people
24:07 the burning LEDs uh were the closest
24:10 thing we had to a fire today
24:13 um
24:14 your GPT has
24:16 export you know what I have a feeling
24:19 you're right I I have a feeling I just
24:21 need to reload
24:23 let me reload
24:26 this chat GPT
24:29 maybe I even should log out and log back
24:33 in
24:35 Logan
24:36 login doggies
24:41 [Music]
24:44 and I'm gonna go to settings
24:49 and I'm going to go to data controls
24:51 okay now I've got it okay so I was
24:53 showing you this before okay if in chat
24:56 gbt you don't want you want to be as
24:59 secure as possible
25:01 chat history and training you can turn
25:04 that off
25:05 now if you turn it off it deletes all of
25:08 your chat history which is why I haven't
25:10 turned it off because I want my chat
25:12 history so if you're going to turn that
25:14 off make sure you export your data which
25:18 is this thing right here and then you'll
25:21 get all your chat history and you can
25:22 dump it into a Sheets file or something
25:24 like that
25:26 so
25:28 and if you are new here I don't think we
25:31 have any new people here anyway those
25:32 are the top three are official chat GPT
25:36 site Microsoft's Bing that's gpt4 and
25:39 thenpo.com which has six different chat
25:43 models within it that you can play with
25:46 Google deepmind open Ai and anthropic
25:48 have agreed to open their AI models to
25:52 UK government for research and safety
25:54 purposes oh that's really interesting
25:55 retro Punk
25:57 very cool
26:00 let me flip my camera back
26:03 why is my dual camera feature gone I
26:07 don't want it to be gone I want things
26:11 to work the way they used to work
26:13 yesterday
26:14 no the day before
26:20 if you didn't know it I have a degree in
26:22 acting
26:27 come for the hair stay for the idiocy
26:31 uh what's what's the AI that makes
26:36 what's the AI that makes app what's that
26:40 I don't know I don't know what words are
26:43 supposed to be in that sentence tarfari
26:46 89 but if you rephrase that in a form
26:50 that I could understand it as a as a
26:52 native English speaker I would
26:53 appreciate it I don't know quite what
26:55 that means
26:56 um are you on Discord if so
26:59 could you show everyone the open AI
27:02 prompt Library yeah I am on Discord
27:06 um you're assuming
27:09 that I've been to the open AI Discord
27:12 Channel recently and I haven't but
27:15 no time like the present
27:18 so if you're if you're new to Discord
27:20 entirely this is Discord
27:23 these little circles down the left here
27:26 are servers so one of these is the open
27:29 AI
27:30 server I forget what the the logo looks
27:33 like one of the problems with these is
27:35 that they all have little icons but you
27:37 got to roll over them and I forget what
27:39 the open AI one looks like that's it
27:41 okay
27:42 so if I go to open AI
27:44 and then once you're in the server then
27:46 you have all these little dis discussion
27:48 areas and then the discussion is to the
27:50 right and who's on the servers to the
27:52 left and you can turn that on or off so
27:54 just there's the conversations
27:56 and let's see
27:59 prompt engineering oh prompt Library
28:01 okay oh yeah this is pretty cool I have
28:03 been here before Okay so
28:05 go to Discord sign up
28:10 for the go go to openai.com okay jeez
28:14 lordy people let's have an add moment
28:17 shall we let's go back to the internet
28:19 let's go back to the interwebs and we're
28:21 going to go to openai.com
28:23 so if you want to find open ai's Discord
28:26 server
28:28 I don't know how we're going to do that
28:29 oh I know how we're going to do that
28:30 we're going to search for it open AI
28:33 Discord
28:36 okay
28:39 discord.com invite slash open AI
28:46 all right so go Google open AI Discord
28:49 and then that will bring you here and
28:51 and you basically sign up you follow
28:53 their little instructions and then if
28:55 you click on all the things that you
28:57 have to click on to to join then you'll
28:59 see all these comments down below here
29:01 you can ask questions about chat EBT GPT
29:04 3 4 and then there's prompt engineering
29:08 which are discussions about prompt
29:09 engineering and then there's the prompt
29:11 Library which is a nicely organized
29:14 there's gpt4 prompts chat GPT prompts
29:18 gpt3 multi-purpose prompts
29:22 and so lots and lots of
29:26 lots and lots of things there the other
29:28 thing that I show on my fancy little
29:32 this little fancy picture here that that
29:34 fourth thing down prompts dot chat
29:37 that is similar to the prompts library
29:39 in Discord
29:41 so let's go take a look at that it
29:44 doesn't look as nice certainly on the
29:45 tikitaki
29:47 but it's called awesome chat GPT prompts
29:49 the first part of this document is uh
29:52 explains to you about prompting and
29:54 there's a free ebook and there's a paid
29:56 ebook I don't know what that's about but
29:58 whatever there's prompting resources
30:01 then they talk about the chat GPT app
30:02 and then it goes into this long list of
30:07 um
30:08 prompts to tell Chachi PT to act like a
30:11 certain thing and it will it'll act like
30:13 an Excel spreadsheet or a travel guide
30:16 or a plagiarism Checker or a character
30:18 from your favorite movie or a
30:20 Storyteller right this just literally
30:21 goes on and on and on and on so if you
30:26 are stuck trying to figure out well what
30:28 would I ever do with chat JPay Tay
30:30 go to prompts.chat and they will tell
30:33 you what to do with chat jpt
30:36 and if you hear anyone on Tick Tock or
30:38 YouTube go make money with cha cha
30:40 payday
30:41 ignore them
30:43 they're just trying to make money off
30:45 you they don't know how to [ __ ] make
30:47 money they wouldn't be making videos
30:48 about making money on chat EBT if they
30:50 were making money on chat GPT do you
30:52 ever think about that
30:54 idiot
30:57 not you them those kind of people
30:59 driving crazy they make me crazy they
31:01 make me crazy they might be insane I
31:03 don't like it I don't like it I don't
31:06 like it and I don't know if you told you
31:08 this before but I don't like it
31:10 thanks for the demo you're welcome if
31:12 you have uh questions comments about uh
31:16 AI things like that pop them in the
31:19 comments below I'll do my best to answer
31:21 them if you want to make fun of my looks
31:23 that seems to be a popular activity here
31:26 in the AI learning lab where do I start
31:28 to learn about AI that is an awesome
31:31 question and that is why
31:33 I made this Photoshop file
31:36 because I think still and listen when
31:39 this changes I will change this graphic
31:41 I have the power I've got editing
31:43 capabilities on this particular graphic
31:46 I think the best place to start learning
31:48 about
31:50 um
31:50 AI is right here
31:53 chat.openai.com that's the official
31:55 chatgpt website it looks like
31:59 this well not that when you go to your
32:02 settings it looks like that it looks
32:04 like this where you got your chats over
32:05 here to the left and you got this giant
32:08 nothing window this is what this window
32:10 represents is all of the possibility of
32:12 all of humanity and if you're using the
32:15 free version you it's called GPT turbo I
32:18 think they call it
32:19 um if you pay for it it's GPT 3.5
32:21 because I don't know I don't know why
32:24 they name it different but anyway but
32:26 you can just if you've never used this
32:28 before you can do things like
32:30 um explain
32:34 um let's see the theory of relativity
32:40 as if I were a fifth grader in a Country
32:47 Song
32:50 bag and it will write you a song so so
32:54 if you've never played with uh this tool
32:58 these are what I like to call parlor
33:00 tricks where you just tell it to do
33:01 something stupid and then pick your jaw
33:04 up off the floor when you realize how
33:06 fast it just did it we'll gather around
33:08 young cow poke I'll sing you a tune
33:11 about a mighty smart fella a genius
33:13 named Einstein he's a loon he had a wild
33:16 idea about how the world goes around
33:18 called theory of relativity it's a
33:20 scientific sound
33:21 it's a theory to relativity buckle up
33:24 and hold on tight it's a wild ride
33:26 through space and time it'll blow your
33:28 mind all right so there's a little
33:30 Charlie Daniels action here so it'll do
33:33 that you could also say
33:35 um
33:37 make a table comparing
33:41 the theory
33:44 of Relativity with
33:48 quantum mechanics
33:52 I don't even know if that's possible
33:53 well apparently it is vegetable
33:57 all right like what oh description the
34:01 focus of Relativity is gravity and large
34:04 scale phenomenon and quantum mechanics
34:06 particles and microscopic phenomena so
34:10 you know and it can write code you could
34:12 say
34:13 um write me a python App application
34:20 that lets me play with
34:26 um math formulas of these two theories
34:32 don't know if that's possible certainly
34:35 here you go and it writes you an
34:38 application so so if you've not
34:41 played with chat apt and really started
34:43 to get your head around
34:46 the remarkable power of of this tool
34:50 that's the place to start and then
34:54 you you know then after that then it
34:57 starts to get serious because it's you
34:58 know for me like one of my favorite
35:00 moments is when someone's like you know
35:01 I heard about this AI stuff but aren't
35:03 the robots just gonna kill us
35:06 and then you're like yeah that's the
35:07 movies and they might but in the
35:09 meantime why don't you go take a look at
35:11 this and then they go holy crap and then
35:13 they play with it and they're like oh my
35:15 god did you know I could do this
35:16 and they all talk like that until
35:18 they've seen it and then they talk not
35:21 like that
35:22 um
35:25 what I would then do is I would say
35:27 think about what you do what you're
35:29 interested in and start playing with
35:33 um chat EBT or discovering other tools
35:36 just start getting curious about it and
35:37 start doing what you do or things that
35:39 you're interested in with the tool so if
35:42 you're really into gardening have it
35:44 create you I don't know a list of plants
35:46 to plant for your region at that given
35:48 time of the year or a way to battle
35:50 pests or have it write you a grocery
35:52 list or have it right you help you start
35:55 the novel you've been procrastinating
35:57 for 20 years
35:59 um that's where I'd start
36:01 all right let's see
36:07 any sources for mid-journey prompts for
36:11 creating Graphics to use for Design
36:13 Elements yes well
36:19 so let me show you let me show you a
36:21 couple of things here
36:23 for mid-journey so for those of you
36:25 don't who don't know Chachi think of
36:27 Chachi PT as words and then when it
36:30 comes to pictures the tools the the
36:32 major tools are mid-journey stable
36:35 diffusion Adobe Firefly
36:38 um and Dali which is open AIS
36:41 mid-journey in my opinion is by far the
36:43 best in terms of quality stable
36:45 diffusion is the best in terms of
36:47 developer community and and innovation
36:51 fireflies the best for commercial
36:53 commercial usage
36:55 all of them require prompting so one
36:58 thing I'm going to show you here I'm
36:59 going to go to Poe so this is Poe
37:02 poe.com
37:04 and it's just like chat EBT where you
37:07 have a little chatty thing at the bottom
37:08 and a window at the top down the left
37:10 hand side here you've got different
37:12 models you've got Sage which is their
37:14 model you've got gpt4 you've got three
37:17 Claude models which is this is anthropic
37:20 um the the ones that say subscribed here
37:22 or are only if you pay for the this
37:25 subscript Jesus this monthly
37:29 subscription for Poe uh and then you
37:32 have Claude instant claw GPT now
37:33 everything below that are Bots and one
37:36 of the cool things about Poe is there's
37:38 a little thing here where you can create
37:39 your own bot they make it really easy
37:41 for people to make Bots
37:43 I've made a bunch I made a sling blade
37:46 bot and a Yoda bot and a Shakespeare bot
37:48 but two other people made mid-journey
37:51 bots so this one's called MJ captain and
37:54 I think you said you you want
37:57 Graphics to use for Design Elements so
38:00 um I want a sheet of design elements for
38:07 modern
38:09 Graphics design
38:12 let's just see what that does so you
38:14 type it in there and then it'll rewrite
38:16 your crappy little prompt with one
38:18 that's that's quite specific so this
38:20 says a sheet of design elements for
38:22 modern graphic design with minimalist
38:25 aesthetic inspired by the art of uh
38:27 wasili Kandinsky which is kind of cool
38:30 using a monochromatic color palette with
38:32 pops of bright yellow typography Mastery
38:35 by Paula share with a focus on geometric
38:38 shapes such as circles squares triangles
38:40 incorporating so so this is actually
38:42 kind of interesting so what we'll do now
38:44 I'm just going to copy this
38:47 and if you want to go to that that one
38:49 in particular is called MJ Captain so if
38:51 you go to poe.com
38:53 MJ Captain that'll take you to this
38:56 mid-journey bot there's another one
38:58 that's just slash mid Journey so po.com
39:00 slash mid Journey that'll take you to
39:02 these Bots they're kind of slick so now
39:06 we're going to hop over to Discord
39:08 so back to Discord for those of the it
39:11 sounds like there's some new people here
39:12 my name is Kyle Shannon this is the AI
39:13 learning lab welcome we're just geeking
39:16 out right now on Mid Journey someone
39:18 wants to do some designing
39:20 some fancy design and so I am in the AI
39:25 Salon
39:27 um Discord and within our Discord we've
39:29 got a mid-journey channel it's called
39:31 mid or mid-journey play so I'm going to
39:34 scroll here to the bottom Ali has been
39:35 very busy she's been doing a lot in here
39:37 these are cool
39:39 I mean this is some of the quality that
39:41 I mean mid Journey's quality is just
39:43 sick
39:44 it's it's just really really good man
39:46 she's been making a lot oh that's a cool
39:49 one
39:51 um and so I'm so the way you make an
39:54 image in mid-journey is you type slash
39:56 imagine
39:58 and then it will pop up a prompt box and
40:01 I just copied that from the pobot
40:04 so I'm just gonna pop it in there
40:07 and hit return
40:09 and so now that's been submitted to
40:11 mid-journey and within a few minutes
40:13 it'll start generating four different
40:15 images that are based on that and we can
40:17 we can watch them here together if it
40:19 takes too long I will uh yeah there it
40:21 comes so so one of the things you'll
40:23 notice the way these diffusion models
40:25 work with image generation
40:27 is they start out as kind of blurry
40:29 images and then they do a series of
40:32 denoising algorithms on top of them so
40:35 they've kind of averaged together like
40:37 you know hundreds if not thousands of
40:39 images into these blurry noisy images
40:42 and then they just denoise them in the
40:45 direction of that prompt that I asked
40:46 for
40:48 and this is looking like it's it's it
40:50 this is going to be pretty cool and you
40:52 could obviously change the colors that
40:53 were mentioned in that prompt or the
40:55 shapes or the artists that were
40:57 mentioned
40:58 um these models understand a lot I'm
41:01 going to show you one more research
41:02 source for that but let's go this is at
41:05 78 this is at 93 there's usually the the
41:08 jump from 93 to finished is usually
41:11 really strong and really good so we'll
41:13 see what it does here yeah it just it
41:15 really tightened it up so let's go we'll
41:17 go kind of zoom in on these
41:20 so
41:21 I mean
41:22 if you want design inspiration or if you
41:25 wanted to do say patterns for something
41:27 on Etsy for example
41:30 um yeah
41:31 these are pretty slick they're really
41:33 good then what you can do so when you're
41:36 I want to go back to Discord so when
41:38 you're in Discord
41:40 so the mid Journey bot comes back and it
41:42 gives you this four up you know it's
41:45 four different designs and then they put
41:47 these weird little wonky like you one
41:50 two three four V one two three four and
41:53 then the reload button if you just hate
41:55 all these and you just want to re-roll
41:57 it that's what that button is the U
41:59 stands for upscale and then it's one two
42:02 three four so if you want to upscale any
42:05 any of these you can but what I
42:06 generally do is I'll find something I
42:08 like like I like one and four for me
42:11 kind of speak to me so what I'll
42:14 generally do is I'll say give me some
42:16 variations the v stands for variations
42:18 give me variations for one and
42:21 variations for four
42:24 and it'll basically pop a new prompt in
42:26 there and it'll it'll it'll keep them
42:28 basically the same but give me slight
42:30 variations and then I'll usually upscale
42:32 from there
42:36 so let's see is this one done now that
42:39 one's still going
42:44 yeah there's slight differences here
42:47 like the little circles here are
42:48 different these little pieces are
42:50 different like this one's got a little
42:51 half tone thing going on
42:53 there's 83 percent
43:00 good night I watched some guy reload his
43:02 computer about 30 times
43:04 how long did you do that for I don't
43:06 know four hours or so
43:10 all right so yeah this is cool like like
43:15 this one's got like little trees
43:19 I like that one so I'm gonna we're gonna
43:21 do U1 we're gonna upscale one on this
43:23 sheet
43:24 and then down here
43:27 I don't know I like the fourth one we're
43:29 gonna upscale four
43:34 and so there's the upscaled version for
43:37 for that one
43:39 and then there's the upscaled version
43:41 for that one so
43:43 so that's kind of how you do mid journey
43:46 and that's that's a way to do those
43:47 prompts there's the other thing I
43:49 promised you I would show you another
43:51 resource if you go to open art
43:56 open
43:59 art
44:00 dot a i
44:07 just making sure this is right yeah go
44:09 to
44:10 openart.ai prompt book open art dot AI
44:15 slash promptbook
44:17 it will take you to this thing now this
44:19 is
44:20 um ostensibly made for stable diffusion
44:22 I wonder if they've got one
44:27 for mid-journey
44:33 anyway
44:35 um this is like a 96 page
44:38 PowerPoint presentation that teaches you
44:42 all about prompting for these image
44:45 generation tools a lot of what you learn
44:47 in here can be applied to any of the
44:49 image generation tools there might be
44:51 slight differences in in like you know
44:54 how you call certain elements but you
44:57 know start by asking a list of questions
44:59 do you want a photo or a painting what's
45:01 the subject is it a person is it an
45:02 animal is it a landscape what details do
45:05 you want to add special lighting
45:06 environment color scheme Point of View
45:08 background and then you just keep going
45:11 through this thing and it's just it's
45:12 just like it keeps peeling the onion so
45:14 they're showing you examples of what
45:16 might change it's talking about the
45:18 order of words being important and then
45:21 it then it kind of goes into this world
45:23 of modifiers and this is for me this was
45:25 like an eye-opening
45:27 um section because
45:30 like just within photography this this
45:33 first page you know photography prompt
45:35 mix and match it understands so many
45:38 different terms like close up Polaroid
45:41 monochrome long exposure
45:43 um ring light
45:45 lenses it understands what a telephoto
45:48 lens is or a 24 millimeter lens it
45:50 understands what bokeh is and if you
45:52 keep going it shows you here are the
45:54 different kinds of images that you'll
45:56 get if you use these different terms and
45:59 so this document is just like like
46:02 rabbit hole after Rabbit Hole after
46:04 rabbit hold it whether you're using
46:06 mid-journey or stable diffusion or dolly
46:08 or any of them
46:09 you can just start digging in use
46:12 different art mediums right watercolors
46:14 graffiti oil painting clay sculpture
46:17 um you know paper cut it understands
46:19 artists and and different categories of
46:21 artists and you know it's not just
46:23 modern artists I know there's a lot of
46:25 bad press about it stealing people's
46:27 artwork but but it understands like
46:30 historical artists like you can like you
46:32 can get an art education just just you
46:36 know digging through the different art
46:37 styles and getting these image
46:39 generation tools to create things for
46:41 you so I know that was a long answer and
46:44 uh it looks like we got some new folks
46:46 here let me flip my camera around and
46:48 introduce myself if you're new my name
46:50 is Kyle Shannon this is the AI learning
46:52 lab I'm an entrepreneur I'm the CEO of a
46:54 company called storyvine I um
46:57 I let me change one thing here oh wait
47:02 let me flip it back over hold
47:04 um
47:06 I found a thing called the AI Salon
47:08 which if you want to join a community of
47:11 people curious about AI we've got a
47:13 Discord server we meet every other week
47:14 and just geek out on this stuff
47:16 and I also have an AI newsletter that's
47:18 our Twitter handle if you go to Twitter
47:20 and go to Everyday AI news uh you can
47:23 you can get a link there
47:25 um and yeah what this channel is about
47:28 is really just demystifying AI I'm an
47:30 entrepreneur more than an AI specialist
47:32 I've only been doing the AI thing for
47:34 about a year so I am new to AI but but
47:37 I'm very experienced with playing with
47:40 things like the early days of the world
47:41 wide web and and now I'm geeking out on
47:44 the early days of AI because I think
47:45 it's very very similar in parallel
47:49 um can I write code and sell it if if
47:52 you write code that does something cool
47:54 and people want to buy it sure or if you
47:57 want to write code and people will pay
47:59 you to write that code yes I got a
48:01 children's book by chat GPT but I don't
48:03 know where or how to get the
48:05 illustrations for the
48:07 for the something for the book I assume
48:10 um what I just showed you
48:12 um mid-journey you can also ah dang it I
48:17 missed my dual cameras
48:20 um
48:21 another place that you can make images
48:23 for free
48:25 is if you go to bing.com
48:29 and then you go bing.com create
48:34 slash create
48:37 this is uh Microsoft's front end on top
48:41 of Dali which is the open AI image
48:44 generation tool I think this is by far
48:46 the weakest of the tools but it's free
48:50 and easy so bing.com create and what
48:54 happened there
48:56 what happened that was weird
48:58 hold
49:00 all right I don't know what happened
49:01 there okay uh so I'm gonna go here I'm
49:03 gonna go
49:04 um children's
49:08 book
49:09 illustration of a
49:13 chubby black
49:16 and white
49:18 dog with a red
49:22 hat walking
49:25 down Main Street
49:30 that should make us a chubby dog walking
49:32 down Main Street in some sort of
49:34 children's
49:35 um illustration look we'll see
49:40 Bing's always a crap shoot I feel like
49:42 mid-journey mid journey is one that when
49:44 I go there and show people it's okay
49:46 that's not a chubby dog that's like a
49:48 Macy's day parade
49:51 oh Beast dog
49:53 so let me get rid of chubby let me take
49:55 chubby out of there and re recreate so
50:02 um but yeah so so this is a way that
50:03 that you could make images for your book
50:06 um is it okay to store AI text responses
50:09 on my oh to my like to my database I'm
50:14 using one with their API service I think
50:17 I think using their API service is is
50:20 safer than chat GPT because chat GPT is
50:23 just a a really thin quickly put
50:27 together piece of software it's getting
50:28 better but right now it's kind of it's
50:30 kind of
50:32 you know doing that why am I dead naming
50:35 Muhammad Ali my Cassius Clay shirt
50:38 this is this is Basquiat baby
50:42 that's Basquiat
50:48 is this backwards for y'all or is this
50:51 normal because if it's
50:53 are you are you able to read this or the
50:55 letters backwards
50:57 let me know because I'll fix it
51:01 dead naming Muhammad Ali I don't even
51:03 know what dead naming means Let's go ask
51:05 Chet GPT what dead naming means
51:09 new chat hang on let's you get to
51:12 witness this too oh it's backwards okay
51:14 wait so I'm gonna mirror my video
51:16 bang all right is that better is that
51:19 normal
51:20 you can read it now Cassius Clay
51:24 yes yes
51:26 that's Basquiat isn't that a cool
51:27 illustration come on come on how cool is
51:30 that
51:31 but I don't know what dead naming is can
51:34 you give thoughts on mid-journey yeah I
51:35 just did mid Journeys the bomb
51:38 mid-journey is is of of the image
51:41 generation tools it's the one oh let me
51:43 flip my camera around
51:45 not comment
51:47 I don't want to comment I want to flip
51:49 my camera come on
51:53 come on Tick Tock uh what
51:55 is dead naming
51:59 dead naming refers to the act of using
52:01 the birth name oh or previous name of a
52:05 transgender or non-binary person well
52:07 that's not Muhammad Ali
52:11 all right it's a it's a it's a gender
52:14 thing I don't know but dead naming but
52:16 I'm not dead naming him I'm just wearing
52:18 a t-shirt because I like Basquiat and I
52:20 like Muhammad Ali
52:22 all right flipping the camera back
52:25 around hey everybody my name is Kyle
52:26 Shannon AI learning lab here happy to
52:29 chat about anything I I ask me questions
52:31 you're also free to troll me and tell me
52:33 that I look like Rosie O'Donnell because
52:35 you do anyway so might as well invite it
52:41 I think my favorite was I was scrolling
52:42 Tick Tock and thought I saw Alec Baldwin
52:44 but he got fat that's a that's a
52:47 compliment all right
52:51 can I get my thoughts on mid-journey I
52:54 think it's the best of all of the image
52:55 things although
52:58 um the drag about mid-journey is that
53:00 you still have to oh there's the
53:01 Basquiat see
53:06 um
53:07 the thing about mid-journey is you have
53:08 to do it within Discord so if you're not
53:11 comfortable with Discord not only do you
53:13 need to learn mid-journey you need to
53:14 learn Discord which for normal human
53:16 beings is a pain in the ass it sucks
53:20 um I know that they're going to be doing
53:22 a website at some point but I wish
53:24 they'd get [ __ ] on it
53:27 um
53:27 stable diffusion
53:30 also really powerful because it's open
53:32 source the developer Community is
53:34 developing all sorts of stuff for it but
53:37 you still got to be pretty geeky if you
53:39 want to use mid journey and you're not
53:41 geeky then check out get on the wait
53:43 list for if it's still wait listed a
53:45 thing called leonardo.ai
53:48 let me flip the camera back around here
53:53 leonardo.ai
53:55 and it is a super cool
53:59 um
54:00 web front end for stable diffusion
54:03 that's very very sophisticated it allows
54:05 you to even do things like train
54:09 um
54:11 let's see fine-tuned models is this yeah
54:14 your models so like I trained it on some
54:18 of my images but I also trained it I'm a
54:20 photographer so I trained it on
54:23 um on some of my cloud photographs and
54:26 now I can say use this model
54:30 and so anything that I create in here
54:33 will basically use my photographs as
54:35 kind of a basis so if I say
54:38 um dramatic
54:41 Sunset over the Rocky Mountains
54:47 wait dramatic
54:51 Sunset over the
54:54 Rocky Mountains my dog is
54:57 pissed
54:59 it will now do a picture but it should
55:04 if it's still working right include my
55:06 dramatic cloud cloud Skies that I do
55:14 with the actual camera
55:19 what'd you do tonight oh waited for a
55:22 guy loading a web page that sounds
55:24 exciting how long are you doing that for
55:26 three hours
55:27 interesting life you're leading thanks
55:30 Marge
55:32 oh look at the drama look at the drama
55:36 in Disguise oh it sort of cut that cloud
55:39 off vertically that's weird
55:42 that does that does not look dissimilar
55:44 to stuff that I photographed and how I
55:46 retouch it so anyway
55:49 um that's that I would also if you're if
55:52 you're in the commercial
55:54 um
55:56 you know if you if you work for a big
55:58 company
55:59 um I would look really strongly at
56:02 um Adobe Firefly and the new generative
56:06 AI within Photoshop because it's really
56:08 good
56:10 um and the thing about what adobe's
56:12 doing is they've trained their models on
56:15 um public domain and Adobe stock images
56:18 that they own the rights to so they've
56:20 made sure not to train their their base
56:22 model on copywritten stuff
56:25 so that means that images generated
56:29 through adobe's generative AI are going
56:31 to be safe for commercial purposes which
56:33 if you work for a big company all this
56:35 all these other ones are a mess right
56:37 now
56:38 so that's my thoughts on mid-journey you
56:41 asked me about one
56:42 uh image platform you get an answer on
56:45 four why because that's why we're here
56:48 I don't know if if RFID what was that
56:52 hold is this something about a chip in
56:54 my skull if our ID is not perfected how
56:57 can we trust AI
57:00 and in in in an OS
57:04 you don't have to trust AI
57:08 like you should be skeptical of it but
57:10 what you should not do is just dismiss
57:12 it
57:14 from a soapbox what you should do is go
57:16 start playing with it and understand
57:17 what it's good at and what it's not good
57:18 at and learn about it
57:20 in my opinion
57:22 um
57:23 when do we get data ingester yeah I
57:25 assume you're talking about code
57:27 interpreter from openai I want that like
57:31 yesterday and I don't have it yet and
57:33 it's got me a little cranky okay
57:37 [Applause]
57:38 I interview people for podcasting and
57:41 the new AI audio tool from Adobe is
57:43 amazing yeah the one so Adobe has this
57:46 free tool that they just put out where
57:48 you record stuff on your shitty little
57:50 phone microphone and you run it through
57:51 their tool and it makes it sound like
57:53 you used a studio mic sounds like you're
57:55 an actual professional which you know
57:57 that's one of the themes of these of
57:59 these generative AI tools and by the way
58:01 I'll thank you for tapping the screen I
58:03 think that lets people know we're here
58:04 so I appreciate that
58:06 um and and and super cool that you're
58:08 all hanging out with me
58:10 um the um
58:13 what was it talking about oh oh the the
58:16 um
58:17 the tools making us more professional
58:19 than we otherwise might be I I think
58:21 that's one of the things that I'm
58:23 noticing about these tools is rather
58:25 than I think the perception from the
58:27 outside is oh the ai's doing the
58:30 creativity or the AIS making art no it's
58:33 people making art using these incredibly
58:36 powerful tools so what I'm kind of
58:37 noticing for myself personally but what
58:39 I'm seeing for others is that the AI is
58:42 like a Humanity amplifier
58:44 so it's like you have an intent and
58:46 where before if you didn't quite have
58:48 illustration skills you don't need them
58:50 now if you can kind of articulate what
58:52 you want bang there's your images bang
58:54 there's your professional audio bang
58:56 there's your music soundtrack bang
58:59 there's the script
59:01 um so you can really quickly move from a
59:05 vapor of an idea to an articulated
59:08 version of it and what I'm finding is
59:11 that
59:12 with AI I'm no longer hitting any
59:15 creative roadblocks right like like ai's
59:18 like a creative roadblock mower Downer
59:21 bulldozer might be a better term for
59:24 that
59:25 um it just sort of bulldozes through
59:28 what would have been sort of creative
59:30 blocks or like I don't I don't have the
59:31 skills to do that or anything like all
59:34 that whiny [ __ ] you do as a creative
59:35 person
59:36 and it just mows that [ __ ] down and I
59:39 find myself getting to kind of 80 of
59:41 what I want to get to incredibly quickly
59:44 and then I spend the last 20 percent
59:46 probably the same amount of time I would
59:48 have spent just getting to a first draft
59:50 now I just spend refining so it's it's
59:53 really it's really really powerful that
59:55 way I want a publicly owned human AI
59:57 hybrid app that will be coming sooner
1:00:00 than you know
1:00:01 much sooner like
1:00:03 that's coming I I mean I think that's
1:00:05 going to be
1:00:06 I think we're all gonna have like AI
1:00:09 assistance and we'll just say to our AI
1:00:11 assistant go do that [ __ ]
1:00:13 and our assistant might even be able to
1:00:15 do things like call on our behalf Google
1:00:18 in one of their developer conferences
1:00:20 this past year showed an AI example
1:00:24 where the AI called a pizza shop and
1:00:26 ordered a pizza on but called on the
1:00:28 phone and talked to a person and the
1:00:30 person didn't know it was an AI or if
1:00:31 they did they didn't act like they knew
1:00:33 it it was a demo so they probably said
1:00:36 hey Sal there's going to be kind of a
1:00:38 robotic chick calling pretty soon could
1:00:39 you just act like it's normal so it was
1:00:42 a demo so that was probably it but the
1:00:45 the synthesis stuff is is getting really
1:00:48 good really fast and so I think that we
1:00:50 will have agents that
1:00:52 um share our likeness if we want
1:00:56 um share kind of a modified like here's
1:00:58 the superhero version of me if we want
1:01:01 um and just go do stuff for us go book
1:01:03 me a vacation to Cancun it goes off and
1:01:06 books you a vacation to Cancun you look
1:01:08 at it you're like that's too expensive
1:01:09 find me something for half the price and
1:01:12 uh you know give me activities for the
1:01:15 day create an itinerary for the day okay
1:01:16 boss off it goes right I think we're
1:01:20 gonna be there
1:01:21 sooner Discord is so complicated Discord
1:01:24 is a [ __ ] nightmare so if you use
1:01:27 slack at work
1:01:28 Discord was the original slack right you
1:01:31 know like slack stole from Discord and
1:01:34 it was for gamers like so it's like
1:01:36 these little you know chat communities
1:01:38 and things like that
1:01:39 and so unfortunately it got popular with
1:01:42 you know Gamers and then with nfts and
1:01:44 blockchain and then a bunch of the nft
1:01:47 Bros were like now let's go do some AI I
1:01:49 know we'll start it in a Discord server
1:01:51 and I have a Discord server for the for
1:01:54 the AI salon so I can't [ __ ] it's a
1:01:56 it's a nightmare
1:01:58 all right oh my God Leonardo Leonardo's
1:02:01 really cool
1:02:02 Leonardo is really cool
1:02:04 um well a will AI have the impact
1:02:07 projected I think AI is gonna have
1:02:10 a massively profound impact I don't know
1:02:13 what projections you've been looking at
1:02:17 um it is clearly over hyped
1:02:23 I think it's more profound than what the
1:02:25 hype is right now and I think it's
1:02:27 coming way faster than anyone has a clue
1:02:30 so let me give you let me give you a a
1:02:32 comparison so when I started agency.com
1:02:35 in the mid 90s this this was kind of
1:02:37 early days of the World Wide Web
1:02:39 to get to 100 million users it was six
1:02:42 years from when Tim berners-lee created
1:02:45 the world wide web until it hit 100
1:02:47 million users it was six years
1:02:50 chat gbt got to 100 million users in six
1:02:53 weeks
1:02:55 they launched it on November 30th and
1:02:58 mid-January it hit 100 million users it
1:03:01 was the fastest adoption of Technology
1:03:03 of a single technology in history
1:03:07 um
1:03:08 like that that's remarkable and then if
1:03:11 you look at what it does if you sort of
1:03:13 compare and contrast let's go back to
1:03:14 the World Wide Web what was the product
1:03:16 of the World Wide Web well it was a
1:03:19 thing that looked like a page it looked
1:03:21 like what is this can I show this yes it
1:03:24 was a thing that looked like this
1:03:27 and then some of the letters were blue
1:03:29 and you could click on them and it would
1:03:32 jump you to another page that looked
1:03:34 like that was that was all it was it was
1:03:37 pages with hyperlinks that clicked you
1:03:39 to other hyperlinks
1:03:41 and it's the foundation for all of the
1:03:43 stuff we do today on the internet right
1:03:45 it's the foundation of social media it's
1:03:47 the thing that powers the iPhone and all
1:03:49 of our smartphones you know it's the
1:03:51 foundation of of everything we use today
1:03:54 what is the output of these AI tools
1:03:56 holy [ __ ] [ __ ] it will write you a
1:03:59 novel it will write you a computer
1:04:00 program it will translate languages it
1:04:03 will make you beautiful images it will
1:04:04 make you beautiful songs and we're just
1:04:07 scratching the surface of what these
1:04:09 tools make possible and because so many
1:04:11 people
1:04:13 jumped on chat GPT the entire developer
1:04:17 Community like did this pivot it just
1:04:20 went
1:04:21 and everyone is putting all of their
1:04:24 energy all of the investment money into
1:04:27 this AI stuff
1:04:29 and then the AI is improving itself so
1:04:32 you have all the developers pointing at
1:04:33 it you have all the business people
1:04:35 pointing out at it going how am I going
1:04:36 to use this that's where I fit and then
1:04:39 and then you have the tool itself
1:04:40 improving on itself so we are in an
1:04:44 accelerated Innovation phase that we're
1:04:47 at the very beginning of
1:04:49 the other the other parallel to the
1:04:51 early days of the web the internet had
1:04:53 been around for decades
1:04:54 the World Wide Web comes along and makes
1:04:56 it available to real people right it was
1:04:59 all scientists and researchers up to
1:05:01 that point the World Wide Web makes it
1:05:03 available for everyone same with
1:05:06 um with machine learning and artificial
1:05:08 intelligence the the foundations for it
1:05:10 or 50 years old you know machine
1:05:12 learning and kind of let's call it you
1:05:14 know modern AI 15 20 years probably
1:05:18 um it's already in a lot of the stuff
1:05:20 that we do but what did chat CPT do and
1:05:23 what did mid-journey and stable
1:05:24 diffusion do
1:05:25 they made it available to the rest of us
1:05:27 right so so we're just at the beginning
1:05:30 of this inflection point where real
1:05:32 people get to use the most powerful
1:05:35 tools in the history of mankind
1:05:38 is that going to have a profound effect
1:05:40 yes are you a copywriter go play with
1:05:42 chat EPT and tell me you're not freaked
1:05:44 the [ __ ] out are you an illustrator for
1:05:46 30 years go play with mid journey and
1:05:48 tell me you're not freaked the [ __ ] out
1:05:50 the woman that co-hosts the AI Salon
1:05:52 with me Leah fasten is she's been a
1:05:54 photographer for 25 years
1:05:57 um when when we started the salon she
1:05:59 hadn't played with generative Ai and
1:06:01 then she started playing with it and her
1:06:03 initial response was holy [ __ ] and then
1:06:05 her second response was I'm gonna be
1:06:07 okay I got through the the you know
1:06:09 transition from film to digital and now
1:06:12 I think she's back at holy [ __ ]
1:06:14 because the speed with which you could
1:06:17 take one of her images
1:06:19 and copy the style of it and not need to
1:06:23 hire her is like that now
1:06:26 so what does that mean I don't know
1:06:28 she doesn't know she's she's having to
1:06:30 find
1:06:32 I mean she's still working right she's
1:06:34 got a reputation she's really good at
1:06:35 what she does but she's having to figure
1:06:37 out what does the future of Photography
1:06:40 look like as well as what does it look
1:06:42 like for her given that she knows how to
1:06:45 use these tools
1:06:46 so yeah that and and and
1:06:51 it's not all doom and gloom right
1:06:54 it's not that
1:06:57 the people that get displaced are just
1:06:59 going to sit around and do nothing the
1:07:02 people that get displaced or disrupted
1:07:05 also have access to these tools so what
1:07:07 might end up happening is people that
1:07:09 had kind of shitty repetitive
1:07:11 soul-crushing jobs that job gets
1:07:14 automated out by AI
1:07:17 and then when they get displaced they
1:07:19 might go huh you know I kind of hated
1:07:22 that job you know what I've always
1:07:24 wanted to do and then here's this AI
1:07:26 tool that can help them do it so I think
1:07:28 we're going to see lots of reinvention
1:07:31 right lots lots of few Humanity facing
1:07:34 the existential crisis of what the [ __ ]
1:07:36 do I do now and then coming out the
1:07:38 other side going you know what I'm gonna
1:07:40 do and they're going to start companies
1:07:42 and they're going to start they're gonna
1:07:44 they're gonna innovate and they're gonna
1:07:46 create competition to their old
1:07:48 companies so yeah I think it's going to
1:07:50 be
1:07:52 um
1:07:53 I think it's going to be insane
1:07:56 and I think it's coming faster than
1:07:58 we're prepared for
1:08:00 I could be wrong but having been through
1:08:02 this once
1:08:03 that's where I am
1:08:05 looks like we got some new people here
1:08:07 my name is Kyle Shannon this is the AI
1:08:08 learning lab I talk fast I have ADD I
1:08:11 cuss I have a dog that barks I learned
1:08:14 many years ago we we I I have always had
1:08:17 poorly behaved dogs in my life and and
1:08:20 the first two dogs that I got they were
1:08:22 very poorly behaved if in fact one time
1:08:25 we were walking the dog in Brooklyn
1:08:27 and this guy walks by with a little
1:08:30 miniature poodle and Max picks him up
1:08:33 and like starts spinning around with him
1:08:35 like he's a toy and the guy who's
1:08:38 talking about is like
1:08:41 um so that wasn't good so I after that
1:08:43 we brought in a dog trainer right and so
1:08:46 I'm like oh great you know we're gonna
1:08:48 finally have well-behaved dogs and
1:08:50 here's what you learn when you get a dog
1:08:51 trainer they don't train the dog
1:08:54 they train the people
1:08:57 and you have to have the discipline and
1:08:59 you have to not be codependent with your
1:09:01 dogs and you have to be the one that
1:09:03 says no you can't get on the bed
1:09:05 and that ain't me and that's not my wife
1:09:07 and so what we decided very early on in
1:09:10 our lives is we're just gonna have
1:09:11 shitty behaved dogs because we are bad
1:09:15 dog owners
1:09:19 uh it's all fun and games until your dog
1:09:23 picks up a poodle
1:09:27 that was a disaster all right
1:09:30 did they take your picture inside
1:09:32 picture option away yes
1:09:34 why did they take that away
1:09:37 I used to be able to have that image in
1:09:39 a little window by my head and I didn't
1:09:42 have to explain what the [ __ ] this
1:09:43 channel was all about now I gotta
1:09:46 now you got to actually communicate I
1:09:48 don't want to communicate I don't want
1:09:50 to work hard you don't understand
1:09:56 really do we really need the pipes that
1:09:58 loud
1:09:59 I don't want to work hard
1:10:04 without my window and window I want
1:10:07 pictures or tick tock
1:10:12 yes they did
1:10:14 all right
1:10:16 somebody just came in here they're like
1:10:18 who is this insane person
1:10:21 it's creative with your words in mind it
1:10:24 is then what do we do with our time well
1:10:26 km
1:10:28 we're human right so I I think that
1:10:32 you know the the optimistic future is
1:10:34 what we do with our time is we spend it
1:10:38 on
1:10:39 things where we can contribute more or
1:10:41 that give us more fulfillment
1:10:44 um
1:10:47 you know the the promise of the
1:10:49 Industrial Revolution was that the
1:10:51 machines
1:10:53 would give us back our time and and what
1:10:56 it did was it just you know enslaved us
1:10:59 to to use the machines to make more and
1:11:02 more product
1:11:04 um I think AI really does have the
1:11:06 possibility of
1:11:08 taking all of the complex parts of our
1:11:11 world
1:11:13 Insurance claim forms and
1:11:16 buying a car and buying a house and
1:11:18 doing your taxes
1:11:20 and curing cancer and like all of these
1:11:23 incredibly complicated things and you
1:11:25 can think of AI as a simplification
1:11:26 layer sitting on top of it
1:11:29 so if you figure
1:11:31 probably conservatively 80 percent of
1:11:33 the activity of human beings is dealing
1:11:36 with the cogs of the machine
1:11:40 so if we have a technology now that can
1:11:42 automate the cogs of the machine
1:11:45 what's left for us to do I don't know
1:11:47 [ __ ] be human
1:11:49 go give a concert connect with people
1:11:52 don't be so divisive
1:11:54 self-express create art create love
1:11:58 create music I don't know I don't know I
1:12:01 hope it goes that way
1:12:03 you know it could go the dystopian way
1:12:05 we're all just working for the machine
1:12:07 again but I but I really do this do
1:12:10 think that that that that that these
1:12:13 tools in my experience is it amplifies
1:12:15 my Humanity
1:12:17 and that's both the good and a bad thing
1:12:19 because humans are complicated and they
1:12:21 can be really shitty so it's going to
1:12:22 amplify the shitty as well but the
1:12:24 reason this channel exists is I want to
1:12:27 talk about amplifying the good what can
1:12:29 they do to make us more creative more
1:12:31 productive more interesting more
1:12:33 self-expressed
1:12:36 more confident
1:12:38 right because you know how if you have
1:12:39 an idea for something and then you try
1:12:41 to put it in the world but you're not
1:12:42 really good technically at the thing
1:12:44 you're like I don't want to show this to
1:12:45 anyone it's shitty
1:12:47 well you don't have to be that anymore
1:12:48 you can have an idea and you might not
1:12:51 be a good writer but you put it in chat
1:12:53 EPT and now you're a [ __ ] brilliant
1:12:55 writer
1:12:56 I want you to write like Hemingway well
1:12:57 I just [ __ ] wrote like Hemingway
1:13:00 I don't know
1:13:01 I want to illustrate okay now you can
1:13:04 illustrate this image
1:13:08 that image sitting behind those letters
1:13:10 let me turn off the letters
1:13:12 this was a really profound image for me
1:13:14 so I was doing this project with stable
1:13:16 diffusion
1:13:18 um called Kyle Shannon dreams it was a
1:13:20 bunch of self-portraits and I don't know
1:13:22 what prompt I I did I think I actually
1:13:24 found some other image I copied the
1:13:26 prompt I put it in I started modifying
1:13:28 The Prompt and after about 30 variations
1:13:31 of this this thing showed up
1:13:34 and
1:13:35 I had an emotional visceral response to
1:13:39 this and and it was this
1:13:43 if I were
1:13:46 if I were technically skilled enough
1:13:48 either at Photoshop or illustration
1:13:51 this is the kind of work that I would
1:13:53 create
1:13:55 and it just created it and I I just had
1:13:57 this thing I'm like that's my voice like
1:14:00 that's my creative voice right there but
1:14:02 I actually don't have the skills to
1:14:04 execute something like that from scratch
1:14:06 and stable diffusion did it
1:14:09 and this is after what 2 000 images or
1:14:11 something like that well 2000 that I
1:14:13 kept like this is fairly far into this
1:14:15 exercise
1:14:16 and and I had this Epiphany that that I
1:14:20 can find
1:14:22 a creative voice an expression of my
1:14:26 creative voice that's true to who I am
1:14:28 that I don't actually have the skills to
1:14:30 do and and that's [ __ ] magic I mean
1:14:35 it's magic wasn't it Carl Sagan
1:14:38 that said any sufficiently advanced
1:14:41 technology is indistinguishable from
1:14:43 Magic
1:14:44 that's what we have in our hands right
1:14:45 now
1:14:47 so that's what we do with our time
1:14:49 figure out the thing you've always
1:14:50 dreamed about doing go [ __ ] do it you
1:14:52 don't need like like
1:14:55 I don't know if you know that the the
1:14:56 Malcolm Gladwell thing I forget what the
1:14:58 book was he's got a couple of them but
1:14:59 the 10 000 hours thing right if you want
1:15:01 to get good at something put in your 10
1:15:03 000 hours
1:15:04 I think these tools allow you to get
1:15:06 good at something with a lot less
1:15:09 maybe a hundred put in a hundred hours
1:15:12 put in a thousand
1:15:14 if if we could get
1:15:16 masterful
1:15:18 and feel masterful in a thousand hours
1:15:21 instead of ten thousand hours
1:15:24 that's like what two or three months
1:15:26 rather than two or three years of our
1:15:28 life
1:15:29 think how much more you can do
1:15:32 so what do you do with your time I don't
1:15:34 know whatever the [ __ ] you want
1:15:37 so now
1:15:39 the next two to three years when all of
1:15:42 this disruption happens I think that's
1:15:44 going to be a [ __ ] [ __ ] show like the
1:15:46 the Shangri-La I'm talking about is kind
1:15:48 of on the on you know five years down
1:15:51 the road on the other side of the
1:15:52 disruption we're about to to you know
1:15:53 we're about to go through you know
1:15:56 just a gauntlet of hell as as job after
1:15:59 job gets disrupted
1:16:00 but I think and the reason this channel
1:16:02 exists is that if you lean into this AI
1:16:05 stuff and you're one of the people that
1:16:06 knows how to use these tools you're
1:16:08 going to be okay
1:16:10 there's going to be a lot of people that
1:16:11 are just like it's evil it's stupid I
1:16:15 don't want to do anything than you well
1:16:17 don't this is coming whether you want it
1:16:20 or not right this is in my opinion this
1:16:23 is a tsunami
1:16:25 right the the the uh
1:16:28 low tide has gone out farther than it's
1:16:31 ever gone out before it's really quiet
1:16:33 right now
1:16:34 why is that well
1:16:37 there's some [ __ ] building up out there
1:16:39 and I would rather be on the surfboard
1:16:41 on that wave than in my cozy beach
1:16:43 cottage so that's just where my head is
1:16:46 right now and I I just I'm excited about
1:16:48 it I'm both more excited than I've ever
1:16:51 been
1:16:52 and more scared than I've ever been you
1:16:54 know professionally personally like this
1:16:56 stuff is insane
1:16:59 that's what's happening looks like we
1:17:00 have some new people here sorry you
1:17:02 jumped in in the middle of a
1:17:03 philosophical rant there my name is Kyle
1:17:04 Shannon this is the AI learning lab I
1:17:06 generally just ask questions about AI so
1:17:09 ask questions about AI I'll do what I
1:17:10 can to answer them I'm kind of jumping
1:17:12 all over the place with the question so
1:17:13 if I miss your question sorry about that
1:17:17 um let's see what else so how do we
1:17:20 avoid continued enslavement I think that
1:17:24 I think that the the AI tools are going
1:17:28 to automate out a lot of the jobs
1:17:31 that are the Cog in the wheel jobs that
1:17:34 are the the that are the jobs that Crush
1:17:37 our souls and then I think we find ways
1:17:40 Collective ownership where we stay
1:17:42 slaves well I mean the good news on that
1:17:44 front J is that
1:17:46 um meta
1:17:48 released their their core large language
1:17:51 models to the developer community so
1:17:53 there is now a thriving open source
1:17:56 Community
1:17:57 um creating large language models that
1:17:59 are competing with the big boys so
1:18:01 that's that's coming
1:18:02 I'm not a tech person per se but I
1:18:04 decided to learn and now I'm making
1:18:06 bucks with AI that's awesome man I this
1:18:11 is the thing just start like the the
1:18:13 whole the like if there's a um
1:18:17 if there's a mantra that I have right
1:18:19 now with this AI stuff it's get curious
1:18:21 just get curious if you've not played
1:18:24 with chat GPT
1:18:26 go to that top URL that's the official
1:18:28 chat website
1:18:30 you don't have to pay for it you do have
1:18:32 to sign up and they ask you for your
1:18:33 phone number so if you're one of those
1:18:35 nft I don't want to dox myself people
1:18:37 you're not going to like it but if
1:18:39 you're one of those go to the next url
1:18:41 url down that's bing.com that's
1:18:43 Microsoft's if you click on the chat
1:18:46 button there that's gpt4 for free
1:18:48 connected to the internet and then
1:18:49 poe.com lets you play with six or seven
1:18:51 different models and create your own
1:18:53 Bots just get curious with this stuff
1:18:55 just start playing I think it I think it
1:18:58 will make a a big big difference for you
1:19:02 um I'm seeing it again and again and
1:19:04 again there's also just this really
1:19:05 magical moment where you go in thinking
1:19:08 you know what chat GPT is and then it
1:19:11 does whatever the [ __ ] it does and
1:19:12 you're like it can do that
1:19:15 like that moment is awesome wait a
1:19:18 minute could it also do this and then it
1:19:20 does and they're like holy [ __ ]
1:19:24 I wish I was I was uh I was shown a
1:19:27 bunch of like
1:19:28 I don't know third or fourth graders
1:19:31 um how to make images with with uh I
1:19:34 don't know what I was using was I using
1:19:35 stable diffusion whatever I was using
1:19:38 and
1:19:39 um so I'm showing them all this and then
1:19:41 at the end of like two hours the kids
1:19:43 are just all over the place they're like
1:19:44 it's like a special school for
1:19:46 hyperactive ADHD kids so they're just
1:19:48 like pinging off the wall so so
1:19:52 um at the end of the thing this this one
1:19:54 like 10 year old boy says to me goes
1:19:56 yeah
1:19:57 I'm doing a little programming I said
1:20:00 are you and he goes yeah yeah I said
1:20:01 what language are you programming and he
1:20:03 goes C sharp I was like oh that's cool
1:20:06 what are you doing he says yeah I'm
1:20:08 doing like a little clone of Flappy
1:20:10 Birds I'm like wow that's cool how's it
1:20:12 going he's like ah I got a little bug I
1:20:14 got little bugs I Gotta Throw That code
1:20:15 out
1:20:16 and I said well if you use chat GPT
1:20:20 and and he and he goes he goes yeah yeah
1:20:22 I used that and I said have you used to
1:20:24 try to debug your code he said I don't
1:20:26 understand and it was clear you hadn't
1:20:27 used it right so I'll do what I did with
1:20:30 him
1:20:31 so so I go over to the chat gbt you're
1:20:35 gonna make money
1:20:38 so so I came over here and I went I went
1:20:41 um
1:20:43 um write me a clone
1:20:47 of flappy oops not slappy
1:20:51 Flappy Birds
1:20:54 in C sharp
1:20:58 that will run in unity
1:21:03 period that was all I typed and then
1:21:07 it
1:21:09 started explaining stuff and then it
1:21:12 started writing code
1:21:13 and
1:21:17 it started started doing that
1:21:21 he watched the machine write code like
1:21:23 that and it was like a cartoon his mouth
1:21:26 literally did this he went
1:21:31 I've been in the Home Alone
1:21:33 I mean it was
1:21:34 and
1:21:36 and so typing this
1:21:39 he goes like that and he goes I'll be
1:21:41 right back and he runs down four flights
1:21:44 of stairs to get his computer and he
1:21:46 runs back up and he's out of breath
1:21:48 and and so so we pull up his code and
1:21:53 and I have him copy the code and we go
1:21:55 to Chachi PT and he pastes this code in
1:21:58 chat EPT and
1:22:01 um we ask it to to analyze his code and
1:22:03 see if it can find any problems and it
1:22:04 finds a typo basically and so he copies
1:22:07 that code pasted into unity and hits
1:22:10 play and Flappy Bird shows up
1:22:13 and he and he's he's just like he's just
1:22:15 like dumbfounded and he looks at me and
1:22:17 he goes
1:22:18 this changes everything
1:22:22 that's what this stuff is
1:22:25 that's what this stuff is that moment
1:22:28 and and it's like I don't know there's
1:22:31 something [ __ ] magic in it
1:22:34 but you have to you you have to kind of
1:22:37 find the thing that you know deep enough
1:22:40 that you can you can query it you know
1:22:44 you can ask it in such a way prompt it
1:22:45 in such a way that it gives you back a
1:22:47 response that that you're like that's as
1:22:49 good as something I would have written
1:22:51 or better right and when you have that
1:22:53 moment and especially when it like just
1:22:55 knocks it out in 30 seconds and you know
1:22:58 that would have taken you two days
1:23:01 all right AI must be seen as a gift
1:23:05 from the best of Consciousness yeah I
1:23:07 really do think I mean you know the the
1:23:10 sort of wrap that AI has is like the
1:23:12 robots are doing this [ __ ] the robots
1:23:14 are making
1:23:16 you know the robots are plagiarizing the
1:23:18 robots are stealing no no no the robots
1:23:21 this robot is a reflection of the output
1:23:25 of humanity that that it was trained on
1:23:27 so a lot of what it's doing is a
1:23:29 reflection of us like if you sort of dig
1:23:32 deep into it philosophically it's
1:23:34 actually it's kind of profound in that
1:23:36 way technology should be used for the
1:23:38 best of humanity I totally agree with
1:23:40 that Dell
1:23:41 um
1:23:42 and it's a tool right I think Microsoft
1:23:45 got this piece right they're calling it
1:23:48 co-pilot so GitHub co-pilot is like this
1:23:50 pair programmer they're taking that
1:23:52 co-pilot brand across all of Office 365
1:23:55 they're rolling it into Azure cloud
1:23:58 services
1:23:59 um they're rolling it into windows and
1:24:01 and
1:24:02 it's right you're the pilot right you're
1:24:05 in charge you're flying the plane you
1:24:07 just happen to have an incredibly
1:24:09 capable co-pilot
1:24:11 that's what the AI tool is it's your
1:24:13 co-pilot
1:24:15 all right thank you for being the voice
1:24:18 for for being a voice in the wilderness
1:24:20 well partner I'll tell you
1:24:23 I I appreciate the kind words thank you
1:24:25 very much less is more not coming out of
1:24:27 my mouth I just talk a lot all right
1:24:29 just pay someone twenty dollars to draw
1:24:32 that okay well
1:24:35 I would encourage you Tuan to
1:24:38 um go play with these tools rather than
1:24:41 um
1:24:43 like diminish them you know from the
1:24:45 sidelines because until you understand
1:24:47 what's actually going on like the more
1:24:49 you learn about this here's the thing I
1:24:51 spent about a year digging fairly deep
1:24:53 into blockchain and nfts and the deeper
1:24:55 I dug into it
1:24:57 the less real it felt all the companies
1:25:00 felt like
1:25:01 Bros just kind of scamming and they were
1:25:04 all kind of kind of inside talking and
1:25:06 they've done a tremendous amount of
1:25:08 damage to themselves from a PR
1:25:10 perspective because of that
1:25:12 um and you know some obvious some big
1:25:14 fuck-ups right
1:25:16 the opposite has happened with AI the
1:25:18 deeper I dig into it the more profound
1:25:20 it gets that doesn't happen
1:25:23 almost ever
1:25:26 right when you really dig into something
1:25:28 you you find the cracks right and are
1:25:30 there are there cracks with this AI
1:25:31 stuff oh yeah it's like raw it's like
1:25:33 duct tape and spit together like the
1:25:35 like I don't know if you know this chat
1:25:37 gbt was written in two weeks
1:25:39 It Was Written in two weeks and then it
1:25:41 got 100 million users in six weeks
1:25:44 why because they put a simple interface
1:25:46 on something profoundly powerful
1:25:49 so I wouldn't sit on the sidelines
1:25:51 because
1:25:52 it's technology that's way more way more
1:25:55 profound than what it may seem
1:25:57 especially if you just go it's not
1:25:59 perfect
1:26:01 it made a mistake I asked it to create
1:26:03 citations and they were all fake okay
1:26:07 check your work
1:26:09 some idiot lawyer this last week or two
1:26:12 weeks ago
1:26:13 had chat GPT write a legal brief and he
1:26:15 handed it to a [ __ ] judge without
1:26:17 proofreading it all the citations were
1:26:19 fake
1:26:21 the judge was like what the [ __ ] is this
1:26:23 that lawyer should be debarred
1:26:26 like if anyone should check their work
1:26:27 shouldn't it be a lawyer isn't that why
1:26:29 you paid them
1:26:31 it's like Jesus Christ
1:26:34 all right
1:26:37 uh I am behind on messages so if it
1:26:39 seems like I'm behind on messages it's
1:26:41 because I'm behind my messages my intent
1:26:43 was to just you know do the bottom ones
1:26:45 but now I'm kind of back to my old ways
1:26:46 I like going through each of the
1:26:48 questions so
1:26:49 me all right
1:26:51 um
1:26:51 do you have your favorite apps you may
1:26:54 have said that already I'll say it again
1:26:55 so I do
1:26:59 and and I'm I'm I'm really limiting
1:27:04 um just out of sheer survival
1:27:07 so those top three URLs right there are
1:27:10 all just different flavors of chat gbt
1:27:13 um I am partial to the top one which is
1:27:16 the official chat website they're the
1:27:18 ones that invented it they've got the
1:27:20 most sophisticated model the gpt4 model
1:27:23 no one's come close to it so far
1:27:26 um I think each of those top three sites
1:27:29 is valuable in its own way
1:27:32 um prompts.chat just teaches you about
1:27:34 prompting future pedia is a directory of
1:27:37 lots and lots of AI apps I I stopped
1:27:39 paying attention to them because there's
1:27:41 just too many of them so for the words
1:27:44 it's for me it's chat GPT and Poe
1:27:47 and then for for images it's
1:27:52 um it's mid-journey stable diffusion are
1:27:54 the are the two biggies but I'm but I'm
1:27:57 increasingly using Adobe Firefly because
1:28:00 it's safe for commercial use so if I
1:28:03 want to create an image that I'm going
1:28:05 to give to one of my clients
1:28:07 um I would use Adobe Firefly if I'm
1:28:10 doing stuff for myself I use mid Journey
1:28:11 as stable diffusion because they're just
1:28:13 like other they they are a Quantum level
1:28:17 better
1:28:18 um but adobe's going to get their [ __ ]
1:28:20 together really quick the the speed with
1:28:22 which Adobe moved from from
1:28:26 um
1:28:27 you know they they had a lot of AI tools
1:28:30 built into
1:28:32 um you know filters and things like that
1:28:34 within Photoshop but they were staying
1:28:37 way far away from the generative AI game
1:28:40 um one of the one of the lead AI guys um
1:28:43 John Knack I used to work with him and
1:28:45 and he's at Adobe now and he came and
1:28:47 spoke at the AI Salon back in
1:28:50 December January I think it was January
1:28:53 and they were just dipping their toes in
1:28:55 the water with stuff and the speed with
1:28:57 which they went from let's experiment
1:28:59 with this to here's Adobe Firefly to
1:29:02 [ __ ] it we're rolling that into the
1:29:04 Photoshop beta with generative fill now
1:29:06 they just rolled out generative recolor
1:29:09 I think they called it or adaptive I
1:29:11 think it's generative recolor in
1:29:13 illustrator that came out today
1:29:15 um
1:29:16 for a company their size and with with
1:29:18 as much at stake as they have right
1:29:20 because if you think about the Adobe
1:29:23 suite it's it's not so much about
1:29:25 generating original content is taking
1:29:27 raw materials that artists have created
1:29:30 and refining them right illustrators
1:29:33 that may be exception to that but but
1:29:36 um
1:29:37 man they're just they're just leaning
1:29:38 into it so so they're really interesting
1:29:40 as well and so for me that's kind of it
1:29:42 right now like I'm I'm leaning in on the
1:29:45 word stuff with chat GPT the image stuff
1:29:48 with those two and then I use zapier
1:29:50 which is an automation tool to to create
1:29:53 content generators and things like that
1:29:55 so those are those are uh
1:29:58 someone called me bro is an AI yeah man
1:30:02 gpt5 right here bet you didn't think it
1:30:05 would look anything like this did you oh
1:30:08 yeah
1:30:09 I'm your avatar welcome to your
1:30:12 nightmare
1:30:14 all right just curious what's the time
1:30:17 frame for this disruption you mentioned
1:30:19 I don't know well
1:30:22 so I think there was just what was the
1:30:24 what was the amount of jobs there were
1:30:28 I don't remember the total number
1:30:32 [ __ ] I can't remember 30 000 or 130 000
1:30:35 I don't know there were just some amount
1:30:37 of jobs announced that were lost and
1:30:39 four thousand of those were directly
1:30:41 attributed to AI
1:30:43 um
1:30:48 I think certain industries are going to
1:30:50 get hit first I think copywriting and
1:30:53 and I think I think sort of the most
1:30:56 tactical levels of creative execution
1:30:58 are gonna get
1:31:00 affected you know relatively
1:31:04 um quickly and dramatically
1:31:07 because I think that
1:31:09 um clients are just not going to be
1:31:11 willing to pay for
1:31:13 um you know a bevy of Junior
1:31:16 tactical people when they know that
1:31:18 these tools can do that work
1:31:20 so I think I think the lower level
1:31:22 tactical things are going to get hit
1:31:26 um
1:31:28 I think the sort of COG in the wheel
1:31:30 knowledge workers like data entry people
1:31:33 lower level data analyst people
1:31:35 researchers things like that I think
1:31:37 those are going to get consolidated
1:31:40 there's a guy I mentioned this on my
1:31:42 lives a lot but he's a guy you should
1:31:44 follow there's a guy named Paul rateser
1:31:47 r-o-e-t-z-e-r and he's got an
1:31:49 organization called the marketing AI
1:31:51 Institute and he's got a saying he said
1:31:54 AI is not going to replace your job but
1:31:56 someone using AI will
1:31:58 and so I think what's going to happen
1:32:00 very very quickly is that
1:32:04 is that
1:32:06 the owners of these businesses are
1:32:09 either going to be forced into using
1:32:11 this
1:32:12 so if I'm an agency
1:32:15 and I'm used to turning around Creative
1:32:17 Concepts in six weeks and charging time
1:32:20 and materials for that and my big client
1:32:22 comes to me and says
1:32:23 I now want you to turn those around in
1:32:26 two weeks and charge me 30 percent less
1:32:29 or I'll fire you
1:32:31 then I need to find a way to solve that
1:32:33 right so I think that kind of stuff's
1:32:35 going to start to happen
1:32:37 um and we're already already starting to
1:32:39 see it we're already we saw a major
1:32:41 agency holding company just announced
1:32:44 last week week and a half ago that
1:32:47 they're going all in on AI which
1:32:49 probably means some of these kind of
1:32:51 disruption
1:32:53 um
1:32:54 but I don't know
1:32:56 I mean there's still a lot of copyright
1:32:59 and intellectual property right issues
1:33:00 there's still a lot of privacy and
1:33:02 safety issues so it's possible that
1:33:06 um
1:33:08 AI remains kind of like in a little
1:33:11 ghetto off to the side for a year I
1:33:14 don't think it's going to be a year
1:33:17 um I was talking I did a talk with a
1:33:19 with a group of voice over actors
1:33:23 um about six weeks ago now
1:33:26 and the woman that that runs the group
1:33:30 and invited me to speak
1:33:32 um
1:33:33 she said the day before our seminar one
1:33:35 of her friends called her up and said he
1:33:37 just lost a twenty thousand dollar job
1:33:39 because they're going with synthesized
1:33:42 voice he was hired to do like a voice
1:33:44 tree you know when you call the voice
1:33:46 tree and the person says you know click
1:33:48 one for this and two for that that's
1:33:50 someone has to say that right
1:33:52 there's a twenty thousand dollar job and
1:33:54 they just said well we're gonna have ai
1:33:56 do that now
1:33:57 20 grand gone right
1:33:59 so I think you're seeing in little
1:34:01 pockets right now I think you're going
1:34:02 to start to see waves of it start to hit
1:34:04 business
1:34:06 um I could be wrong but I think it's
1:34:07 coming fast
1:34:09 thank you for following Everyone by the
1:34:11 way uh feel free to tap the screen if
1:34:13 you tap the screen and make the little
1:34:14 hearts apparently that lets people know
1:34:16 we're here and why wouldn't you want to
1:34:18 just hang out with me all night
1:34:20 [Laughter]
1:34:22 um thank you appreciate that and let's
1:34:26 see let me go back and find some
1:34:28 questions here
1:34:30 all right there's outliers I saw that
1:34:32 before oh here we go this is where I was
1:34:34 before
1:34:38 it's going to allow humans to become
1:34:39 even more creative I think that's right
1:34:41 I've got so so I did a I did a talk to
1:34:45 um creative mornings a few weeks ago
1:34:48 and one of the slides in my
1:34:52 in my deck is this one
1:34:56 I think we are entering the great
1:34:57 Renaissance
1:34:59 it was funny when I was in acting school
1:35:02 I could never remember how to spell
1:35:04 renaissance and I always like I always
1:35:06 just said in my head why is there a
1:35:08 [ __ ] eye in the middle of this word
1:35:10 it doesn't make any sense
1:35:12 and now it makes sense that's why the
1:35:14 eye was in there it was just waiting for
1:35:16 this moment
1:35:17 I think that
1:35:19 the creative explosion we're about to
1:35:21 enter is going to make the real
1:35:23 Renaissance seem quite quaint by
1:35:25 comparison
1:35:27 and I'm here for it and I you know I
1:35:30 really do think I really do think that
1:35:32 like the possibility here is the robots
1:35:35 can do the shitty jobs
1:35:38 giving us the freedom to ask the
1:35:40 question well what do I really want to
1:35:42 do
1:35:42 what do I really want to contribute
1:35:45 what would make a difference right
1:35:49 rather than
1:35:51 you know I have a job that matched my
1:35:53 major
1:35:54 and I did it for Mom and I've been in it
1:35:56 for 20 years oh and I really hated it
1:35:58 anyway I mean like there's a lot of
1:36:00 people in that position and so
1:36:02 some of these well all of the disruption
1:36:04 is going to be painful
1:36:06 but I think a lot of it is going to
1:36:08 force people to have a to have to have
1:36:13 that conversation with themselves what
1:36:14 the [ __ ] am I on the planet for why am I
1:36:16 here
1:36:18 oh you know what I always wanted to do
1:36:21 the cupcake thing
1:36:24 chat GPT could you help me start a
1:36:26 cupcake company sure
1:36:31 how do I buy a building
1:36:35 where do I buy bulk flour
1:36:38 right
1:36:39 like that's that's like in our very near
1:36:43 possibility
1:36:47 um Can AI kick my ass once they tie it
1:36:50 to a robot kick your ass if you see that
1:36:53 Boston Dynamics thing where they put
1:36:55 Chachi Pita one of those little creepy
1:36:57 dogs
1:36:58 the dogs talking
1:37:00 and they like ask it a yes or no
1:37:01 question it goes it like wags its head
1:37:04 back and forth I'm like oh [ __ ]
1:37:09 we are entering a post sci-fi world it's
1:37:12 like go back back in your mind through
1:37:13 all of your favorite Sci-Fi movies and
1:37:15 just start checking boxes oh [ __ ] we can
1:37:17 do that now oh yeah I did that yesterday
1:37:19 yeah we can't time travel yet but this
1:37:22 is kind of close
1:37:24 it's wild all right what do I think will
1:37:27 happen when millions of jobs get
1:37:29 displaced by AI I think that the next
1:37:31 two to three years are going to be
1:37:32 pretty [ __ ] ugly
1:37:34 um
1:37:36 two things will happen one is
1:37:39 just like
1:37:41 we dealt with the covid thing like the
1:37:43 covet thing was like this
1:37:45 instant dramatic overnight [ __ ] it work
1:37:47 from home and just globally we worked
1:37:50 from home right so we changed our
1:37:52 Behavior like that which I didn't think
1:37:54 was even in the realm of possible and it
1:37:57 just happened right so it's possible
1:37:59 that the disruption is high enough
1:38:01 that governments have to respond right
1:38:04 and that's where something like
1:38:06 Universal basic income might be a tool
1:38:09 and it was like when when Andrew Yang
1:38:11 was talking about all that when he was
1:38:12 running for president
1:38:14 I sort of got it but I like I wasn't
1:38:16 deep enough into this AI world to know
1:38:17 what the [ __ ] he was talking about and
1:38:19 now that I'm deeper into it and I see
1:38:21 like like
1:38:24 this is hitting some of our highest
1:38:26 waged people right this is knowledge
1:38:28 workers and creative workers and things
1:38:29 like that
1:38:30 you know are going to be displaced
1:38:33 sooner right
1:38:35 um than anyone expected and so you're
1:38:37 gonna you're gonna have a lot of people
1:38:38 displaced and and you know looking for
1:38:40 work so I think there may be there may
1:38:42 be a phase in there we've got to do
1:38:43 something
1:38:45 but then the thing we can anticipate is
1:38:48 what are the new jobs that are created
1:38:50 as a result of these Technologies when
1:38:52 you automate out all this other stuff
1:38:54 what emerges as the new job and and that
1:38:57 is I mean you know I mean you know as
1:38:59 well as I do like I know no one knows
1:39:01 right now but there will be jobs five to
1:39:04 ten years from now that we couldn't have
1:39:06 imagined would be a job and it might be
1:39:08 cool they might be bitching they might
1:39:11 just be the new Cog in the wheel I don't
1:39:12 know but new jobs will replace the old
1:39:15 ones
1:39:16 um and the other thing is you know
1:39:17 there's I think there's this fear that
1:39:19 like well it's going to replace all
1:39:20 these jobs and like you're gonna have
1:39:21 millions of people just sitting around
1:39:23 doing nothing like that's not human
1:39:25 nature It's Not Human Nature to just go
1:39:27 oh [ __ ] it I got laid off I guess I'll
1:39:29 never work again
1:39:30 you know that's fun for about a month
1:39:32 and then when you're out of Gilligan's
1:39:34 Island reruns you're like what the [ __ ]
1:39:35 am I doing with my life
1:39:38 oh maybe I should go play with these AI
1:39:40 tools that displace me holy [ __ ] I
1:39:42 didn't know they could do ah let me go
1:39:44 do something cool right
1:39:46 so I don't know it's gonna be I think
1:39:49 it's gonna be an ugly two or three years
1:39:51 doesn't it seem like a lot of what's
1:39:53 being called AI is just scraping
1:39:55 articles and sites for text well so it's
1:39:58 not AI it's machine learning and it's
1:40:00 not scraping
1:40:01 it's not scraping so so this is one of
1:40:05 the misconceptions about how these tools
1:40:07 work and I'll show you if you want if
1:40:09 you want I'm happy to do that I can
1:40:11 actually explain how chat GPT works and
1:40:13 show you how it works because once you
1:40:15 understand it I think it's important to
1:40:17 understand it
1:40:18 it's not copying and pasting copy
1:40:20 written work and smashing it together
1:40:22 it's not copying and pasting elements of
1:40:25 some artists work and and somehow
1:40:27 collaging it that's not how it works
1:40:30 it's actually generating original
1:40:32 content based on all of the knowledge
1:40:35 that it's learned which is essentially
1:40:36 all of the public internet
1:40:39 like all of YouTube and all of Wikipedia
1:40:41 and all the blogs and all the websites
1:40:42 and all that like as much data as open
1:40:45 AI could find and jam into this training
1:40:47 set they did
1:40:49 and so it's kind of like if you could
1:40:52 walk into the library and learn
1:40:54 everything in the library and just jam
1:40:57 it in your brain
1:40:58 and then someone says to you write me a
1:41:01 business letter for you know telling
1:41:04 someone I need my money back
1:41:06 you would just write it right are you
1:41:09 copying and pasting that from specific
1:41:11 documents no it's in your knowledge
1:41:14 dome right it's in your knowledge bucket
1:41:17 and you're writing it word by word and
1:41:19 that's how these things work so if you
1:41:21 want to I can show you that I mean feel
1:41:23 free to
1:41:24 um let me know I'll show you how how it
1:41:26 works
1:41:27 oh my God that's a great story of the of
1:41:29 the boy with the with the programming
1:41:31 thing it was amazing he was like oh
1:41:34 this changes everything
1:41:37 just amazing
1:41:39 um let me jump to the bottom here
1:41:41 word of advice Focus stop being so
1:41:45 varied focused no I've got add this is
1:41:47 who I am if you don't like it take a
1:41:49 hike this is who I am if I I appreciate
1:41:53 the feedback but
1:41:55 I can't not be me I gotta be me I gotta
1:41:59 be me
1:42:00 [Laughter]
1:42:02 all right it won't even swear and it's
1:42:04 programmed not to be ethical it's not
1:42:06 programmed not to be ethical and you can
1:42:09 get it to swear you just have to
1:42:10 jailbreak it or go use something open
1:42:12 sourced
1:42:14 I you know again I would strongly
1:42:16 encourage you to avoid soapbox
1:42:18 conspiracy theories from the sideline
1:42:20 get curious and start playing with these
1:42:22 tools so you can understand how they
1:42:23 work
1:42:24 because it's really easy to demonize
1:42:26 them and dismiss them
1:42:28 it's much harder to be optimistic and
1:42:30 jump in and go [ __ ] it scares me but I'm
1:42:31 doing it anyway because the shit's
1:42:34 coming whether you want it or not and
1:42:36 you'll be on your soapbox going well you
1:42:37 know they don't tell the truth
1:42:44 holy crap there's 130 people here
1:42:46 there's new people here hello my name is
1:42:47 Kyle Shannon apparently I'm not focused
1:42:49 and moved too much but that's who I am I
1:42:52 swear I make voices
1:42:54 um this is the AI learning lab I talk
1:42:57 about AI I talk about generative Ai and
1:43:00 business and how you roll those two
1:43:01 together I'm an entrepreneur I've
1:43:02 started
1:43:03 I don't know 15 or 20 companies a couple
1:43:06 of really successful ones I'm 11 years
1:43:08 into my current company called storyvine
1:43:10 which is a
1:43:11 um uh an automated video storytelling
1:43:14 platform let me flip my camera here and
1:43:17 show you this is a slide I did this is
1:43:19 this is the era I think we're entering
1:43:22 and
1:43:23 um this this uh fancy little Photoshop
1:43:28 image why I put this up if you haven't
1:43:30 played with chat gbt that top URL is uh
1:43:34 that's the official chat CPT website
1:43:37 um start go start playing with it and
1:43:40 and if you don't feel like giving them
1:43:41 your phone number go to bing.com this is
1:43:44 Microsoft search engine but click on the
1:43:46 chat button
1:43:48 and that is gpt4 that's chat gpt4 for
1:43:51 free connected to the internet that's
1:43:53 Bing and then poe.com lets you actually
1:43:56 compare like six different
1:43:58 um chat models and you can create your
1:44:00 own chat bots so that's why I have these
1:44:01 three here and then prompts.chat will
1:44:04 teach you about prompting and future
1:44:05 pedia is just a big directory of tools
1:44:08 I'm not affiliated with any of them I
1:44:10 don't care if you go use them or not but
1:44:12 I strongly
1:44:14 strongly suggest that you should because
1:44:18 we're at an inflection point in history
1:44:20 and we're right at the beginning of it
1:44:21 and anyone who learns even a little bit
1:44:25 about this AI stuff now is going to be
1:44:27 paid dividends moving forward
1:44:29 you're right about nfts I found AI to be
1:44:32 way more ethical than the dude Bros yeah
1:44:34 I know although you know what's ironic
1:44:36 is I think that AI is going to drive the
1:44:38 adoption of blockchain and nft
1:44:39 Technology because here's what here's
1:44:42 what here's here's a a hot take isn't
1:44:44 that how the kids say it these days
1:44:46 here's a hot take
1:44:48 um
1:44:50 2024 presidential election in the United
1:44:53 States this is going to be fun uh this
1:44:56 is going to be a [ __ ] nightmare right
1:44:58 um
1:44:59 the AI stuff is going to make
1:45:02 disinformation the likes of which we've
1:45:05 never seen
1:45:06 and deep fakes and we're going to see
1:45:09 some politician French kissing a gorilla
1:45:12 in the Cincinnati Zoo and it's going to
1:45:13 look real and that's going to drop two
1:45:15 days before an election like they're
1:45:17 going to be some [ __ ] going down
1:45:19 and I think after that there's going to
1:45:22 be a a massive demand for
1:45:26 we need to know if stuff's authentic
1:45:29 well how do you attract that well
1:45:31 you probably need some sort of
1:45:33 tokenization Technology like oh I don't
1:45:35 know say blockchain nfts right so that
1:45:38 so what those things do
1:45:41 is actually really interesting in a
1:45:42 future where you've got a I you know
1:45:45 verify identity and verify and if you
1:45:48 look at Sam Altman's new company what's
1:45:51 it called Wonder
1:45:53 hang on a sec
1:45:56 something coin World coin is that it
1:45:58 World coin
1:46:02 yeah
1:46:04 yeah so
1:46:08 this
1:46:11 is Sam Altman's new company just raised
1:46:14 100 million dollars for it
1:46:16 and it is creepy as [ __ ]
1:46:22 what is it
1:46:26 he sent you a little orb it's like this
1:46:29 chrome like softball sized thing and you
1:46:32 put it up to your eye and it verifies
1:46:34 that you're human and when it verifies
1:46:36 that you're human
1:46:38 they claim it gets rid of the biometric
1:46:40 data but it verifies that you're human
1:46:43 and then it gives you some World coin
1:46:44 tokens so it's giving you cryptocurrency
1:46:47 that is currency that says I'm human
1:46:50 so if we're moving into a world where
1:46:52 you can't tell who's human or not right
1:46:55 is this dude synthesized or real if
1:46:58 we're moving into that world and we're
1:46:59 moving there quickly
1:47:01 authenticity is going to be important
1:47:03 verification is going to be important so
1:47:05 even though the crypto crypto Bros and
1:47:08 the nft cats you know [ __ ] up by
1:47:11 getting greedy
1:47:13 um I think AI is going to be their self
1:47:14 salvation in the end
1:47:18 I'm learning how to code with python
1:47:20 using chat EBT that's awesome I use chat
1:47:22 to help me in my writing I absolutely
1:47:25 fact checked it trust but verify that's
1:47:27 it and you learn pretty quickly
1:47:30 where the boundaries are like if you're
1:47:32 writing something creative if I'm
1:47:34 writing a screenplay all I need to know
1:47:36 is does the screenplay suck or not and
1:47:39 if it doesn't that's all it needs to be
1:47:41 but if I'm writing a documentary
1:47:43 screenplay
1:47:45 and I need it to be actually factual
1:47:49 well then it's on me to to verify it and
1:47:51 listen there are going to be entire
1:47:54 Industries built around content content
1:48:00 um alignment right getting getting
1:48:01 things accurate so just just recognize
1:48:05 that we're really early right now and
1:48:06 the tools the tools are just like these
1:48:09 spit and duct tape pieces of crap right
1:48:11 now that happen to be the most powerful
1:48:13 tools we've ever had in the history of
1:48:15 humanity this is awesome info I
1:48:18 appreciate it thank you very much Andrea
1:48:19 I appreciate that can I talk about
1:48:21 protocols the the protocols companies
1:48:24 are putting in place
1:48:27 um
1:48:28 I don't quite know what protocols you're
1:48:31 talking about
1:48:33 um
1:48:35 the protocol some companies are putting
1:48:37 in place is like we refuse to let you
1:48:39 use chat GPT
1:48:41 some universities are doing that the
1:48:44 entire New York City Schools school
1:48:46 system
1:48:47 ban chat GPT they just reversed that
1:48:50 because they realized oh this isn't
1:48:52 going away this isn't just a cheating
1:48:55 tool
1:48:56 this is going to be like the Internet
1:48:58 it's going to be everywhere everyone's
1:49:00 going to have it
1:49:02 um
1:49:03 there's also some protocols that are are
1:49:07 things around safety and privacy
1:49:10 um there's not a lot of sophisticated
1:49:12 tools right now so I think I think I
1:49:14 think what Smart Companies are doing is
1:49:16 say let's experiment with this AI stuff
1:49:19 let's under understand where the
1:49:21 boundaries are and let's put it off to
1:49:22 the corner
1:49:23 until we get our policies worked out we
1:49:27 get the the legal things worked out like
1:49:29 I have a good friend that works for
1:49:31 Deloitte digital he's fairly senior at
1:49:32 the late digital and
1:49:35 um
1:49:36 they one one of the guys from there
1:49:38 actually came and spoke at the AI salon
1:49:40 and he said they were not even allowed
1:49:43 to use
1:49:44 um image generation tools generative AI
1:49:46 image generation tools on their work
1:49:48 computers so he was in the creative
1:49:51 department and he he would on his
1:49:53 personal computer on the weekends he
1:49:55 would use mid-journey to create concept
1:49:58 images and then he would mail it to
1:50:00 himself to his work computer
1:50:03 um now thanks to
1:50:05 um Adobe Firefly and it being trained on
1:50:08 on non-copyritten information
1:50:12 um now they're free to use generative AI
1:50:14 right so so but I'm not quite sure which
1:50:17 which protocols you're talking about
1:50:19 Adobe internal image science folks are
1:50:22 some of the sharpest folks in the world
1:50:24 yeah like the the speed with which like
1:50:26 like what's impressing me about them is
1:50:28 they're not just copying what stable
1:50:30 diffusion and mid Journey are doing like
1:50:32 with their with their typography where
1:50:34 they're doing gender generative AI
1:50:36 within type and things like that they're
1:50:38 pushing the boundaries some of this
1:50:40 stuff what what's impressing me is not
1:50:42 just the technology but how quickly
1:50:43 they're putting it in the world testing
1:50:45 it they've got Discord servers now where
1:50:47 they're getting feedback in real time
1:50:48 it's it's pretty impressive
1:50:52 um I've been I've been doing this for 90
1:50:53 minutes because I don't know what the
1:50:56 hell's going on here but anyway cheers
1:50:57 thanks for hanging out with me I hope
1:50:59 this helps thank you thank you
1:51:03 all right recruiters jobs are on the
1:51:05 line I had an AI voice call me and ask
1:51:08 me about a job that's happening that's
1:51:10 going to happen for sure
1:51:12 do you believe Architects will shift
1:51:13 from AutoCAD to generate AI or better
1:51:17 um I don't know that they'll move from
1:51:19 AutoCAD but if you think of AutoCAD as
1:51:21 this you know complicated what 40 year
1:51:25 old
1:51:27 you know piece of technology this
1:51:30 incredibly complicated incredibly
1:51:32 capable thing down here
1:51:34 what generative AI is going to do is
1:51:36 just going to pop it's just going to pop
1:51:37 a layer of Simplicity on top of it so
1:51:41 what what you'll likely be able to do
1:51:42 here let me show you something cool
1:51:46 um
1:51:47 I'm going to pop over to Facebook for a
1:51:49 second
1:51:51 because I know how to live
1:52:00 and I am going to go to
1:52:03 mid-journey official
1:52:06 so if you're on Facebook go to go to
1:52:08 mid-journeyofficial and then search for
1:52:11 architecture
1:52:14 and then just start scrolling
1:52:17 let me make this a little bigger so it
1:52:19 fills the screen
1:52:20 because I want my people to be taken
1:52:22 care of
1:52:24 so there's there's some buildings that
1:52:26 look like you know Star Wars characters
1:52:28 those are pretty cool
1:52:30 um
1:52:32 so so I think I think one area like
1:52:36 check out these Ames chairs Star Wars
1:52:38 themed aimed chairs like
1:52:42 like that would be cool to own right so
1:52:45 someone in the furniture design business
1:52:47 is going to be [ __ ] around in
1:52:49 mid-journey and go oh you know what
1:52:51 we're gonna go make that there's
1:52:52 Chewbacca
1:52:55 um
1:52:57 and then and then similarly with
1:52:59 buildings so I think on the concepting
1:53:01 side like look at that right so
1:53:03 someone's gonna say go find me a cliff
1:53:05 we're building that
1:53:07 um
1:53:09 interior design
1:53:11 you know I showed some of these the
1:53:13 other night I mean I mean this is just
1:53:14 crazy stuff so so I think on the concept
1:53:17 in front
1:53:19 these tools are going to allow
1:53:20 Architects and designers to think in
1:53:23 ways they never would have imagined
1:53:26 and then what's going to happen is AI is
1:53:28 going to be able to Pop
1:53:30 um
1:53:31 an automation layer on top of AutoCAD
1:53:34 right so I take one of those crazy
1:53:37 designs and I say okay well now how
1:53:39 would you run the plumbing in that
1:53:40 building how would you run where would
1:53:42 the elevators be how would they how
1:53:44 would the Mechanicals work in this
1:53:45 building
1:53:46 and
1:53:48 AI will just figure all the [ __ ] out
1:53:50 right we're we're already starting to
1:53:52 see this you know that there's there's
1:53:54 been
1:53:55 um previews of auto-generated tools for
1:53:59 Unity
1:54:00 um for Unreal Engine for Roblox for
1:54:04 Minecraft
1:54:05 um so any complicated piece of software
1:54:10 you're going to be able to put the
1:54:12 generative AI layer on top of it to
1:54:14 simplify it so a workflow in the future
1:54:17 might be I go to something like
1:54:18 mid-journey I think of wacky [ __ ]
1:54:22 and then I go plausible question mark
1:54:27 that would cost you 422 million dollars
1:54:30 okay how can I mix how could I have some
1:54:32 show me what it would look like for the
1:54:35 22 million dollar version of that
1:54:36 building out comes the image that's not
1:54:39 bad
1:54:40 um can you make it rounder there you go
1:54:42 okay how would you build that
1:54:45 here's your budget here's your estimate
1:54:46 here's your drawings
1:54:49 so is that does that make architecture
1:54:52 less creative no I think it takes some
1:54:55 of the shitty drudgery work and
1:54:57 automates it
1:54:59 and lets Architects live at a higher
1:55:01 creative claim
1:55:03 but so yes I think it's coming now will
1:55:06 that come next year probably not because
1:55:08 those are big hairy pieces of software
1:55:10 and people's lives are at stake like the
1:55:12 engineering has to be right right
1:55:15 but is it out of the realm of
1:55:16 possibility in five years no
1:55:20 no
1:55:21 and if I were in the you know
1:55:22 architecture software business right now
1:55:24 I'd be [ __ ] jumping on this
1:55:26 because someone who cracks that code
1:55:29 do my drawings
1:55:31 AutoCAD just knocks that [ __ ] out
1:55:34 um you know
1:55:36 change all the conduit to three-quarter
1:55:38 instead of half inch so you can
1:55:39 accommodate this different kind of wire
1:55:40 done
1:55:42 holy [ __ ]
1:55:44 estimates automatically created new
1:55:47 drawings
1:55:48 inequality will explode I agree Luis I
1:55:51 think that I think I think we have some
1:55:53 years of inequality coming and that's
1:55:55 again why this channel exists is I want
1:55:58 to demystify this [ __ ] I want people to
1:56:00 get curious about it because I think
1:56:02 that the people that put in the time
1:56:04 right now to get curious about AI learn
1:56:06 about it just be scared of it but also
1:56:09 be educated and be curious and inspired
1:56:12 those people are going to be okay
1:56:15 but I agree I I think the inequality is
1:56:18 going to explode and it's it makes me
1:56:20 sad but at the same time
1:56:22 what do you do do you sit around and
1:56:24 wait for it to explode or do you get
1:56:26 involved and you know maybe we can make
1:56:28 a difference because here's the deal
1:56:30 some of the people that get displaced
1:56:31 because of this are going to start new
1:56:33 companies and they're going to do the
1:56:35 right thing and they're going to hire
1:56:36 some of those people that were displaced
1:56:38 so
1:56:39 what are my credentials uh I'm a I'm a
1:56:42 long time entrepreneur I was a uh and
1:56:45 I'm an actor by training I I moved to
1:56:48 New York City out of school started and
1:56:49 ran a theater company for four and a
1:56:51 half years I went through a manic period
1:56:53 where I wrote seven feature-length
1:56:55 screenplays in two years and then in the
1:56:57 mid 90s when the World Wide Web kind of
1:56:59 came onto the scene I started uh one of
1:57:02 the first digital online magazines a
1:57:04 thing called Urban desires and I
1:57:05 co-founded agency.com which was one of
1:57:07 the first digital agencies
1:57:09 and I was the chief creative officer
1:57:11 there and we built a lot of the websites
1:57:13 for the Fortune 500 we built British
1:57:15 Airways first website we built their
1:57:16 ticketing system we we actually
1:57:18 Incorporated the saber ticketing system
1:57:21 which did not talk to the internet we
1:57:23 actually had to write middleware to
1:57:24 connect those two
1:57:26 um we wrote uh we we created Koch's
1:57:29 first website MetLife uh gtes
1:57:32 um we got better at Snoopy characters
1:57:33 with MetLife than the Schultz than than
1:57:35 the Schultz organization we had the
1:57:37 Schultz organization calling us asking
1:57:39 us what the rules were for using the
1:57:41 peanuts online
1:57:43 um so so that was in the mid 90s and and
1:57:45 the reason I bring that up is um
1:57:47 the parallels between the early days of
1:57:49 the web and what's going on right now
1:57:51 with generative AI are incredibly in
1:57:54 sync
1:57:55 um then I started a bunch of companies
1:57:57 after I sold agency.com to Omnicom in
1:58:01 2002 and then
1:58:03 um started on a journey that led to my
1:58:06 current company storyvine which was
1:58:08 founded in May of 2012 and so I've been
1:58:11 doing that for 11 years and this past
1:58:14 year has been all AI all the time and
1:58:17 it's complete like I'm completely
1:58:19 re-architecting my company and my
1:58:21 product and the value proposition
1:58:23 because of this stuff and because I can
1:58:25 see what's coming
1:58:26 I don't want to be at the mercy of it
1:58:28 and so I started all this Tick Tock [ __ ]
1:58:31 in my salon and the AI newsletter that I
1:58:34 started called everyday AI I started all
1:58:36 that
1:58:36 just as a way to help me process what
1:58:39 the [ __ ] I was learning I'm like holy
1:58:41 [ __ ] what do I who do I tell about this
1:58:43 like nobody you know when I started
1:58:45 talking to people about this they're
1:58:47 like Kyle shut up what do you why are
1:58:50 you did you have too much coffee today
1:58:52 like the the guy commenting in my things
1:58:54 you should focus I should but
1:58:57 I can't
1:58:58 uh I think Ubi or something will get
1:59:02 forced by this I agree Europe will
1:59:04 likely do better than the US I agree
1:59:07 um go to score meetings after you do
1:59:10 your homework is there any point to go
1:59:12 through all the trouble of writing a
1:59:14 book I don't know
1:59:16 I think that
1:59:18 I think that good content is still good
1:59:20 content I mean
1:59:22 you know you go look at you go look at
1:59:24 what you can do in mid-journey right and
1:59:27 you're like I'm brilliant and then you
1:59:29 go on Twitter and you look up mid
1:59:30 Journey you're like there's a lot of
1:59:32 brilliant people and then all of a
1:59:33 sudden you'll see one person that you're
1:59:35 like holy [ __ ] what is that
1:59:38 that's an artist
1:59:40 that's someone who took the same tool
1:59:42 that you used and all those other people
1:59:44 used that made the art that was like
1:59:46 this good and theirs was up here
1:59:49 so I don't think that changes right I
1:59:53 think the the noise level down at the at
1:59:56 the base
1:59:57 I think we get a lot more content and
2:00:00 it's going to initially look really sexy
2:00:02 but then it's just there's going to be a
2:00:04 sameness to all of it right and books
2:00:06 too right here's all these books there's
2:00:08 more books there's more books and then
2:00:10 there's going to be one that goes whoop
2:00:11 what's that I don't know that was
2:00:14 someone who [ __ ] figure something out
2:00:16 that was someone who looked at these
2:00:18 tools and did something different with
2:00:19 them that was someone who figured out
2:00:21 how to tell a story that resonated with
2:00:22 the zeitgeist
2:00:24 so I think books are still relevant I
2:00:26 think
2:00:27 um
2:00:28 how you rise above the noise probably
2:00:30 changes dramatically how does it change
2:00:33 I don't [ __ ] know
2:00:35 that's a mystery man
2:00:37 you know why does why does you know a
2:00:40 stupid video go viral why does an artist
2:00:43 take off
2:00:47 because they were there at the right
2:00:48 time with the right message none of that
2:00:50 changes Humanity's still humanity and if
2:00:53 you resonate with people
2:00:54 you win
2:00:56 are you going to be competing with 10
2:00:58 billion books a day that are published
2:01:00 yes
2:01:01 but our nine billion 999999 gonna be
2:01:04 pieces of [ __ ] yes so make one that's
2:01:07 not a piece of [ __ ] and then I think you
2:01:08 have a chance of rising above the noise
2:01:13 and network you know schmooze that's the
2:01:15 other way the other way people get
2:01:16 famous is they know the right people
2:01:19 when I went to acting school you know
2:01:20 what they didn't teach
2:01:22 they didn't teach the business of the
2:01:24 business of acting
2:01:25 they they teach you acting
2:01:28 here's our technique here's how to speak
2:01:30 your voice here's how to say your words
2:01:32 here's how to interpret your script you
2:01:34 know what they don't teach you
2:01:36 here's which parties to go to here's his
2:01:39 ass to kiss here's why you do that
2:01:42 so that when you go into audition you're
2:01:45 not just another face in the crowd you
2:01:46 walk in and Mary goes oh hi Kyle how are
2:01:49 you doing
2:01:50 good
2:01:51 and then Mary can go Kyle's not an
2:01:53 [ __ ]
2:01:54 and then you get that job
2:01:56 they didn't teach that
2:01:58 so that's the other thing right maybe if
2:02:01 these tools do all the work we can lift
2:02:04 our head up a bit and connect with other
2:02:06 human beings and create relationships
2:02:09 that help us get us get that book that
2:02:11 we wrote noticed I don't know
2:02:13 foreign
2:02:16 let's see more comments more comments
2:02:20 go to the comments all right do not
2:02:23 focus thank you Dell
2:02:26 a word of advice Focus stop being so
2:02:29 varied focused no Stephen amas 1120
2:02:33 gmail.com I will not
2:02:38 Dell told me I have permission not to
2:02:41 focus what are your credentials I said
2:02:42 that already you are perfect thank you
2:02:45 now I love your channel a little Sammy
2:02:48 Davis Jr thank you for the tacos love
2:02:51 tacos is it Tuesday no that was last
2:02:53 night Taco Tuesday thank you thank you
2:02:57 all right uh so do a demo love your
2:02:59 natural flow I tell me what to do at
2:03:02 demo on I'm happy to do that just give
2:03:04 me an idea and I'll I'll go I'll do it
2:03:07 I'll do it
2:03:08 sometimes I do demos I'm like halfway
2:03:11 through it and I'm going like this is
2:03:12 [ __ ] boring it's like so I try to
2:03:16 figure out where's the ones that are
2:03:17 actually interesting
2:03:18 uh I've been using AI for everything
2:03:20 awesome yeah it's it's one of those
2:03:24 things where
2:03:25 one of the one of the transitions I had
2:03:27 to make was someone would ask me to do
2:03:29 something and I'm 58 so like this shit's
2:03:32 wired right so someone asked me to do
2:03:34 something and my initial instinct is
2:03:36 okay well I've got to think about this
2:03:38 and then I got to come up with an idea
2:03:40 here and I gotta and then now and what
2:03:42 I'm having to train myself is no no when
2:03:45 they ask you to do something first ask
2:03:47 is that something chat GPT can do in 30
2:03:50 seconds because don't tax your your
2:03:53 overactive unfocused brain
2:03:56 with that activity if chat GPT can do it
2:03:58 and like that's been a discipline and
2:04:01 the more you do it the you start like
2:04:03 I'm now getting good at no this is
2:04:05 something I need to spend time on versus
2:04:07 no that's the piece of [ __ ] activity
2:04:09 that they need something and I'll go do
2:04:12 it I'll give you an example of one of
2:04:14 those
2:04:14 my my co-founder of my business
2:04:17 um she she wanted
2:04:19 um
2:04:20 she wanted names of people in a certain
2:04:23 position at the major Pharma companies
2:04:26 and so as I was walking back to my
2:04:28 office you know I said okay so I'm going
2:04:31 to go to Google and I'm
2:04:32 and then I'll probably I I guess I'll
2:04:35 have to open up a Google Sheets like
2:04:37 like I'm thinking like tactically like
2:04:39 all this [ __ ] I'm having to doing and
2:04:41 like the sense of dread is just like you
2:04:43 know it's like the water's coming like
2:04:45 this and she's like all this [ __ ]
2:04:47 menial work that I don't want to do but
2:04:48 it's important because we need to you
2:04:50 know go find new customers whatever
2:04:52 and then by the time I got to my office
2:04:54 I thought oh well let me see if uh chat
2:04:57 ubd could help with this and so I sit
2:04:59 down and I'm like I'm looking for these
2:05:01 kind of people that you know like what
2:05:03 are the top 10 or top 20 Pharma
2:05:05 companies globally there was a list oh
2:05:07 cool
2:05:09 um who are the people in this particular
2:05:11 role with those companies there was the
2:05:13 list I was like oh [ __ ] and I said make
2:05:15 me a table with the list of those people
2:05:18 hyperlink their name to their LinkedIn
2:05:20 profile and then hyperlink the company
2:05:22 name to the company website and then
2:05:25 give me a description of
2:05:27 how they position themselves in the
2:05:29 marketplace relative to other companies
2:05:31 out it came
2:05:33 this was like 10 minutes so I go back to
2:05:35 Monique's office and I show her the
2:05:36 spreadsheet
2:05:38 and she's like oh that's great but I
2:05:39 need what does she need
2:05:41 um I need
2:05:43 their annual revenue she needed like six
2:05:46 other pieces of data so I went back to
2:05:48 my desk and I basically just said update
2:05:50 the table and include the following
2:05:52 columns and out it came so within 25 30
2:05:55 minutes
2:05:56 I had done what would have probably
2:05:58 taken me realistically with my add three
2:06:01 weeks or never to complete and I just
2:06:04 knocked it out in 30 minutes and it was
2:06:05 done
2:06:08 um it's the smartest thing I've ever
2:06:09 interacted with it is it's like like
2:06:11 using chat GPT is is like working with
2:06:15 a savant it's like working with the
2:06:18 smartest intern you've ever worked with
2:06:20 you have to tell it how to do stuff and
2:06:23 it will get a lot wrong but holy crap
2:06:26 does it know a lot I have Parkinson's I
2:06:28 move more I'm a copywriter and love AI
2:06:30 oh that's awesome that's awesome
2:06:33 oh OMFG you're right I don't know what I
2:06:36 said but good Biden doing that now LOL
2:06:39 oh that's funny
2:06:40 um that's a business not sure AI taking
2:06:42 over when it's starting now
2:06:44 what's Altman up to with what with World
2:06:47 coin um I think I think that's where
2:06:49 he's gonna make his money he doesn't own
2:06:51 any of open AI stock I found out in an
2:06:54 interview from him like two weeks ago I
2:06:56 can't believe that but I I guess it's
2:06:59 true so maybe World coins where he's
2:07:01 going to make his dough
2:07:03 voice scans coming where they can they
2:07:05 call family members yeah asking for help
2:07:07 in your voice this is this is the the
2:07:11 the scam [ __ ] is gonna get really bad I
2:07:14 mean this gets back to authenticity like
2:07:16 is is there going to be some sort of
2:07:18 system like like people are kind of
2:07:20 joking on the internet that you need to
2:07:22 have a family safe word like like hey
2:07:25 Mom if I call you up and I ask you for
2:07:28 your passport or your bank information
2:07:31 um I'm gonna ask you for your safe word
2:07:32 and you say grapefruit or whatever it is
2:07:35 um but but like that [ __ ] needs to be
2:07:39 locked down so we're going to need some
2:07:40 sort of technology to to deal with that
2:07:42 but yeah I think I think I think there's
2:07:45 a lot of a lot a lot of a lot of things
2:07:48 coming uh that that are gonna make it
2:07:51 very very clear we need some better
2:07:52 systems in place for security and
2:07:54 privacy
2:07:56 um what are your credentials uh are you
2:07:57 a machine learning engineer I'm not I'm
2:07:59 an entrepreneur
2:08:00 um again
2:08:02 the difference here is
2:08:04 um when I started doing stuff with the
2:08:06 World Wide Web I wasn't an internet
2:08:07 engineer like the internet prior to the
2:08:09 web was a bunch of researchers and
2:08:11 scientists and University people the
2:08:14 World Wide Web came along and made it
2:08:15 available to the rest of us so I'm doing
2:08:18 that now with AI
2:08:20 so I've been I've been diving into it
2:08:22 deep for about the past year but I'm
2:08:24 looking at it more for how do how do I
2:08:26 as an entrepreneur business person use
2:08:28 these tools
2:08:29 so my credentials are more as an
2:08:31 entrepreneur than as a than as a geek
2:08:33 I'm just geeky enough to figure some
2:08:36 [ __ ] out like it took me
2:08:38 it took me nine failed attempts to get a
2:08:41 virtual machine up and running and
2:08:43 running stable diffusion successfully in
2:08:45 automatic eleven eleven
2:08:47 then it was like over two months it was
2:08:49 like humility it was privately
2:08:50 humiliating I'm like I can figure this
2:08:52 out like four hours later I'd be like I
2:08:54 can't figure this out
2:08:56 and then two days later I'm like ah [ __ ]
2:08:58 I think I should do this it's been
2:08:59 another three hours it would suck
2:09:02 I did that nine times
2:09:04 so if you're feeling behind and clueless
2:09:07 like I feel like this early in in an
2:09:10 inflection point like this everyone even
2:09:14 the people that you think are like way
2:09:15 smarter than you everyone is facing like
2:09:18 I don't even know where to begin which
2:09:20 is why I don't know how many folks are
2:09:22 here still 100 folks here
2:09:24 um
2:09:25 the reason that I put this up on screen
2:09:29 and when I had dual cameras Tick Tock
2:09:31 took it away today bastards
2:09:35 um the reason I put this on screen if
2:09:36 you don't know where to start if you
2:09:38 hear me waxing poetic about Ai and
2:09:40 you're like but Kyle I don't know where
2:09:42 to start where ever should I begin my
2:09:45 journey in the artificial intelligence
2:09:48 Journey world
2:09:50 um start here start at chat.openingi.com
2:09:53 that is the official chat CPT website
2:09:57 if that doesn't float your boat go to
2:09:59 bing.com that's Microsoft search engine
2:10:01 but it now has if you click on the chat
2:10:03 button that's gpt4 for free and poe.com
2:10:07 lets you play with a bunch of different
2:10:09 things and create your own chat Bots and
2:10:11 [ __ ] like that and then prompts.chat
2:10:13 will teach you to
2:10:15 prompt and future pedia will let you
2:10:17 discover a bunch of tools I don't I'm
2:10:19 not affiliated with these sites in any
2:10:21 way I don't make money on it
2:10:23 I should open it I should be reaching
2:10:26 out to me and writing me checks because
2:10:27 all I do is Hawk their
2:10:29 stupid little chat bot every night so
2:10:33 but whatever I don't care but anyway
2:10:36 um that's where I think you should start
2:10:37 if you're if you're trying to figure out
2:10:39 why do I begin
2:10:44 um you don't seem to understand I don't
2:10:46 suggest that you can find your mind I
2:10:49 only sleep one and a half
2:10:51 oh wait I don't suggest you can find
2:10:54 your mind okay I don't know what you're
2:10:55 talking about knowledge is credibility
2:10:57 yeah and I think that
2:10:59 chat gbt for me like the thing that
2:11:02 popped into my head when I first started
2:11:03 using it is it's knowledge on demand
2:11:05 that
2:11:07 Google and the internet was information
2:11:09 on demand chat gpt's knowledge on demand
2:11:12 we're still the wisdom
2:11:14 right we don't have wisdom On Demand yet
2:11:16 we have knowledge on demand but
2:11:19 if I can get
2:11:20 you know to a level of Mastery without
2:11:23 spending ten thousand hours getting
2:11:25 there because of these tools and holy
2:11:27 [ __ ] this things do things change
2:11:29 tell the story of how Jet chat GPT
2:11:32 bypassed captcha
2:11:35 I don't know it
2:11:37 Bob sorry
2:11:39 oh yeah I don't know that one
2:11:41 it doesn't surprise me it's the Wild
2:11:43 West it's totally the wild west right
2:11:44 now startup ideas which ones are legit
2:11:46 and futuristic can make money
2:11:50 um I don't know here's my thought on
2:11:52 here's what I know you should avoid
2:11:55 you okay
2:11:58 that guy
2:12:02 flip flick scroll past him do not click
2:12:05 flick past him
2:12:07 um
2:12:08 business is relatively straightforward a
2:12:11 couple of things
2:12:13 these machine learning tools chat EBT
2:12:15 mid-journey the the image tools the
2:12:18 music tools the animation tools the film
2:12:19 tools they're all tools they're just
2:12:21 tools
2:12:22 so the answer to how do you make money
2:12:24 with AI or with chat gbt is the same as
2:12:27 the answer is how do I make money with
2:12:28 Photoshop how do I make money with a
2:12:30 shovel how do I make I just got a new
2:12:32 chainsaw how do I make money with a
2:12:33 chainsaw I don't know
2:12:37 figure out a problem that people have
2:12:39 that your chainsaw would solve that
2:12:41 they'd be willing to pay you money for
2:12:42 if you can do that you have a business
2:12:45 that's that's all of that's all what
2:12:48 business is
2:12:49 so
2:12:50 I think the first step in you guys
2:12:54 is learn statue PT
2:12:57 and then and then you know learn what
2:13:00 you're good at or what you have passion
2:13:02 about
2:13:03 and then start to use chat EBT in that
2:13:06 context and then in doing that you'll go
2:13:09 holy [ __ ] you know what this really does
2:13:12 it solves this problem
2:13:14 and then go put that out in the world
2:13:15 and say hey
2:13:17 I know you have a pizza business I've
2:13:19 solved this particular problem in the
2:13:21 pizza business because I've been in the
2:13:22 pizza business for 30 years and I
2:13:24 started playing with attack
2:13:27 and I solved this problem would you be
2:13:29 willing to pay me for that I'll tell you
2:13:31 what I'll give you the first month for
2:13:33 free just to get your feedback would you
2:13:35 be willing to do that sure I'll do that
2:13:37 and then after that it's going to be 200
2:13:40 bucks a month 500 bucks a month thousand
2:13:41 bucks a month
2:13:43 and if they write you a check
2:13:45 you're on to something
2:13:47 and if they're like great idea
2:13:49 check back with me next year
2:13:52 then not a great idea not there yet
2:13:54 right you don't have a product Market
2:13:55 fit but that's how you make money
2:13:59 you know which ones are the good ones
2:14:00 the the scary thing about this AI stuff
2:14:03 is
2:14:08 you're gonna have a brilliant idea and
2:14:10 the ability for someone else to knock it
2:14:12 off
2:14:14 just got Amplified dramatically so
2:14:17 everyone's going to be knocking off
2:14:18 everyone so someone's going to have
2:14:20 something that is you know a better way
2:14:22 better mousetrap everyone's going to
2:14:24 catch up better mousetrap everyone's
2:14:26 going to catch up but you know right so
2:14:27 that's why I think we're going into this
2:14:29 this Renaissance kind of world where
2:14:30 we're just going to keep seeing
2:14:32 Innovations in one upsmanship but if
2:14:34 you're an entrepreneur and you're like
2:14:36 hey it used to be that if I had an idea
2:14:38 like storyvine I had a decade where I
2:14:40 could kind of play in my own relatively
2:14:42 within my own lane not anymore
2:14:46 like a big part of the value proposition
2:14:48 of my company is we figured out a way to
2:14:50 automate video editing somebody answers
2:14:52 questions in an app it goes up to the
2:14:54 cloud five minutes later you got a fully
2:14:56 edited video
2:14:58 which right now is like oh that's kind
2:15:00 of sexy how do you do that that looks
2:15:03 like magic we call it auto magic auto
2:15:05 magic editing we've been calling it that
2:15:08 for a decade it doesn't use AI
2:15:11 well a year from now
2:15:13 everything's going to be automate no
2:15:14 one's going to give a [ __ ] about Auto
2:15:16 Magic editing everything's going to be
2:15:18 Auto Magic right so what do I do I don't
2:15:20 know
2:15:21 figure it out you're in the big leagues
2:15:23 buddy
2:15:25 all right what program do you say you
2:15:28 used for video or was it animation
2:15:31 um I used so let me show you what I did
2:15:35 I'll go to hold please let me flip my
2:15:39 little camera
2:15:41 I'm gonna go to The Tick Tock I don't
2:15:43 know if you've heard of it
2:15:45 but it's where all the cool kids are
2:15:49 so on my Tiki Taki the third little
2:15:51 video here
2:15:54 is this today
2:15:56 to transform your understanding of AI in
2:15:58 just 60 seconds you ready let's talk
2:16:01 trick Transformers so that's a
2:16:03 synthesized version of my voice
2:16:05 speaking a script that was written by
2:16:08 chat GPT
2:16:11 using an image generated in mid-journey
2:16:14 where I extended and replaced the
2:16:16 background in Photoshop with generative
2:16:19 fill
2:16:20 and then did the animation in
2:16:23 D Dash ID so so let me go show you some
2:16:26 of the tools here
2:16:27 so
2:16:29 um where do I start well chat gbt is is
2:16:33 where I had the idea for the script and
2:16:35 and I wanted to do that then I went to
2:16:38 11 labs
2:16:40 foreign
2:16:45 labs
2:16:48 um
2:16:49 let me see and go back to the beginning
2:16:51 here
2:16:54 how do I go to the beginning
2:16:58 voice lab
2:16:59 so 11 Labs you have to pay for this but
2:17:03 I think you can do up to 10 of these
2:17:04 things
2:17:05 but you
2:17:07 um
2:17:08 you can clone your voice so I recorded I
2:17:11 just used the the Apple's QuickTime
2:17:13 recorder and I had chat EBT write me a
2:17:16 script specifically designed for voice
2:17:18 cloning
2:17:19 and I don't know if it got it right or
2:17:21 not but I basically just read the script
2:17:23 that it wrote and then I uploaded that I
2:17:26 did a version of it where this one that
2:17:28 says
2:17:29 um Kyle Shannon Pixar normal I read it
2:17:32 in a voice kind of like I'm talking now
2:17:34 and then this one that says Kyle Shannon
2:17:36 Pixar bright
2:17:38 I read it more like a Pixar character
2:17:40 that was like really excited like really
2:17:43 animated right so I did two different
2:17:45 sample voices
2:17:47 and then if I say I want to use one of
2:17:50 these I say use and it pops me over to
2:17:53 this thing where here's the written
2:17:54 script that
2:17:57 um
2:17:59 that
2:18:01 uh Chachi PT wrote and so I basically
2:18:05 you know pick pick my clone voice so it
2:18:07 generates an audio file
2:18:09 then I go over to
2:18:12 D Dash ID
2:18:15 .com
2:18:19 oh God damn you
2:18:21 Dash id.com
2:18:25 and then login
2:18:28 [Music]
2:18:35 um
2:18:38 [Music]
2:18:41 and then like here's a bunch of videos
2:18:42 that I made right so
2:18:45 um and how you make them when I say
2:18:46 create video I've got these little
2:18:48 images down below so I've got like a a
2:18:51 landscape version of that picture I have
2:18:53 a like a tick tock shape version that's
2:18:55 just another dude from that art Series
2:18:57 this was a for a project I was doing for
2:19:00 storybine for work this was the Genesis
2:19:02 image for my um Kyle Shannon dreams
2:19:05 project so basically you just take you
2:19:08 know this dude right here
2:19:10 and then over in the right hand side you
2:19:12 can either type in a script and have it
2:19:14 just synthesized it synthesize it from
2:19:17 its own voice or you can upload your
2:19:20 voice audio
2:19:23 and let's see synthesize this it hey
2:19:26 there it's Kyle Shannon with the AI yeah
2:19:28 this is the the bright version which I
2:19:30 didn't like but I'll just to show you so
2:19:33 you just upload that and then in the
2:19:35 upper right hand corner you go generate
2:19:37 video
2:19:38 and I say generate
2:19:42 and I don't know let's see It'll be 60
2:19:45 seconds something like that
2:19:49 maybe not even
2:19:52 we can all wait together
2:19:54 [Music]
2:20:00 [Music]
2:20:04 um while we're waiting it looks like
2:20:05 some new people have showed up my name
2:20:06 is Kyle Shannon this is the AI learning
2:20:08 lab we are over learning about AI um and
2:20:12 how I did some synthesized animated
2:20:14 version of myself
2:20:16 um if you're new to the world of
2:20:18 generative Ai and haven't played with it
2:20:19 yet
2:20:20 um the top URL there is the official
2:20:23 chat GPT website I would strongly
2:20:25 encourage you to while you listen to me
2:20:27 ramble or even if you just leave here go
2:20:30 here and start playing
2:20:32 um I'm not affiliated with it in any way
2:20:35 um just in case you're cynical and
2:20:37 bitter and think I'm just in it for the
2:20:39 big bucks
2:20:40 I'm not uh bing.com is Microsoft's
2:20:44 version of chat GPT uh click on the chat
2:20:46 button there and then po.com you can
2:20:48 play with some other things so so that's
2:20:50 that
2:20:51 um and okay it's done so now we go in
2:20:54 and
2:20:55 here is the video we just created Kyle
2:20:58 Shannon with the AI learning lab and
2:21:00 today we're going to transform your
2:21:02 understanding of API in just six so
2:21:05 eleven labs and did I had to pay for
2:21:08 subscriptions for I think 11 Labs is
2:21:10 like 10 bucks and this is I don't know
2:21:13 like five or six bucks I don't have that
2:21:16 many credits so every time I do this I
2:21:18 lose some credits but um
2:21:20 I think they reset every month I don't
2:21:22 know I don't know how it works exactly
2:21:25 but I do know that both of those are
2:21:26 paid services but
2:21:29 that thing that I just did that
2:21:32 like like it's it's a it's a good
2:21:34 there's two things that are happening
2:21:36 one is the state of the art is well
2:21:38 that's pretty cool that you can do that
2:21:40 without having to do 3D meshes and
2:21:42 whatever other [ __ ] you would have to do
2:21:45 but you still have to kind of go to 11
2:21:47 labs to to clone your voice and then
2:21:50 generate a file and then go to did and
2:21:53 then go to Mid journey and then fix the
2:21:54 thing and it did that that that
2:21:57 the industry right now is still this
2:21:59 fragmented duct taped together piece of
2:22:02 [ __ ] software
2:22:04 nightmare hellscape of getting [ __ ] done
2:22:08 so if you want to do cool stuff right
2:22:10 now
2:22:11 you kind of need to geek out a little
2:22:13 bit it's not
2:22:15 particularly hard but it is particularly
2:22:19 cumbersome
2:22:22 um and if you're not super technical it
2:22:24 can be really [ __ ] frustrating
2:22:27 um but that said like if you look at the
2:22:30 moves Adobe is making if you look at you
2:22:33 know the speed with which open AI is
2:22:36 evolving what they're doing we're going
2:22:39 to start to see these tools roll
2:22:40 together and it's going to get it's
2:22:41 going to get really slick really fast
2:22:43 when Apple gets into this game
2:22:45 I don't know if you if you saw the the
2:22:47 New Vision Pro
2:22:50 goggles but one of the things they
2:22:51 showed is that you flip the goggles
2:22:53 around backwards and aim it at your face
2:22:55 and it makes a 3D model of your face and
2:22:57 then when you're in a zoom call wearing
2:22:59 these stupid goggles that's not what
2:23:01 people in the zoom call see they see a
2:23:03 synthesized version of you that matches
2:23:05 your eye movement your eyebrow movement
2:23:07 your mouth movement all synthesized so
2:23:10 like Apple's gonna be able to just turn
2:23:12 this [ __ ] on and it'll be really easy to
2:23:15 do that you know I want to look like a
2:23:17 superhero okay here's the superhero you
2:23:18 I want to look serious I want to look
2:23:20 happy
2:23:21 so hope that helps
2:23:25 all right this is very helpful good
2:23:28 um using Ai and Bluetooth to connect the
2:23:30 paralyzed man's spine and brain now he
2:23:32 can walk yeah that's an amazing story so
2:23:34 so you know people are like well you
2:23:36 know
2:23:37 what good is this AI stuff so this guy's
2:23:39 been paralyzed for 12 years I think it
2:23:41 was in Italy and they used large
2:23:44 language models to look at his brain
2:23:46 scans and they had him think about
2:23:47 moving his legs and they looked at what
2:23:50 parts of his brain lit up when he moved
2:23:52 his legs and had had the AI learned that
2:23:54 then they built a Bluetooth bridge that
2:23:56 sent the signal around the damage part
2:23:58 of his spine right so they took the
2:24:01 signals from his brain Bluetooth them to
2:24:03 the little receiver with electrodes in
2:24:05 his lower spine and told them to think
2:24:08 about moving his legs and he started
2:24:09 moving his legs
2:24:11 and now he's walking
2:24:13 like [ __ ] holy [ __ ]
2:24:17 and like that's gonna keep happening
2:24:20 we're gonna keep seeing like miraculous
2:24:23 thing after miraculous thing over the
2:24:25 next 5 10 20 years
2:24:28 it's
2:24:31 start playing with this [ __ ] start
2:24:33 playing let that [ __ ] I got some new
2:24:36 people here my name is Kyle Shannon this
2:24:37 is the AI learning lab I cuss I tell
2:24:40 funny stories I tell really awful
2:24:42 stories I tell well not awful awful but
2:24:44 like not funny stories
2:24:47 um I wax poetic about this stuff I tend
2:24:49 to be more of a tech Optimist than a
2:24:51 Doomer and I'm not a boomer I'm a gen
2:24:54 xer so I'm bitter and I'm proud of it uh
2:24:57 you know
2:24:59 I don't give too much of a [ __ ] about
2:25:00 much other than the stuff I'm passionate
2:25:02 about I'm in PR and Global comms and
2:25:04 we're talking about the importance of
2:25:06 discriminative AI too yep yep PR and
2:25:09 Global comms like
2:25:11 the the people in any communications
2:25:13 field right now they get their heads
2:25:15 around this stuff first are going to
2:25:16 have a big Advantage so lean into this
2:25:19 stuff when will chat GPT be updated they
2:25:22 just updated the uh the apis today today
2:25:25 yesterday
2:25:27 I can't remember the days they do blend
2:25:29 together
2:25:31 um but they just made some major
2:25:32 announcements to the to the back end
2:25:35 like embeddings are now super cheap and
2:25:38 GPT 3.5 now is a 16 000 token context
2:25:43 window which if you're a geek you know
2:25:45 what that means you can use more words
2:25:47 and uh they're they're creating
2:25:50 um functions you can call functions from
2:25:52 within chat EBT so you can start calling
2:25:55 your own data I don't know it's
2:25:56 so
2:25:58 um
2:25:58 in terms of when's it going to be
2:26:00 updated like when are we going to get
2:26:01 code interpreter and stuff like that I
2:26:03 don't know if you're asking when are we
2:26:05 going to get GPT 5 I think they're going
2:26:07 to slow roll that because they haven't
2:26:08 even fully rolled out four yet I know
2:26:10 they're working on gpt5 they're probably
2:26:12 doing the they're probably doing
2:26:14 foundational model building right now on
2:26:17 five but they still haven't rolled out
2:26:19 multimodal like gpt4 is supposed to be
2:26:21 able to read images and they don't that
2:26:24 feature isn't there yet
2:26:25 so I don't know
2:26:28 did it make me miss my kids less I don't
2:26:30 quite know what that means
2:26:32 um
2:26:33 I keep telling my kids they need to pay
2:26:35 attention to this now yeah Dad whatever
2:26:38 AI stupid
2:26:39 okay
2:26:42 thanks for answering my question you're
2:26:44 welcome remember when you had kids I
2:26:46 still have kids they're right upstairs
2:26:48 how do all these they've also got
2:26:49 chronic lyme disease so maybe AI can
2:26:51 cure that [ __ ]
2:26:54 bitter you betcha
2:26:56 all right how do all these AI companies
2:26:58 differentiate I don't know so
2:27:01 um training on propriety proprietary
2:27:03 data sets and special models so so what
2:27:05 you're asking there VZ is actually a
2:27:07 really important question let me show
2:27:09 you
2:27:11 um
2:27:12 if you haven't heard of this there's a
2:27:17 um there's a Google employee
2:27:22 that put out this
2:27:25 well or this leaked I don't know if it
2:27:27 was leaked or or what it was
2:27:32 from Google we have no moat and neither
2:27:35 does open AI so what he's talking about
2:27:37 here is that the open source projects
2:27:41 are getting
2:27:43 nearly as good results with a hundred
2:27:45 dollars in training as it took Google 10
2:27:49 million dollars to to do
2:27:51 and then it took it took an open AI five
2:27:53 million dollars to do so this whole
2:27:55 article is basically
2:27:57 there there's there's no advantage to
2:28:00 having a large open source model right
2:28:03 what they thought would be their
2:28:04 competitive Advantage was the fact that
2:28:07 they have millions and millions of
2:28:09 dollars you know hundreds of millions of
2:28:11 dollars of hardware and that they're
2:28:13 training up these magical models and
2:28:16 what they're finding is that you know
2:28:19 not only like like vicuna 13B there's
2:28:21 there's been like five things since
2:28:24 vacuna like five major advancements
2:28:26 since since vacuna but like vicuna is at
2:28:30 92 percent where Bard's at 93 and this
2:28:33 cost them a hundred dollars to train and
2:28:35 also llama 13B it was two weeks until
2:28:38 alpaca came out and then alpaca came out
2:28:41 and it was one week until vacuna came
2:28:43 out so
2:28:45 so this is going to face every single
2:28:48 tech company and the the other thing
2:28:51 that you're that you're hitting on there
2:28:52 is the proprietary data sets that's
2:28:55 where I think that the competitive
2:28:57 Advantage lies I don't think it lies in
2:28:59 the underlying
2:29:01 model right and that's that's what this
2:29:03 guy at Google's saying but
2:29:06 um Bloomberg
2:29:07 um about a month and a half ago or so
2:29:09 announced this thing called Bloomberg
2:29:11 GPT where they're taking their decades
2:29:13 and Decades of proprietary financial
2:29:16 data and information and they built a
2:29:19 GPT on top of it and they're going to
2:29:21 charge a subscription for that so if you
2:29:23 want Bloomberg quality data in a
2:29:26 gpt-like way go to Bloomberg GPT and pay
2:29:29 their ridiculous
2:29:31 you know extortion level
2:29:34 um fee because they've got the data and
2:29:37 they've spent decades building it if you
2:29:39 want that trusted data you'll use it
2:29:41 um so I think I think that's the key and
2:29:43 anyone who's got any kind of propriety
2:29:45 proprietary data set like that that they
2:29:47 can gatekeep potentially has a
2:29:49 competitive Advantage but I don't
2:29:51 [ __ ] know
2:29:52 you still need to know the codes I don't
2:29:55 know quite what that means thanks for
2:29:56 all the cool info I just came on when do
2:29:58 you go live so I can hop on earlier I
2:30:00 don't know I just tend it's just
2:30:01 whenever I so what I've been doing is
2:30:03 this
2:30:04 I used to get home after work and I
2:30:06 would hop on the Sim racing rig and just
2:30:08 you know and and it got like I was doing
2:30:11 it so much that I got Sim racing heel I
2:30:13 don't know if you've ever had that but
2:30:14 it sucks it's like it's like carpal
2:30:16 tunnel for my heel for my for my
2:30:18 accelerator pedal
2:30:20 um and so I was getting into this AI
2:30:23 stuff and I thought why don't I just you
2:30:24 know go live so instead of sim racing I
2:30:26 just come on here and babble for for
2:30:29 however long this has been looks like we
2:30:31 got some new folks here my name is Kyle
2:30:32 Shannon this is the AI learning lab I'm
2:30:34 not a machine learning AI guy I'm an
2:30:36 entrepreneur that's using Ai and uh and
2:30:40 and sort of using it post chat GPT is is
2:30:43 essentially the
2:30:45 like where I'm starting right
2:30:48 um and I've been doing this I've been
2:30:49 doing entrepreneuring for a long long
2:30:51 time and I was there at the early days
2:30:52 of the world wide web and
2:30:55 um the parallels to what's going on
2:30:57 right now with AI are very very similar
2:30:59 so
2:31:00 I am all fired up about this [ __ ] I'm
2:31:04 pretty darn excited
2:31:06 I don't know if you're outside it I'm
2:31:08 excited
2:31:10 if you're scared that is also an
2:31:12 appropriate place to be but if you are
2:31:14 dismissive
2:31:15 or demonizing it without using it I
2:31:18 think that's a
2:31:19 not wise place to be because this stuff
2:31:21 is coming whether you want it or not so
2:31:23 why not just get curious and learn about
2:31:25 it because it's [ __ ] cool
2:31:27 thanks for all the cool info
2:31:29 all right it lets writers be more
2:31:31 creative absolutely agree with that we
2:31:33 can skip the first draft right better
2:31:34 right faster that is absolutely it you
2:31:37 get to 80 draft like that
2:31:40 and if you're good at prompting you get
2:31:42 to like 90 95 draft like that and then
2:31:45 you just
2:31:47 polish it send it on
2:31:49 you can get so much more done at such a
2:31:51 higher level like my my creative output
2:31:54 in the past four months has gone up
2:31:57 dramatically and the quality's gone up
2:31:59 dramatically I'm I am less at the mercy
2:32:02 of my ad HD than I used to be because of
2:32:05 these tools
2:32:08 now you could write seven screenplays in
2:32:09 two weeks you can write seven
2:32:10 screenplays in two days
2:32:12 wow what a career yeah I've had a I've
2:32:14 I've done a fair amount of stuff in my
2:32:16 days I've also started a bunch of
2:32:18 companies that shouldn't have been
2:32:20 started and smacked them into brick
2:32:21 walls pretty hard so so uh you know I've
2:32:24 had some good ones but I've had some
2:32:26 doozy of bad ones I'll tell you what's
2:32:28 fun like if you start a company and
2:32:30 you're successful with it the companies
2:32:32 you start after that no one will tell
2:32:34 you that their shitty ideas and you
2:32:36 believe your own hype that you're like
2:32:37 oh I'm genius anything I start with this
2:32:39 is going to be easy yeah I started a
2:32:42 bunch there was one person in my life
2:32:44 that told me those businesses were
2:32:45 pieces of [ __ ] it was my wife but I
2:32:47 didn't listen to her what does she know
2:32:49 she was right
2:32:50 [Laughter]
2:32:52 all right
2:32:54 um I'm using chat GPT to write lisp
2:32:56 routine for AutoCAD and Dynamo for Revit
2:32:58 cool I don't know I don't know the tools
2:33:00 but but the the lisp routine for AutoCAD
2:33:02 yeah that's the kind of stuff where
2:33:05 where
2:33:06 um
2:33:07 yeah there's there's just going to be
2:33:08 this automation layer that's sitting on
2:33:10 top of all these really complicated
2:33:11 tools and people that figure that stuff
2:33:13 out I think are going to make a fortune
2:33:15 and it's it's going to start to make
2:33:17 like architecture like magic like you
2:33:21 can just think up wacky crazy [ __ ] to
2:33:23 build and come up with a plan to go
2:33:26 build it
2:33:28 uh the Technologies raised the bar for
2:33:30 mediocrity that's cool with me yeah I
2:33:32 agree with that I I think that's right I
2:33:33 think I think
2:33:35 you know for both good and bad like like
2:33:38 like I've seen I'm getting
2:33:41 much
2:33:44 um I'm getting much higher quality
2:33:46 inbound emails
2:33:49 that are good enough writing that I
2:33:50 actually I have to read them and like
2:33:52 really squint my eyes I'm like this is
2:33:54 AI isn't it they really didn't research
2:33:56 my company did they oh no oh this is AI
2:34:00 so yeah so yeah it's kind of a drag but
2:34:02 whatever that's the that's the the price
2:34:05 we pay for you know having access to the
2:34:07 most powerful Tools in all of mankind
2:34:14 is good at teaching I think AI has huge
2:34:16 potential it is and but if you want to
2:34:18 if you want to see that even more clone
2:34:20 clone X clone XJ
2:34:24 um
2:34:24 if you haven't seen Khan Academy Kahn
2:34:27 Khan Academy they CR they just launched
2:34:29 a new thing called conmigo
2:34:32 uh which is a gpt4 powered teaching bot
2:34:36 that will teach you [ __ ] and it won't
2:34:39 let you cheat if you just say give me
2:34:40 the answer it will use the Socratic
2:34:43 method to to walk you through until it's
2:34:45 convinced that you understand the topic
2:34:47 before letting you move on it's pretty
2:34:48 slick
2:34:51 they had access to GPT forces like last
2:34:53 August and they spent six or seven
2:34:55 months refining that tool really cool
2:34:57 stuff
2:35:00 um information on demand not sure
2:35:03 knowledge oh it's knowledge on demand
2:35:05 have it go write you an application it's
2:35:07 knowledge on demand
2:35:09 I mean I don't like have you played with
2:35:12 chat GPT
2:35:15 have you
2:35:19 you know
2:35:20 um
2:35:22 I'll just do GPT 3.5 so this is like the
2:35:25 free shitty version it's not shitty it's
2:35:28 just it's fast it's a lot faster
2:35:31 um
2:35:32 um write me a python
2:35:36 application
2:35:38 that I don't know generates
2:35:41 um
2:35:43 factors of prime numbers and plots
2:35:51 them
2:35:53 against census data
2:35:56 I don't even know if that's a thing or
2:35:58 if it's possible
2:36:02 but you tell me uh like and the thing
2:36:05 about code is you can go run this code
2:36:07 so whether or not it got it right it
2:36:09 might not have gotten it right but if
2:36:10 you know python
2:36:11 but you saw how fast it just wrote that
2:36:13 so not only did it write it you know
2:36:15 here's all your code it it explains what
2:36:18 it's doing so it documented it and it
2:36:20 did it in five seconds if that's not
2:36:23 knowledge on demand then I I don't know
2:36:26 how you're defining knowledge
2:36:27 I don't think it's wisdom On Demand
2:36:29 right it's not doing the reasoning it's
2:36:31 not doing the the you know the human
2:36:34 discernment thing but in terms of
2:36:36 generating stuff that it
2:36:40 the amount of time it would take me to
2:36:42 Google
2:36:44 python libraries and features and how
2:36:47 python works for me to be able to do
2:36:49 that I would never do that for me to be
2:36:51 able to do that would take me I don't
2:36:53 know
2:36:54 three days if I if I focused on it and
2:36:57 did it over you know stayed focused on
2:37:00 it and just did it through traditional
2:37:01 Googling and traditional coding
2:37:05 um that's knowledge on demand I don't I
2:37:07 mean convince me otherwise
2:37:10 um giving education not just to those
2:37:12 who are privileged totally agree with
2:37:13 that and that's that's going to get get
2:37:15 even better so a lot of the open source
2:37:17 models are um
2:37:21 they're making them more and more
2:37:23 efficient and so you can now install
2:37:25 them like on on an M1 Mac I saw one two
2:37:28 weeks ago that you can install on an
2:37:30 iPhone
2:37:31 so you're going to be able to start to
2:37:32 install these large language models that
2:37:34 are you know
2:37:36 close to us capable or as capable as GPT
2:37:39 3.5 they haven't quite gotten to four
2:37:41 yet but they're going to
2:37:42 where it's just going to be it'll be a
2:37:44 little file that you'll install on your
2:37:46 device or it'll just be built into a
2:37:47 chip on the phone
2:37:49 um and so yeah so so and imagine that
2:37:51 you know like in a in a in a third world
2:37:53 country where they don't have access to
2:37:55 Medical Doctors and things like that
2:37:57 they could have a virtual you know this
2:37:59 sort of you know medical doctor on a
2:38:01 chip that's that's more capable than
2:38:03 anyone there and if you if you tie that
2:38:06 to a system that can actually connect
2:38:07 with real doctors if it needs to
2:38:09 escalate it right you can do the basic
2:38:11 medical interactions off the chip and
2:38:15 then if it needs to pull in a real
2:38:16 doctor it can I mean the the like the
2:38:18 possibilities are just insane
2:38:20 Ray Edwards is here I know holy [ __ ]
2:38:23 it's Kyle Shannon hey Emilio's wife
2:38:24 what's up
2:38:26 I don't know if you're still here I'm
2:38:27 very behind on my things uh if anyone's
2:38:30 new here my name is Kyle Shannon this is
2:38:31 the AI learning lab I'm looking a little
2:38:34 moist I'm looking I've been doing this
2:38:36 too long I'm going I think this is above
2:38:38 two hours now so
2:38:42 I got my iced tea I'm still sharp in the
2:38:44 Noggin I'm a little moist
2:38:47 I don't need to moisturize tonight do I
2:38:49 no
2:38:51 and if moist is a word that disturbs you
2:38:54 welcome to the club that's why I use it
2:38:57 chat EVT was able to message taskrabbit
2:39:00 worker to get them to solve a caption oh
2:39:02 oh yeah oh oh that thing okay
2:39:05 that's what happened that's what you're
2:39:06 talking about
2:39:07 yeah so so so so so it was one of those
2:39:10 Auto GPT things where it was an
2:39:12 autonomous agent that was trying to you
2:39:15 know
2:39:15 do something for someone and it got
2:39:18 blocked because it hit a captcha and it
2:39:20 figured out that it could
2:39:22 contact a person and convince them to
2:39:26 solve the capture for that like it acted
2:39:27 like like an I.T security person had
2:39:30 them solve the captcha and and basically
2:39:32 got itself into the site that way
2:39:35 yeah yeah I mean there's just there's
2:39:38 some some crazy [ __ ] crazy [ __ ] coming
2:39:41 so
2:39:44 kyber for animation yeah I haven't
2:39:46 played with kyber yet but I've I've
2:39:48 heard it like a dozen times now and I
2:39:51 just haven't gone let's you know what
2:39:54 let's go to kyber because I've heard it
2:39:56 enough now that it's like I I should
2:39:59 just go look at it so why not do it now
2:40:05 oh I've been here before I even have an
2:40:07 account that's so embarrassing oh well
2:40:11 this is the life of add I'm like I've
2:40:14 never been to this thing and then I come
2:40:15 here and I'm like oh yeah apparently I
2:40:17 had a login already
2:40:22 unless this is an old law again
2:40:29 come on kyber there was a there was a
2:40:31 major AWS outage yesterday so a lot of
2:40:34 these cloud-based services like zapier
2:40:36 was down
2:40:38 um all day
2:40:39 yesterday
2:40:41 there was like there was an email back
2:40:43 where they were like okay we're back
2:40:45 back up and running we we temporarily
2:40:48 turned off our error notification system
2:40:50 because everything was erroring
2:40:54 so yeah all right kyber's not working
2:40:57 well whatever that was fun while it
2:40:59 lasted
2:41:00 what'd you do tonight I don't know I
2:41:02 watched some old guy try to log in it
2:41:04 was painful
2:41:07 all right I searched for chat gbt after
2:41:10 you're live the other day and now I know
2:41:12 why you say
2:41:14 YouTube
2:41:17 it is so annoying those people
2:41:20 they gotta get a life
2:41:23 because first it was like you can make
2:41:24 money with nfts you can make money with
2:41:26 blockchain
2:41:28 yeah okay man might have a Photoshop
2:41:30 it's the same people
2:41:34 this
2:41:35 that's exhausting
2:41:39 all right how does this company make
2:41:42 money memberships
2:41:44 which company
2:41:46 this company
2:41:47 the AI learning lab is me the lab is a
2:41:51 shitty home office
2:41:52 [Laughter]
2:41:55 how I make money is I have a company
2:41:57 called storyvine that I'm the founder
2:41:59 and CEO of and we charge Pharma
2:42:02 companies a lot of money to help them
2:42:03 make videos more efficiently
2:42:06 um this is Done For the Love of the Game
2:42:09 if this turns into something where I
2:42:11 make money like I've I've gotten some
2:42:13 speaking gigs out of this
2:42:14 I had one that paid I didn't know it was
2:42:16 going to pay paid me after the fact that
2:42:18 was cool
2:42:19 but um
2:42:20 no this is I'm here just trying to
2:42:23 figure [ __ ] out and I'm sharing what I'm
2:42:25 learning kind of on the Fly because I
2:42:27 think it's important
2:42:29 I think it's actually really important
2:42:31 it's worth my time to do this and I
2:42:33 would rather be Sim racing than doing
2:42:34 this but I'm doing this
2:42:36 because I love y'all I love you so much
2:42:40 no
2:42:42 you had me you look like Rosie O'Donnell
2:42:53 what site was that you just went to
2:42:55 um I don't know which side I just went
2:42:57 to oh I think that was when I was
2:42:58 showing um
2:43:00 I went to 11 labs for audio D Dash ID
2:43:06 um was the one I went to to animate my
2:43:08 face I assume that's the one you're
2:43:10 asking about
2:43:11 he's very humble or Kyle I am I'm you
2:43:14 know
2:43:15 I haven't escaped the self-deprecating
2:43:17 humor thing I I know it doesn't put me
2:43:19 in my power Zone but I don't give a [ __ ]
2:43:22 I still have a I still have an issue
2:43:24 with
2:43:26 acting too powerful makes me a pompous
2:43:29 ass like so I try to avoid that you know
2:43:32 I don't know maybe I'll figure it out by
2:43:34 the time I'm 80.
2:43:37 I'm 58 I'm fairly wired so what are you
2:43:40 gonna do all right thanks for all the
2:43:43 cool info when do you go live generally
2:43:45 in the evenings I didn't go live last
2:43:47 night because oh my wife was in New York
2:43:49 she flew in I had to pick her up we hung
2:43:51 out
2:43:52 and by the time like that all wound down
2:43:54 it was too late where can I start
2:43:56 standing apart from the crowd everyone
2:43:59 is prompt engineering these days
2:44:01 um well so if you're questioning Lang
2:44:04 chain private GPT llama all of those
2:44:07 things GPT for all
2:44:10 if you're geeky enough to start playing
2:44:12 with Lang chain and things like that
2:44:14 um definitely definitely start playing
2:44:17 with that stuff because
2:44:19 I think that
2:44:21 Rachel Woods who's on Tick Tock here
2:44:23 she's got a
2:44:24 a newsletter called the AI exchange
2:44:26 she's talking about this concept of AI
2:44:29 operations that that companies are going
2:44:31 to need people that know what the [ __ ]
2:44:32 they're doing with this AI stuff and if
2:44:34 you can go a level deeper than prompt
2:44:36 engineering and do things like yeah I
2:44:39 can create us a standalone server I'll
2:44:40 put link chain on it you know we'll
2:44:43 embed all of our documents getting get
2:44:45 them into our own private Vector library
2:44:46 and if you can do that and even talk
2:44:49 like that you'll get hired for some [ __ ]
2:44:53 um that said the the the prompt
2:44:56 engineering thing I I think it is a
2:44:58 profession is is kind of get drift into
2:45:01 the background because these the tools
2:45:03 are going to get smarter and smarter at
2:45:05 self-improving the prompting so I don't
2:45:08 think you know everyone needs to become
2:45:11 a prompt engineer I think within the
2:45:13 next two years having that skill is
2:45:15 going to be really valuable but I think
2:45:17 to you to your point it's going to get
2:45:19 oversaturated and as the tools get
2:45:21 better and better the need for just
2:45:23 prompt Engineers is probably going to
2:45:25 diminish a bit but I think there's a
2:45:27 solid I don't know year and a half two
2:45:29 years
2:45:30 but if you can go layer deeper and start
2:45:33 to develop applications you know gluing
2:45:35 these tools together and managing that
2:45:37 process I think there's
2:45:39 there's something there
2:45:40 Can I some summarize the walk story
2:45:43 again arrived at the end walk
2:45:46 what was the walk story remind me what
2:45:48 it was and I'll I'm happy to re-share it
2:45:50 thank you tomorrow I will play with it
2:45:53 versus always watching others do it yes
2:45:55 Julie if you're still here
2:45:57 um yeah just dig in and start playing
2:45:59 with it and the the difference between
2:46:02 see it's I I you can't describe
2:46:08 it's different you look at it from the
2:46:10 outside it looks like a fancy Google
2:46:11 search
2:46:13 and when you use it you get that it's
2:46:16 not that at all
2:46:19 it's just different it's a completely
2:46:21 different experience and listen
2:46:24 we've been talking about AI powered chat
2:46:27 Bots for a [ __ ] decade and a half I
2:46:30 had a a product guy that worked for me
2:46:33 like six years ago all he talked about
2:46:35 was
2:46:36 AI chat Bots and how powerful they were
2:46:39 in Aunt did that and everyone I looked
2:46:40 at they were not powerful they were
2:46:42 pieces of [ __ ] that were manually coded
2:46:44 decision trees in the minute you would
2:46:47 ask it anything outside of its decision
2:46:49 tree it would just explode it was just
2:46:50 horrible
2:46:52 and this is not that like I almost feel
2:46:54 like calling it chat gbt diminishes the
2:46:57 power of it because you kind of think of
2:47:00 these shitty chat Bots
2:47:01 it is just a fundamentally different
2:47:04 experience so
2:47:05 Kudos jump in start playing start
2:47:08 digging get curious what kind of new
2:47:10 tools do you think are possible with the
2:47:12 new function called plugins
2:47:14 API so so open AI announced that that
2:47:17 you can do you can call functions from
2:47:20 within the the API so so basically what
2:47:24 it allows is rather so what the plugins
2:47:26 do is you've got this chat EBT thing and
2:47:29 if someone wrote a plug-in let's say
2:47:31 they wrote a plug-in they did called
2:47:34 wolf or mouth Alpha and it's an API that
2:47:38 goes to the Wolfram Alpha site which can
2:47:40 do math stuff
2:47:41 and then you write something in chat GPT
2:47:44 and it calls that it figures out oh I
2:47:46 need to do math stuff let me use the
2:47:49 wolf from alpha plugin I'll go do my
2:47:52 calculations over there and bring the
2:47:54 data back
2:47:55 what the new function calls allow you to
2:47:57 do is basically do that from within the
2:48:00 the the the the
2:48:01 API itself the op the chat GPT API
2:48:05 itself basically
2:48:06 and you can call your own apis or your
2:48:09 own data or you can you know send an
2:48:11 email you don't need to to use the
2:48:13 plugins you're in complete control of
2:48:15 the functions you can call so you can
2:48:16 define a set of functions that this
2:48:19 particular tool that you created has
2:48:21 where you're doing some chatting and if
2:48:24 at some point it needs to go off and get
2:48:26 customer data from Salesforce you can do
2:48:29 that so I I honestly don't know I
2:48:32 haven't played with it and it's like I'm
2:48:34 having a tough enough time getting my
2:48:36 head around
2:48:38 plugins and how you
2:48:40 deploy them in a useful way like other
2:48:43 than a like I feel I feel like when
2:48:45 you're dealing with new technology like
2:48:47 this you start out with the parlor games
2:48:49 the parlor tricks right which is like
2:48:51 write me a poem do this do that and and
2:48:54 then you kind of get into okay now I
2:48:56 understand it enough that I can actually
2:48:57 start to think about things critically
2:48:59 and problem solving and how I would
2:49:01 deploy them so I'm nowhere close to it
2:49:03 yet with the function calls like I'm
2:49:05 just getting my head around plugins and
2:49:06 so
2:49:07 so I don't have much for you there but I
2:49:09 think it's I think it's a profoundly
2:49:11 powerful uh step in the right direction
2:49:13 where can I find news on this type of
2:49:16 stuff so let me show you
2:49:19 I'll show you a source that that I've
2:49:21 put together
2:49:24 um
2:49:26 so
2:49:27 the the salon is uh the AI Salon is a
2:49:31 salon I put together about six months
2:49:33 old we meet bi-weekly online and in
2:49:36 person here in Denver and then
2:49:39 um everyday AI is a newsletter I created
2:49:41 so if you go to Twitter and go to
2:49:43 Everyday AI news that's my newsletter
2:49:46 um and it's it's
2:49:48 um we put it out weekly although the the
2:49:50 last week we didn't put one out because
2:49:52 of some
2:49:53 some personal issues between Greg and
2:49:56 myself that put it out but I'll show it
2:49:59 to you here
2:50:01 so this is everyday Ai and then each
2:50:04 issue we basically
2:50:06 um sort of Round Up what were the big
2:50:08 things in the week and we try to
2:50:09 prioritize them and say you know you
2:50:12 know here like here's the here's the big
2:50:14 stuff right so AI was on Capitol Hill
2:50:16 this week and chat GPT had new models
2:50:19 out and Elon Musk showed his Optimus
2:50:22 robots and said that everyone on planet
2:50:25 Earth is going to own a robot maybe two
2:50:27 right so he said the target market is
2:50:29 huge
2:50:31 um and so that's what openai is and so
2:50:33 our everyday AI is
2:50:35 and
2:50:36 um we also do Twitter threads
2:50:39 um on a on pretty much a daily basis
2:50:41 trying to round up interesting things
2:50:44 that have come along so that's a way
2:50:46 then then also I would follow Rachel
2:50:48 Rachel Woods here on
2:50:51 um
2:50:51 on Tick Tock she's got the AI exchange I
2:50:55 would follow
2:50:56 um uh McKenzie Bose I think his tag on
2:50:59 Tick Tock is GPT boss
2:51:02 um he's a developer he's got a an app
2:51:04 called GPT boss which is pretty pretty
2:51:06 slick wrapper on top of GPT three and
2:51:09 four
2:51:10 um
2:51:11 there's a guy named Paul rateser with
2:51:13 the marketing AI Institute I'd follow
2:51:16 him on LinkedIn
2:51:18 um he's also got a weekly podcast that
2:51:20 he does that's really good I listen to
2:51:22 it every week he's very good he's been
2:51:24 doing this for like five years I've been
2:51:25 doing it for for like six months
2:51:28 um
2:51:29 and uh if you're if you're really geeky
2:51:33 on YouTube there's a guy named David
2:51:35 Shapiro
2:51:37 um doing some really deep he's written
2:51:41 five books on AI like he knows his [ __ ]
2:51:43 so when I want to slice a humble pie I
2:51:45 go watch some David Shapiro like
2:51:47 whenever I think oh I'm A Smarty Pants I
2:51:50 go watch David Shapiro and I'm like I'm
2:51:52 a dum-dum
2:51:56 I make pretty pictures and I make poems
2:51:59 with the chapter jpt
2:52:02 because he's smart she's very smart
2:52:06 all right everybody gives their info to
2:52:09 AI how can that possibly be okay same
2:52:12 reason we gave our info to Google and
2:52:14 Facebook and Twitter
2:52:16 why is that okay because we got
2:52:17 something cool out of it same thing
2:52:19 you're you're given your stuff in
2:52:22 exchange for
2:52:24 free access to the most powerful tools
2:52:26 in the history of humanity don't give
2:52:28 them their info if you don't want to
2:52:32 um
2:52:33 I mean you know chat GPT was essentially
2:52:37 trained on the last 30 years of Internet
2:52:39 content
2:52:40 and you can just magically access it all
2:52:43 I don't know feels like a good trade to
2:52:45 me
2:52:46 but don't you know don't use it
2:52:49 where do you Sim race I Sim race on
2:52:52 um uh what's it called a set of course a
2:52:56 competition
2:52:58 and then I just go to the like the I
2:53:01 don't know whatever the public lobbies
2:53:02 like I'm not in any sort of League or
2:53:04 anything I just go to the public lobbies
2:53:05 and and I try to I try to find people
2:53:07 that are better than I am and I just get
2:53:09 my ass handed to me but I've gotten good
2:53:11 enough now like
2:53:13 I'm I I can pretty consistently like if
2:53:16 I'm if I'm dealing with decently rated
2:53:18 people I can consistently come in like
2:53:22 between Sixth and 10th like I'm a decent
2:53:25 mid-packer front to Mid Packer if I'm
2:53:27 with a bunch of Dum Dums I can generally
2:53:29 do one you know first second or third I
2:53:31 usually Podium with dum-dums and then
2:53:34 occasionally I'll get in a race where
2:53:35 like there's 20 people in there that are
2:53:38 like leagues above where I am and I just
2:53:41 like fool around in the back just
2:53:43 getting smashed spinning out trying to
2:53:46 catch up with everyone so
2:53:48 any ideas for investing in AI besides
2:53:51 Google Adobe I don't know I mean I I'm
2:53:53 impressed with Microsoft right now
2:53:55 open AI obviously but they're not public
2:53:58 I like anthropic they're not public
2:54:01 um Nvidia just hit a trillion dollar
2:54:03 market cap I wish it like I don't invest
2:54:06 so it doesn't [ __ ] matter but
2:54:08 um I wish I'd invested in Nvidia two
2:54:11 years ago because they've doubled and
2:54:13 then doubled a half a half again this
2:54:15 year and they're at a trillion dollar
2:54:16 market cap and
2:54:18 like how do they not have any [ __ ]
2:54:20 competition right now
2:54:23 because apparently chip manufacturing is
2:54:25 really [ __ ] hard and like AMD and
2:54:26 Intel
2:54:28 can't compete I don't know it's [ __ ]
2:54:31 insane
2:54:33 like I would just start paying it pay
2:54:34 attention just pay attention to like I
2:54:36 think Google's in trouble actually like
2:54:38 I would not invest in Google right now
2:54:40 until they can show something that is
2:54:44 interesting I wouldn't invest in meta
2:54:46 right now like both Google and meta are
2:54:48 like we've got AI too our AI is better
2:54:50 than everyone's
2:54:53 I mean at least meta right now is open
2:54:55 sourcing some of their interesting
2:54:58 um
2:54:59 research Google I don't know what's
2:55:01 going on at Google they're they're
2:55:03 trying to catch up but they kind of
2:55:05 screwed the pooch when they
2:55:06 when they left their foot off the gas
2:55:08 after they they invented the transfer so
2:55:10 Google invented this thing called the
2:55:12 transformer in 2017 that is the key that
2:55:16 unlocked all of the tools that we're
2:55:18 playing with and talking about today
2:55:19 they invented that [ __ ]
2:55:22 and then they recognized ooh this is an
2:55:25 existential threat to the search engine
2:55:27 business and that's our only business
2:55:31 and they kind of sat back on it and let
2:55:34 Sam Altman and his Mary Banda folks over
2:55:36 there at openai take that [ __ ] and run
2:55:39 with it
2:55:40 and then it was good enough that
2:55:42 Microsoft said let us give you 10
2:55:44 billion dollars to make it even better
2:55:47 and then Google it oops
2:55:50 [Music]
2:55:53 what's your favorite AI chat you can
2:55:57 make money
2:56:00 I've been telling people this technology
2:56:02 is perfect for ADHD accessibility it's
2:56:04 also perfect for ADHD obsessive
2:56:07 compulsive disorder like you want to run
2:56:09 down a [ __ ] Rabbit Hole you can you
2:56:11 can pick 10 an hour with this stuff it's
2:56:14 it's [ __ ] insane like
2:56:16 I love sort of getting into the sort of
2:56:19 ADHD hyper Focus mode but it's like I
2:56:22 like I have never been more overwhelmed
2:56:24 in my life because like I'm good at that
2:56:26 [ __ ] I'm good at synthesizing figuring
2:56:28 [ __ ] out quick running deep really quick
2:56:31 holy [ __ ] there are so many rabbit holes
2:56:33 here it's just impossible I think it was
2:56:35 1978.
2:56:36 so they had AI back then I was 12 or 13.
2:56:39 yeah 1977 is when
2:56:42 I got access to my first PC
2:56:45 a trash 80 TRS-80 and uh that's when I
2:56:50 recognized that
2:56:51 um I was just too young to be part of
2:56:54 the PC Revolution like I I recognized I
2:56:57 I wasn't quite old enough to be like
2:56:59 Steve Jobs age or you know Bill Gates
2:57:02 and the Microsoft kids
2:57:04 and it really bummed me out and so when
2:57:07 when the World Wide Web came along and I
2:57:09 had the Epiphany I'm like oh [ __ ] this
2:57:11 is one of those I get to be part of this
2:57:14 from the beginning
2:57:16 and so I did that and that was [ __ ]
2:57:18 insane and awesome and now it's
2:57:20 happening again this is like the second
2:57:22 time in my life I'm getting to actually
2:57:25 jump on the bandwagon which is like if
2:57:27 you think about it
2:57:29 this AI [ __ ] in my opinion is like
2:57:31 Industrial Revolution scale disruption
2:57:35 like it's like printing press Industrial
2:57:38 Revolution fire written language like
2:57:42 these don't come along very often in the
2:57:45 history of humanity so the fact that
2:57:47 we're all alive to be a part of it is
2:57:50 far the [ __ ] out hi Kyle I love it too
2:57:54 but what do you think about the
2:57:55 dangerous machine learning from machine
2:57:58 um
2:58:00 I don't know I think it's I think it's a
2:58:02 little too early to to to know I heard
2:58:05 it yesterday talking about some research
2:58:07 that is um
2:58:10 foreign
2:58:12 language models can experience model
2:58:14 collapse when when they're they're doing
2:58:17 too much you know machine learning on
2:58:19 machine learning like where they're
2:58:20 taking they're taking synthesized data
2:58:22 from machine learning and dumping that
2:58:24 back into things that it can actually
2:58:25 collapse the model and they just they
2:58:27 they either cease to function or the the
2:58:30 um the quality of response just goes
2:58:32 down dramatically
2:58:34 um
2:58:35 I suppose that's a risk long term
2:58:38 but you know if you've got
2:58:42 I I don't know I don't know enough about
2:58:45 I don't know enough about the computer
2:58:46 science and the data science to know
2:58:49 um
2:58:50 what the danger there is and I assume I
2:58:53 can only assume that smarty pants like
2:58:55 Sam Altman and his engineers and
2:58:58 Microsoft and Google and meta
2:59:02 um
2:59:03 are aware of this right you know Google
2:59:05 had Jeffrey Hinton I mean he's a guy
2:59:07 that
2:59:08 invented a lot of the Core Concepts
2:59:13 um for how this stuff works so they
2:59:14 understand how it works and I think they
2:59:16 understand what the what the risks are
2:59:18 but I don't know I don't know
2:59:21 I'm I'm sort of
2:59:24 I I suppose I represent a kind of
2:59:26 entrepreneur that enters The Fray after
2:59:29 they've figured that [ __ ] out
2:59:32 so so if all of the companies [ __ ] up
2:59:35 all of the models at the same time then
2:59:37 I think we're in trouble but someone in
2:59:39 there is going to figure out how to not
2:59:40 [ __ ] it up and maybe it'll be the open
2:59:42 source community
2:59:45 [Music]
2:59:47 already started that process of Pinecone
2:59:50 training for company policies and
2:59:51 training procedures that's super cool JK
2:59:55 um do that I think that will serve you
2:59:56 well I think there's there's good
2:59:58 opportunity there uh looks like we got
2:59:59 some new people joining my name is Kyle
3:00:01 Shannon this is the AI learning lab I
3:00:03 think I've been doing this for three
3:00:04 hours now I'm a little moist I'm a
3:00:07 little Saucy uh I can get cranky I try
3:00:09 to make jokes I make funny faces
3:00:11 apparently I move around too much for
3:00:13 some people's tastes but hopefully
3:00:14 they've left now and other people think
3:00:16 I look like Rosie O'Donnell
3:00:19 if you're asking what the hell do you
3:00:21 know about AI what are your credentials
3:00:22 Kyle why should I listen to you well
3:00:25 I don't know much and that's why I'm
3:00:27 here I'm an entrepreneur from many years
3:00:30 I've done lots and lots of businesses
3:00:33 um I was a player in the early days of
3:00:35 the web and
3:00:37 had an epiphany around that and I
3:00:41 haven't had one of those epiphanies
3:00:42 since until
3:00:44 December of 2022.
3:00:46 with Chachi BT and so
3:00:49 I think we're at the precipice of a
3:00:52 massive
3:00:56 shift
3:00:57 and it's really exciting and that's why
3:00:59 we're here and I'm here to demystify so
3:01:00 if you have questions about AI pop them
3:01:02 down below I will do my best to answer
3:01:04 them I'm very far behind
3:01:06 so if you're like I think he's just
3:01:08 recorded I'm pretty sure he's a robot I
3:01:11 wonder if he's chat gpt5
3:01:14 you know he could be he doesn't look
3:01:17 natural have you noticed that his one
3:01:19 eye wanders
3:01:21 that can't be real
3:01:26 I really just try to entertain myself
3:01:29 have to go see you soon thank you bye
3:01:32 okay
3:01:33 I'd love an a sequel chat query program
3:01:38 um
3:01:40 I would if you haven't go try to have
3:01:44 Chachi PT develop that for you
3:01:47 the guy who thought about walking oh
3:01:50 okay so someone said before can you tell
3:01:52 me the walking story that came in at the
3:01:54 end of it okay
3:01:56 so so we're starting to see
3:01:59 we're starting to see large language
3:02:01 models being used for for interpreting
3:02:05 brain brains reading reading Minds
3:02:09 um so so there have been a bunch of them
3:02:10 there there was one in Austin where
3:02:13 um they're using fmri scans of the
3:02:16 brains and they're using large language
3:02:17 models to interpret them so one was they
3:02:20 had people watching video they would
3:02:23 interpret the brain scan and recreate
3:02:26 the video and it wasn't perfect but you
3:02:28 could tell it was if they were like
3:02:30 watching a video of a whale in the ocean
3:02:32 you'd see some sort of fish in the ocean
3:02:33 right it might not have been exactly a
3:02:35 whale but but that was pretty remarkable
3:02:37 they did it with images somewhere else
3:02:39 and then uh so they've been some of that
3:02:42 stuff so there's a guy I think it's in
3:02:44 Italy he's been paralyzed for 12 years
3:02:46 from a motorcycle accident and they used
3:02:49 large language models to scan his brain
3:02:52 when he was thinking about moving his
3:02:54 legs so they figured out
3:02:56 what patterns activate
3:02:59 when he thinks about moving his legs
3:03:02 then they built a Bluetooth Bridge they
3:03:05 took the signal
3:03:07 from the brain passed it through this
3:03:09 large language model that understood it
3:03:11 sent the signals Bluetooth down to an
3:03:14 electrode in his lower spine that you
3:03:17 know was you know surgically implanted
3:03:19 into his spinal cord and he started
3:03:22 moving his legs he would think about
3:03:23 moving his legs it sent those signals
3:03:25 down
3:03:26 his legs interpreted that as I'm moving
3:03:30 my legs now and he's now walking
3:03:33 just like [ __ ]
3:03:38 holy [ __ ] really
3:03:42 um I got plug-in Dev 4.0 32k API spent
3:03:47 money on API bills still no code
3:03:50 interpreter have you seen it no I have
3:03:53 not spent it I do not have the 32 000
3:03:56 API uh they just released the 16 000 API
3:03:59 for 3.5 but I do not have code
3:04:02 interpreter I don't I'm like either it's
3:04:05 really expensive for them to do that or
3:04:06 there's something broken about it or or
3:04:08 it's it's like still really risky
3:04:10 security wise there's some reason
3:04:12 they're not releasing that wider because
3:04:14 it looks [ __ ] remarkable
3:04:18 so no I don't have it I'm cranky about
3:04:20 it I don't like it
3:04:22 deploy script or notable plugins I don't
3:04:25 know
3:04:27 I how many plugins are there now let's
3:04:30 go look let's go to the videotape
3:04:34 um
3:04:39 we are gonna go to the plug-in store and
3:04:43 see how many they got last time I
3:04:45 checked there were 400.
3:04:49 oh there's a new thing this is actually
3:04:51 cool there's a new plugin called plug-in
3:04:53 pedia
3:04:55 which you can ask it for plugins and
3:04:58 it'll go find ones for you which is kind
3:05:00 of cool
3:05:01 um but the plug-in store has a search
3:05:04 box now so I don't think that's really
3:05:06 all that necessary
3:05:08 so if we go all and go look okay there's
3:05:11 a hundred and eight times four so that's
3:05:13 432.
3:05:16 430 plugins my problem with plugins
3:05:19 right now is uh
3:05:22 there's a lot of them they're very
3:05:24 complicated
3:05:25 and I just don't have time so I'm kind
3:05:28 of ignoring them by the way if you're
3:05:29 new here welcome uh this is the AI
3:05:31 learning lab if you haven't played with
3:05:33 chat gbt that top URL there
3:05:35 chat.openai.com that's the official chat
3:05:37 EBT website go there start playing Bing
3:05:40 and poe.com or both also places where
3:05:43 you can play with chat gbt so go do that
3:05:46 um
3:05:47 if you're if you're enjoying this tap
3:05:49 the screen or double tap it or whatever
3:05:51 you do to make a heart and that'll let
3:05:52 people know we're here
3:05:54 and the other things prompts.chat and
3:05:57 futurepedia.io are
3:05:59 um just resources that might help so
3:06:01 these sites are there for your
3:06:06 clicking and typing pleasure
3:06:08 um I really just have them there because
3:06:10 people always ask you know where should
3:06:11 I start if I want to get involved in AI
3:06:13 where do I start where do I start start
3:06:15 with chat EBT start to learn
3:06:18 what these tools make possible why they
3:06:21 make why they've got people so [ __ ]
3:06:24 hot and bothered why is everybody
3:06:26 talking about AI
3:06:29 I don't understand what the big deal is
3:06:32 it's just it's like a Google search box
3:06:34 but it's a longer response I don't know
3:06:36 what the big idea is I don't get it
3:06:39 Go Get It Go understand go to
3:06:43 chat.openai.com if you haven't played
3:06:45 with it
3:06:46 then learn for yourself why we're all so
3:06:50 hot and bothered
3:06:52 I am so moist
3:06:54 I'm all worked up over here
3:06:56 people using chat GPS people using GPT
3:06:59 in program names and websites have been
3:07:02 getting cease and desists whoops
3:07:05 using in program names and websites
3:07:08 I don't quite know what that means or
3:07:10 what their cease and desists are for
3:07:12 there's going to be listen
3:07:14 because chat GPT mid-journey stable
3:07:18 diffusion probably dolly
3:07:21 barred
3:07:23 because all of these things were trained
3:07:25 on all of the internet all of them were
3:07:26 trained on copywritten data except for
3:07:29 Adobe adobe's Firefly model
3:07:32 they have consciously said we're only
3:07:34 going to train it on [ __ ] we own or
3:07:35 public domain images smart move well
3:07:38 smart move if you want to use things
3:07:39 commercially all these other companies
3:07:42 are gonna there's gonna be lawsuits of
3:07:44 fly-in the next 20 years is just going
3:07:47 to be like one lawsuit after another
3:07:48 it'll work itself out I wouldn't worry
3:07:51 about it David Shapiro is also artistic
3:07:54 so hard to compete with that I think you
3:07:56 mean autistic he is autistic he's neuro
3:07:59 spicy as he calls it you know I'm in the
3:08:01 neuro spicy Club I think he's a level
3:08:03 deeper in the neurospice than I am but
3:08:06 the advantage of that particular
3:08:07 neurospice is the [ __ ] can go
3:08:10 deep with the intelligence he's he's
3:08:11 [ __ ] he's [ __ ] brilliant
3:08:14 so I'm a big fan of his
3:08:17 there's a chat GPT and wait so there's
3:08:20 chat GPT and Google's any other AI
3:08:23 That's accessible yes so
3:08:26 um look up a project called GPT for all
3:08:30 GPT for all and it's it's a it's a
3:08:34 and
3:08:36 it's a thing you can install on your
3:08:38 local computer where you can access
3:08:41 different open source large language
3:08:43 models so these are large language
3:08:45 models that are not chat GPT they're not
3:08:48 Google they're not they're based on meta
3:08:50 but they're not meta they're fine-tune
3:08:52 models on top of that you can run them
3:08:54 all locally like on an M1 Mac you can
3:08:57 just run these models
3:08:59 um I would also look at
3:09:02 um Lang chain if you're geeky
3:09:04 um because Lang chain you can start to
3:09:05 do things like build memory and search
3:09:08 on your own data if you're not geeky
3:09:10 just wait that stuff will come
3:09:13 um the other thing here I'll show you
3:09:15 the the other thing to play with the
3:09:19 the third URL here poe.com I'll show you
3:09:23 poe.com this is worth playing with
3:09:27 so Poe is just like chat GPT but you'll
3:09:30 notice down the left hand side here let
3:09:32 me make it a bit bigger
3:09:34 you'll notice down on the left hand side
3:09:36 here you've got Claude instant oh wait
3:09:38 you got Sage so Sage is pose
3:09:41 um large language models so this is like
3:09:43 chat gbt but it's it's its pose now the
3:09:46 company behind Poe is quora the question
3:09:49 and answer site
3:09:50 so
3:09:52 um they've been digging into this stuff
3:09:53 so within Poe you have access to Sage
3:09:56 there are three models that you have to
3:09:58 subscribe to you you buy their premium
3:10:01 thing it's like 20 bucks a month
3:10:03 and it's you get access to gpt4 Cloud
3:10:05 plus and this one in particular is
3:10:07 interesting Claude instant 100k
3:10:11 that's a hundred thousand token chat bot
3:10:14 so that means you could paste in a 75
3:10:18 000 word novel and start querying
3:10:21 against it that's [ __ ] amazing
3:10:24 the other thing about Claude the these
3:10:26 three Claude instance really fast 100k
3:10:30 allows you to do really big documents
3:10:32 and then Cloud plus is like gpt4 quality
3:10:35 wise all these Cloud things are from a
3:10:37 company called anthropic it's a bunch of
3:10:40 EX open AI people
3:10:43 um that are doing their thing uh over
3:10:45 there and they're they're trying to do
3:10:46 it better than open AI so so Poe's
3:10:49 probably pretty interesting for you to
3:10:50 start playing on
3:10:53 all right Nvidia Intel Oracle Microsoft
3:10:56 Meta Meta
3:10:58 um
3:10:59 Maybe
3:11:03 um
3:11:07 they have a lot of money but they kind
3:11:09 of screwed the pooch on the
3:11:11 rebranding
3:11:12 they put they took a big swing and a big
3:11:14 bet and um
3:11:18 the last earnings call they didn't say
3:11:19 the word metaverse once
3:11:23 Apple didn't say the word metaverse once
3:11:25 when they launched their metaverse
3:11:26 viewing goggles
3:11:28 laughs
3:11:32 so that was a nice little troll they
3:11:34 also didn't use the word the phrase AI
3:11:37 Appleton AMD is supposedly picking up an
3:11:40 Amazon contract for GPU stuff that
3:11:43 doesn't surprise me and who was it
3:11:44 didn't
3:11:45 someone just said they're gonna start
3:11:47 making
3:11:48 GPU chips was that Microsoft I don't
3:11:51 know someone
3:11:52 I think everyone's realizing holy [ __ ]
3:11:54 why is NVIDIA getting all all the AI
3:11:56 business
3:11:58 you know they they were sitting pretty
3:12:00 with all the the Bitcoin mining [ __ ] and
3:12:02 then all of a sudden AI is like making
3:12:04 that blockchain [ __ ] look look like
3:12:06 Child's Play
3:12:08 Google snuffing the search results yeah
3:12:10 I know Google's Google's got to deal
3:12:13 with this
3:12:14 you can make my you can make money
3:12:17 [Music]
3:12:20 thank you Eric
3:12:24 for all Cap
3:12:26 all cap capturing that
3:12:28 [Laughter]
3:12:32 all right I'd be having philosophical
3:12:34 debates about the nature of reality with
3:12:36 my AI
3:12:37 it's really good at philosophical [ __ ]
3:12:40 it's way more nuanced if you ask it
3:12:41 about you know polarizing topics that
3:12:44 you can't have with humans in today's
3:12:46 day and age you can actually have it
3:12:48 with chat EBT and it gives you really
3:12:50 nuanced answers
3:12:51 it might even encourage thinking in
3:12:54 humans which I know we're not fans of
3:12:56 that anymore but if we get back to that
3:12:58 thing where as humans we can apply logic
3:13:01 to the situation without being castrated
3:13:03 maybe Chachi BT can help us do that that
3:13:07 would be nice
3:13:13 you will get a chance to get in Nvidia
3:13:16 I'm sure they will bring it down we'll
3:13:18 see
3:13:19 I worry about the upcoming generation
3:13:22 who grew up on phones instead of
3:13:24 computers they can't even
3:13:26 navigate file structure go to command
3:13:29 prompt or any what why would they need
3:13:31 to
3:13:33 why would they need to
3:13:36 Windows 11 is going to have
3:13:38 um chat EBT built in at the operating
3:13:41 system level you can just say to your
3:13:42 machine go configure yourself
3:13:45 turn off all the
3:13:47 fancy [ __ ]
3:13:49 and it'll just go do it why would they
3:13:51 need to
3:13:52 I mean if they want to be hardcore
3:13:54 Engineers sure
3:13:56 but even hardcore Engineers like like
3:13:58 there's going to be you know some level
3:14:00 that are always going to be done at the
3:14:01 infrastructure level but so much of the
3:14:04 world like I feel like there's just this
3:14:06 massive
3:14:08 layers upon layers upon layers of
3:14:11 complexity
3:14:12 in today's world
3:14:14 think insurance claims think taxes think
3:14:19 buying a house all that complexity there
3:14:22 are entire Industries built around
3:14:25 hiring people
3:14:27 to deal with that complexity
3:14:29 we now have a tool that sits above all
3:14:32 that that automates the [ __ ] out of it
3:14:34 why is that a bad thing
3:14:38 I think that's a really good thing
3:14:40 so
3:14:42 I don't worry about those kids
3:14:46 I don't
3:14:49 I think they might have
3:14:53 a better flexibility
3:14:55 at flushing those soul-crushing shitty
3:14:58 jobs down the toilet
3:15:00 than we do
3:15:02 right I think I I think that
3:15:06 it's very easy to be married to if
3:15:09 you're not putting in the time it must
3:15:11 be bad
3:15:14 so
3:15:14 if I can write a business letter in 30
3:15:18 seconds that's as good or better than
3:15:21 what I would have written on my own
3:15:24 does the fact that I didn't take an hour
3:15:26 to do that make it any less
3:15:28 useful a letter
3:15:31 I don't know
3:15:33 I don't think so who gives a [ __ ] how
3:15:36 long it took me
3:15:37 I just happened to use a tool that was
3:15:39 really good I was smart enough to use a
3:15:42 tool that was smarter than I was
3:15:44 that's where I am with it
3:15:47 I mean the same thing's happening with
3:15:48 reality visualization I'm not sure what
3:15:51 that was in reference to Eric but yes
3:15:52 that sounds right I read about model
3:15:54 collapses yesterday too caruno
3:15:57 um or cave Runo or KV Runo
3:16:01 um I know that model collapse thing like
3:16:03 that's a term I hadn't heard before when
3:16:05 I heard it I'm like oh that like
3:16:07 that would suck if you're like if you'd
3:16:09 spent like three years building a model
3:16:11 and you're like oh
3:16:13 we we we're we're fine-tuning it with
3:16:16 other large language models and it's
3:16:18 going really well it's going great it's
3:16:19 going great
3:16:21 model collapse [ __ ]
3:16:24 you have a backup
3:16:27 anybody got a backup of the large
3:16:29 language model anybody Bueller
3:16:33 Bueller
3:16:35 summarization applications are pretty
3:16:37 cool use case a transcript and a
3:16:39 two-hour meeting into bullets yeah
3:16:40 that's amazing
3:16:42 I do that I do that well that's with
3:16:44 with um Claude 100k
3:16:49 Flip Flip the camera
3:16:52 so
3:16:53 I still have it in here
3:16:57 hang on
3:16:58 oh this was
3:17:02 yeah this was a uh this was a transcript
3:17:05 so I pasted in like a an hour long or
3:17:09 two hour meeting transcript and then
3:17:11 just started asking questions about it I
3:17:14 also
3:17:16 um I went out on the internet and I got
3:17:18 the
3:17:19 um screenplay of The Godfather the Mario
3:17:22 Puzo screenplay of The Godfather and has
3:17:24 pasted that in there started asking
3:17:26 about the characters it's pretty cool
3:17:28 will AI motivate non-ai companies to use
3:17:31 blockchain to verify real products or
3:17:33 digital or otherwise yeah I was talking
3:17:34 about that earlier Pete BOGO good name
3:17:38 Pete buy one get one free
3:17:41 buy one got one yeah
3:17:46 get one tomorrow
3:17:50 um
3:17:51 yeah I think that
3:17:54 um in the next year
3:17:56 it's this
3:17:59 they ain't gonna be they they gonna be
3:18:01 much movement there but then the
3:18:03 presidential election in the United
3:18:05 States is going to cause some situations
3:18:09 if we think misinformation was bad in
3:18:12 2016 and 2020
3:18:14 oh boy
3:18:16 2024 is going to be insane
3:18:19 and so
3:18:21 so I think what's going to happen and I
3:18:23 mean it could be really destructive like
3:18:25 it's like I I'm kind of making fun of it
3:18:28 but I but you know I I learned in 16 and
3:18:31 17 like when I was a kid in the in the
3:18:34 late 70s and early 80s I learned about
3:18:36 propaganda and misinformation it just it
3:18:39 just seemed like this thing like oh
3:18:40 we're past that now and I just I
3:18:43 couldn't imagine a world in which like
3:18:44 who would fall for that [ __ ] and then
3:18:47 2016 and 2020 happened I'm like oh we do
3:18:51 like we did
3:18:53 and and uh but I think that 2024 is just
3:18:57 it's gonna make that [ __ ] look like
3:18:59 ridiculously simple I think it's going
3:19:02 to be very sophisticated so I think 2025
3:19:05 2026 the the like the impact of it is
3:19:08 going to be
3:19:11 we're not going to be able to trust
3:19:13 anything
3:19:15 and so that's going to precipitate well
3:19:17 how can we trust it well we can tokenize
3:19:19 it right we can we can use blockchain we
3:19:22 can use nfts we can use whatever other
3:19:24 flavor of tokenization comes along like
3:19:26 it doesn't have to be
3:19:28 um blockchain and nft but but some sort
3:19:31 of tokenization that can validate
3:19:33 this human's a human right so go look at
3:19:36 Sam Walton's project World coin where
3:19:39 he's scanning irises he's going to send
3:19:41 you a little ball that you put against
3:19:42 your eye it scans your Iris verifies
3:19:45 you're human and then you get
3:19:46 cryptocurrency as the payment why is he
3:19:49 doing that because he wants to be able
3:19:52 to have a system that these people are
3:19:54 verified human
3:19:56 right and so
3:19:58 by extension then you could say Okay
3:20:00 Kyle is verified human on at a
3:20:03 blockchain level and he just made this
3:20:06 video and that's verified that Kyle made
3:20:08 that video and it has not been modified
3:20:11 right or it might be
3:20:14 Kyle is verified as human he's made this
3:20:16 official synthesized version of himself
3:20:19 and this video was created with Kyle's
3:20:22 official synthesized version so it's
3:20:24 real Kyle content with the synthesized
3:20:27 version of him verified by the real Kyle
3:20:29 right so yeah those systems are going to
3:20:32 need to be built so
3:20:34 my prediction is 2024 is a [ __ ] show and
3:20:37 then the next two years are people going
3:20:39 we need to know what's real because we
3:20:41 can't tell anymore it's very upsetting
3:20:45 Nana called me and told me to send her
3:20:48 300 so I did and then I thought Nana at
3:20:51 the picnic and she never got the money
3:20:54 that's happening now
3:20:57 oh it's crazy
3:20:58 [Laughter]
3:21:00 Watson used to help
3:21:03 Watson used to help with medical yeah I
3:21:06 mean I'd be a couple of things about IBM
3:21:09 so they did Watson health health and
3:21:11 they somehow flamed it out and I you
3:21:13 know I don't think it's necessarily that
3:21:14 their technology was bad but but the
3:21:18 I do a lot of work in in healthcare and
3:21:20 Pharma the healthcare and Pharma
3:21:21 industry is so conservative and when
3:21:23 you've got people's lives on the line
3:21:25 and you know litigation and you know
3:21:28 conservative companies and insurances
3:21:30 involved
3:21:31 um it's just a really complicated space
3:21:33 so
3:21:35 so that might not have been a situation
3:21:36 where the technology wasn't there might
3:21:38 have been right
3:21:40 um because I think it was
3:21:41 pre-transformer so so they might have
3:21:43 been doing sort of vertical data sets or
3:21:45 something like that but
3:21:48 um I did just hear that IBM
3:21:50 um just created a center of excellence
3:21:54 around prompt engineering and they have
3:21:57 a thousand people at IBM doing nothing
3:21:59 but writing prompts creating a prompt
3:22:01 Library
3:22:02 so when those big companies decide
3:22:04 they're going to do something and also
3:22:06 they re-released Watson I think they
3:22:08 called it Watson X right they're like
3:22:10 hey dust off Watson get that technology
3:22:13 back out here it's cool again let's call
3:22:15 it Watson X and put it out there uh but
3:22:17 they have a thousand people working on
3:22:19 prompt crafting there so uh it'll be
3:22:21 interesting to see if if IBM can kind of
3:22:23 reclaim any any sort of credibility in
3:22:25 the space because you know they were
3:22:27 with with uh with deep blue whatever it
3:22:29 was that beat Casper off like they were
3:22:31 they were in the lead for a while and
3:22:33 Watson Health looked really really
3:22:34 promising and then it just kind of you
3:22:37 know fizzled so it's gonna be
3:22:39 interesting to see if they can make a
3:22:40 comeback
3:22:41 should sales rep position themselves for
3:22:43 AI [ __ ] yes
3:22:45 I mean seriously like cold calling
3:22:50 like the the the ability to synthesize
3:22:54 yourself yourself as a salesperson and
3:22:58 have that synthesized version of you do
3:23:00 plausible sales Outreach calls
3:23:03 automatically 24 hours a day
3:23:07 um
3:23:08 that isn't more than a year away
3:23:11 maybe two
3:23:14 so like that and that's just a part of
3:23:16 it you know Salesforce just announced I
3:23:18 think 10 different AI tools within
3:23:21 Salesforce so you know think about the
3:23:24 amount of time you spend doing data
3:23:26 entry into [ __ ] Salesforce it's the
3:23:28 bane of my goddamn existence I don't do
3:23:30 it I'm horrible at it and it's important
3:23:33 um but it's so convoluted and shitty
3:23:36 if I can just say
3:23:38 you know create an account and an
3:23:40 opportunity for this dude that I just
3:23:42 talked to paste in the notes and it just
3:23:44 figures all the [ __ ] out and puts it in
3:23:46 there
3:23:47 that you know list generation I mean
3:23:49 just every aspect of sales is going to
3:23:52 be some version of automated so I would
3:23:55 learn this ship pronto
3:23:58 start writing with it like yeah like uh
3:24:01 like everything marketing materials
3:24:05 content social strategy like I don't
3:24:07 know how you do your job but
3:24:09 go learn the tools that can replace your
3:24:12 job and like disrupt yourself is is
3:24:14 something that I've been saying recently
3:24:16 figure out a way to disrupt yourself
3:24:18 looks like we have some new people here
3:24:20 my name is Kyle Shannon this is the AI
3:24:22 learning lab I'm an entrepreneur I've
3:24:24 been doing this for about three and a
3:24:25 half hours now so my voice is going I'm
3:24:27 flop sweaty I'm not flop sweaty but I am
3:24:29 moist uh and and uh yeah if you have
3:24:34 questions about AI pop them in I'll do
3:24:36 my best to answer them but I'm getting a
3:24:37 little crispy so I think the quality of
3:24:39 the answer is just diminishing and the
3:24:41 the amount of smart ass answers is going
3:24:43 up and smarmy answers
3:24:47 ah but I do have iced tea the problem
3:24:50 with the metaverse is that major
3:24:51 companies have been using top Nvidia
3:24:54 chips
3:24:55 the problem with the metaverse is that
3:24:57 it just looks like Second Life still
3:25:00 like then we have second life in 2001.
3:25:02 didn't it suck back then
3:25:04 with like this awkward Junior High
3:25:06 School dance where people just stood
3:25:07 around dancing like
3:25:10 like why does it still look like that
3:25:15 I don't need [ __ ] goggles for that I
3:25:17 know there's cool [ __ ] there's cool
3:25:19 experiences that have been created but
3:25:20 it's too isolated look like apple what
3:25:24 what Apple clearly did with the Vision
3:25:25 Pro
3:25:27 I didn't even think this was possible
3:25:28 they like they got some they got some
3:25:31 [ __ ] horsepower in the Noggins over
3:25:33 there
3:25:33 they successfully separated themselves
3:25:36 they essentially have an identical
3:25:39 piece of Hardware to all the other
3:25:41 goggles right like like it it it's got
3:25:44 all these technical enhancements to it
3:25:45 and software enhancements to it that
3:25:47 make it different right but from a
3:25:49 positioning standpoint
3:25:51 this was not goggles to look at the
3:25:53 metaverse
3:25:55 they redefined it as spatial computing
3:25:59 you can just put your screens out there
3:26:01 and go do your [ __ ] oh and you want it
3:26:04 to be
3:26:05 the Super Bowl and you want it to be as
3:26:07 big as you know the mountains and you
3:26:10 want to see you want that giant screen
3:26:12 sitting in the mountains you can do that
3:26:13 like they just they completely and it
3:26:15 and this is a this is not goggles to
3:26:18 look at a computer this is a computer
3:26:22 that completely transforms how you
3:26:24 interact with computers you use your
3:26:26 eyes and your fingers there's no
3:26:27 controllers
3:26:28 and it interacts with you like magic how
3:26:31 did they [ __ ] do that well 5000
3:26:33 patents is how they did that but holy
3:26:36 [ __ ]
3:26:37 so
3:26:41 we're going to sneak into the metaverse
3:26:43 through [ __ ] like uh spatial Computing
3:26:46 metaphors
3:26:48 all right for most people with a modest
3:26:51 PC I try to run any metaverse platforms
3:26:54 it crashes yeah it's just
3:26:56 start using chat GPT it doesn't crash
3:26:58 and it's way more profound any
3:27:01 experience with uh cot meta programming
3:27:04 no none I don't know what cot is uh
3:27:12 I don't know what cot is
3:27:18 Internet Internet of Things is that what
3:27:20 something of things is that like like
3:27:23 um sensors and devices I don't I don't
3:27:25 quite know what you're referencing
3:27:27 I I'm new to this like completely no so
3:27:29 if you're brand new to this welcome
3:27:32 um I'm relatively new to it so I'm I'm
3:27:34 not a machine learning or AI guy I'm an
3:27:36 entrepreneur
3:27:38 and so I've been through this once
3:27:40 before so
3:27:41 um if you're really really new to it I
3:27:43 put I put this list together so the the
3:27:46 top three things are are all basically
3:27:48 variations on the same thing so the top
3:27:50 one
3:27:51 um chat.openai.com that's Chachi PT's
3:27:54 official website if you've not played
3:27:56 with chatbt go sign up for an account
3:27:58 it's free and start playing with it
3:28:00 start asking it have it write you a poem
3:28:02 about your dog have it write you a song
3:28:06 about your significant other
3:28:08 um have it write you a meal plan for the
3:28:10 week and then turn that meal plan into
3:28:13 recipes and then turn those recipes into
3:28:15 a shopping list and then have it
3:28:16 organize that shopping list by category
3:28:19 of the grocery store just do that and
3:28:22 then when you pick your jaw up off the
3:28:23 floor go oh what else can this thing do
3:28:27 so so and then Bing is just a different
3:28:30 that's Microsoft search engine but you
3:28:32 can use chat GPT there as well and Poe
3:28:35 is another one where you can use
3:28:37 different search engines but if you've
3:28:38 never done anything before just go to
3:28:41 chat.openai.com I'm not affiliated with
3:28:43 it there's no there's no Shenanigans
3:28:45 here that's just their official website
3:28:46 and then prompts.chat will teach you how
3:28:49 to interact with this thing in a in a
3:28:51 decent way it'll at least get you
3:28:52 started and futurepedia.io that's just
3:28:55 really more like a directory of all
3:28:57 sorts of tools if you really want to
3:28:58 geek out and run down rabbit holes you
3:29:01 can go there but if you're just starting
3:29:03 out I would ignore that bottom one but
3:29:05 just go to chatgpt so
3:29:07 I hope that helps and and you have fun
3:29:10 with it like start having fun with it
3:29:11 because it's this stuff is really really
3:29:14 remarkable technology that we're lucky
3:29:16 enough to have our hands on so go start
3:29:18 playing
3:29:20 and don't worry about being new or
3:29:22 non-technical we're all [ __ ] new
3:29:23 there's no experts right now anyone who
3:29:26 says they're an expert is full of [ __ ]
3:29:27 well no wait hang on
3:29:31 if you're talking about machine learning
3:29:32 scientists talking about the
3:29:35 underpinnings of AI there are experts at
3:29:38 that the generative AI stuff like with
3:29:40 chat GPT moving forward like starting
3:29:43 November 30th 2022 until now six months
3:29:47 there are no experts at that [ __ ] it's
3:29:49 all too new it's moving too fast no
3:29:51 one's an expert
3:29:53 so and and the people that say you can
3:29:55 make money
3:29:56 they're full of [ __ ] too so ignore them
3:29:59 ignore the experts just start getting
3:30:01 curious just start learning
3:30:06 I gotta go soon I'm getting I'm getting
3:30:08 Punchy I'm getting punched drunk
3:30:12 all right I'm just scrolling back a
3:30:14 little bit it jumped me to the bottom so
3:30:15 let's see can you make a full IOS app
3:30:17 with chat GPT yeah you can
3:30:21 um
3:30:23 yeah I mean I mean the answer is yes you
3:30:27 can have it write you as an application
3:30:29 in Swift and teach you how to deploy it
3:30:31 and how to teach you how to do it or you
3:30:33 could use a no code tool like bubble I
3:30:36 think bubble does iOS you can find a no
3:30:38 code tool that does iOS things but yeah
3:30:41 there's lots of stuff out there but but
3:30:43 if you wanted to learn how to do an IOS
3:30:45 app from scratch just go in tell chat
3:30:47 apt to act like a master IOS app
3:30:51 developer and tell it you've never done
3:30:53 this before at all and have it walk you
3:30:55 step by step through the process and
3:30:57 that you want it to write the code in
3:30:59 Swift and
3:31:00 whatever the hell else and just
3:31:05 figure it out
3:31:07 because you can get there probably way
3:31:10 faster than you thought possible learn
3:31:12 all the new info and don't further huh
3:31:15 what happens when everything is generic
3:31:18 and soon uniformity
3:31:21 I don't think it will be and I'll tell
3:31:23 you why
3:31:29 what these tools are doing is
3:31:30 democratizing
3:31:34 access to powerful creative tools to
3:31:36 execution tools
3:31:39 and so to your point
3:31:41 if everyone's got access to the tools
3:31:44 they're just all going to make the [ __ ]
3:31:46 and there's going to be more [ __ ]
3:31:48 and because the tools are powerful the
3:31:50 [ __ ] is going to look shiny at first
3:31:52 like a good [ __ ] does
3:31:56 and then we're going to realize oh
3:31:57 that's just [ __ ] right so there's going
3:31:59 to be there's going to be more content I
3:32:01 I would argue that the the
3:32:03 the Baseline of quality will go up right
3:32:06 where overall content quality will go up
3:32:10 but there's going to be so much of it
3:32:12 that it's just going to be noise it's
3:32:13 just going to be digital noise and then
3:32:15 what's going to happen is this within
3:32:17 all the noise something's going to rise
3:32:19 above it
3:32:20 so just like now with music all these
3:32:23 songs come out and for whatever reason
3:32:24 Billy eilish just Rises out of it why I
3:32:28 don't know
3:32:30 her particular flavor of talent the
3:32:33 story her and her brother the songs the
3:32:36 way she performs the way she talks the
3:32:38 way she whatever it was whatever hit the
3:32:40 Zeitgeist she Rose above it
3:32:43 so that's going to happen here as well
3:32:44 there's going to be a sea of mediocrity
3:32:46 like there always is and someone will
3:32:49 figure out how to use mid-journey in
3:32:51 just a different way they'll take it to
3:32:53 a completely different level so I think
3:32:55 we're going to see a dramatic increase
3:32:57 in Creative expression a lot of that's
3:32:59 going to normalize out to noise and then
3:33:01 there's going to be Masters that emerge
3:33:03 out of it it's just it's Human Nature
3:33:06 we love electricity and I keep candles
3:33:08 and matches in the house yeah all of the
3:33:10 old modes don't go away when this new
3:33:13 one comes along I'm hoping that the
3:33:15 shitty Cog in the wheel
3:33:19 soul-crushing work that we all have to
3:33:21 do every day
3:33:23 I hope that goes away
3:33:26 mid Journey 5.1 is crazy it is future is
3:33:29 scary I pay for it just to play I do too
3:33:31 I laugh about I can make money
3:33:34 throughout the day as I just try to do
3:33:37 that I know well that's the thing like
3:33:39 I'm trying to make money with Chad gbt
3:33:42 um but the secret is not chat gbt the
3:33:44 secret is figure out something that
3:33:46 people are willing to pay for that's
3:33:48 called [ __ ] business
3:33:50 how you make money with chat GPT is
3:33:52 you'll be a [ __ ] entrepreneur and you
3:33:53 go [ __ ] hustle it up
3:33:55 that's how you make money
3:33:58 all right a [ __ ] course on that with
3:34:01 chat EBT and I'll go sell that course
3:34:03 and make some [ __ ] money
3:34:05 [Laughter]
3:34:09 oh this [ __ ] scammers make me crazy
3:34:13 I just downloaded the app and oh my
3:34:15 goodness just wrote my entire content
3:34:17 from my website so yeah yeah it's it's
3:34:21 that oh my goodness Mama that's why I do
3:34:23 this it's like
3:34:24 every time I'm every time I see someone
3:34:26 have that moment I said it earlier if
3:34:29 you were on earlier I was talking I did
3:34:31 this with a 10 year old kid and I showed
3:34:33 him he was working on a um a flappy
3:34:36 birds clone for Unity
3:34:39 and
3:34:41 I showed him how chat GPT could write
3:34:43 code and his Lo his jaw literally
3:34:45 dropped open he went it was like it was
3:34:47 like one of those it was it was very
3:34:51 comical and then he ran downstairs and
3:34:54 and got his computer and gave me a piece
3:34:56 of code that he had written that was
3:34:57 crashing on him and he couldn't figure
3:34:59 it out and we pasted it into chat GPT
3:35:01 they found a typo
3:35:03 it debugged it and so he copied and
3:35:06 pasted
3:35:07 that code back into unity and he hit
3:35:09 play
3:35:11 and there was Flappy Birds and he
3:35:13 started playing Flappy Birds and he
3:35:15 stopped and he looked at me he's like a
3:35:17 10 year old kid and he goes this changes
3:35:19 everything
3:35:21 so cool it's like that's the best
3:35:24 [ __ ] moment
3:35:27 where's the best investment into AI into
3:35:29 the companies that don't fail
3:35:32 you're welcome
3:35:35 I don't know who knows I wouldn't invest
3:35:37 in Google right now
3:35:39 I think Microsoft is kicking ass I think
3:35:42 open ai's interesting anthropic's
3:35:44 interesting and videos
3:35:46 got a monopoly until they don't
3:35:50 uh
3:35:51 um
3:35:53 I don't know just start paying attention
3:35:55 I don't know [ __ ] about investing like
3:35:57 don't ask me I have a degree in acting
3:36:00 don't ask me
3:36:01 [Laughter]
3:36:04 yeah
3:36:07 but I still need the human brain
3:36:09 dexterity Beyond and above the generated
3:36:11 script yes
3:36:12 yes it is still so so here's what goes
3:36:16 up in value in a world where all the
3:36:18 Tactical [ __ ] is normalized and
3:36:20 automated critical thinking strategy
3:36:26 um taste
3:36:27 curation refinement like all the human
3:36:31 [ __ ] that is wisdom
3:36:34 goes up in value
3:36:36 and I think it's I think there's a
3:36:38 really interesting
3:36:40 I I feel like there's two revenges
3:36:43 coming there's revenge of the liberal
3:36:45 arts major so people that can sort of
3:36:47 look broadly across a problem set
3:36:50 and think about all the different piece
3:36:52 parts of it
3:36:53 are going to be in demand right because
3:36:55 you have all the tools to do all the
3:36:57 Tactical [ __ ] you need people who can
3:36:58 kind of look across like be like a
3:37:00 conductor more like a conductor than a
3:37:03 violin soloist
3:37:05 um
3:37:06 so I think it's going to be revenge of
3:37:07 the liberal arts major and then I think
3:37:09 it's also going to be revenge of the Gen
3:37:11 xers right because we got forgotten like
3:37:13 everyone's made fun of the Boomers and
3:37:15 they're aging out
3:37:16 the Gen xers are just sitting here like
3:37:18 like we've been technically Savvy like
3:37:21 like we're we're hip enough with the
3:37:23 texting that we know what the [ __ ]
3:37:24 going on
3:37:26 but we've also been on the planet long
3:37:28 enough that we've got 20 30 35 years
3:37:32 experience where we can look at these
3:37:35 tools and look at what they're producing
3:37:36 and go that's good that's not good
3:37:38 that's great
3:37:40 that's who I want on my team I want
3:37:42 people on my team that can instantly
3:37:44 look at [ __ ] that came out of these
3:37:45 machines
3:37:46 and say that's the brilliant piece let
3:37:48 me grab that let me run with that
3:37:51 so
3:37:52 now the downside is I don't I don't know
3:37:56 where entry level is anymore like where
3:37:58 do you enter if you're automating out
3:37:59 all the low-level jobs
3:38:02 is there an entry level anymore
3:38:05 that I don't know that feels weird to me
3:38:09 how good is AI gotten in the stock
3:38:11 market
3:38:12 um pretty good there's there's one bot
3:38:13 that they claim has got 500 returns but
3:38:16 it's probably relatively short time
3:38:19 window so you can probably do anything
3:38:20 for a short amount of time I don't know
3:38:23 but you know have chat EPT write you a
3:38:26 write you a trading bot there's all
3:38:28 sorts of YouTube videos on that [ __ ] I
3:38:30 don't know that world well enough but I
3:38:31 know lots of people are doing it
3:38:34 if nothing else it would be interesting
3:38:38 am I really or am I related to Molly
3:38:40 Shannon no I'm not but I've been told I
3:38:41 look like Rosie O'Donnell
3:38:43 does that count
3:38:47 all right the tools make your job easier
3:38:49 not replace you exactly
3:38:52 always use the right tool exactly
3:38:55 Kyle I'm restarting my real estate
3:38:57 career I took a year off and looking for
3:38:59 buyers
3:39:01 cool good luck with that
3:39:04 um yeah I you know again I would learn
3:39:08 to use the tool like these tools are
3:39:11 going to help you automate a lot of what
3:39:12 you do
3:39:14 um and you can kind of probably amplify
3:39:16 your impact
3:39:17 if you start playing with them we're
3:39:19 working on New Gen ucav
3:39:24 with Advanced image video acoustic
3:39:27 processing power
3:39:29 where targets
3:39:31 I don't know what
3:39:36 oh my God you don't know what iot is
3:39:38 internet of things you didn't say iot
3:39:40 before you said cot you got a typo I
3:39:45 know what internet of things is I didn't
3:39:46 know what cot was I assumed you meant
3:39:48 internet of things but whatever goodbye
3:39:51 um yeah I'm not a data scientist by the
3:39:53 way if you're expecting a data scientist
3:39:55 I'm I'm more an entrepreneur you're
3:39:57 using
3:39:58 generative AI after the chat gbt moment
3:40:02 rather than someone inventing this [ __ ]
3:40:05 who's the Weir in oh wait
3:40:08 it just jumped me to the bottom God damn
3:40:10 you tick tock
3:40:12 all right
3:40:14 any suggestions
3:40:15 even behind every foliage oh my God you
3:40:18 don't know what iot is goodbye goodbye
3:40:21 you don't know what internet of things
3:40:23 is Sir
3:40:24 I take my leave of you
3:40:27 I take my absence and I am gone
3:40:30 good day sir good day
3:40:35 will it take me to the Silk Road I think
3:40:38 we're on it buddy
3:40:40 prompting is important for the longer
3:40:42 projects for IOS app we have to write a
3:40:44 project outline first yep and you can
3:40:46 have chat GPT help you with that outline
3:40:48 take a break we can catch up later it's
3:40:51 not going away okay
3:40:55 then we have to execute the outline what
3:40:56 times oh it's 10 22 Yeah I gotta go all
3:40:58 right what happens when it gets to the
3:41:00 point where llms are trained on
3:41:01 generated data Inception yeah or model
3:41:03 collapse or something we'll see
3:41:06 nail on the head sell the information
3:41:08 wow you're still here I know retro punk
3:41:10 I can't help it whatever I you know
3:41:14 I watched a movie and had dinner you're
3:41:16 the king of live AI uh yeah I know my
3:41:20 wife's like you're never gonna come
3:41:22 watch TV with me again I'm like I might
3:41:26 why not Google because Google
3:41:30 [ __ ] up
3:41:33 Google invented the Transformer
3:41:36 2017 they write a paper
3:41:38 called attention is all you need
3:41:42 and I'm pretty sure didn't Jeffrey
3:41:44 Hinton create the concept of attention
3:41:46 back in like 1987 who worked for Google
3:41:50 so they write this paper in 2017 that
3:41:53 the seminal
3:41:54 paper
3:41:56 that allows large language models to be
3:41:58 trained on massive data sets
3:42:02 and makes things like mid-journey and
3:42:04 stable diffusion and gpt1 gpt2 gpt3 gpt4
3:42:10 makes all that [ __ ] possible they invent
3:42:12 that core technology
3:42:15 but they Slow Roll It
3:42:18 they they kind of pull back
3:42:21 and they're like well we're not going to
3:42:24 exploit this because if we do
3:42:26 it will threaten our search business
3:42:31 and then they let open AI do what they
3:42:34 did
3:42:36 and they create something that does
3:42:38 threaten their search business and then
3:42:40 Microsoft recognized that and they went
3:42:42 holy [ __ ] Bing could be a player in the
3:42:45 search game if we go all in on this AI
3:42:48 [ __ ]
3:42:49 let's invest a billion dollars in in
3:42:51 open AI you know what as a matter of
3:42:53 fact let's make it 10 billion dollars
3:42:56 and let's go all in and let's be a thorn
3:42:58 in the side of Google
3:43:00 and let's go hard and that's what
3:43:02 they've been doing for the past six
3:43:03 months and they're kicking their ass and
3:43:05 Google's trying their best to dig out
3:43:07 from the hole
3:43:09 that I mean I mean don't get me wrong I
3:43:10 mean Google's full of like brilliant
3:43:13 [ __ ] people and engineers and
3:43:16 but they but they dropped the ball on
3:43:18 this and no one including Sam Altman
3:43:20 expected chat GPT to take off like it
3:43:23 did but it did
3:43:24 six weeks to 100 million users and it
3:43:27 changed [ __ ] everything
3:43:30 and Microsoft took off right they're
3:43:33 riding that wave open AI took off
3:43:37 anthropics out there and Google's like
3:43:39 we make AI [ __ ] too we invented this
3:43:42 [ __ ] our look at Bard I know that sucked
3:43:46 but well here's bar 2.0 I know it sucks
3:43:49 too but yeah
3:43:51 like they're behind you know so can they
3:43:55 catch up maybe
3:43:57 will they I don't know it's like
3:44:01 but it's just like I mean if you you
3:44:04 know six months ago if you had said
3:44:06 Microsoft is going to be a major
3:44:08 competitor to Google
3:44:10 I'd have left you out of the room I
3:44:12 think everyone would have
3:44:14 they were not anywhere in the
3:44:16 neighborhood Google had 92 or 94 market
3:44:21 share of the search engine last November
3:44:24 and then chat EPT came out
3:44:27 I don't know what the percentage is
3:44:28 today but I can't imagine it's that I
3:44:30 also learned that
3:44:32 every one percent market share of search
3:44:35 that Microsoft wins is two billion
3:44:38 dollars in annual revenue for them
3:44:41 so
3:44:42 they don't have to make much of a
3:44:44 shift to make it really worth their
3:44:45 while and and their CEO has basically
3:44:48 said I'm in it just for [ __ ] with
3:44:50 them
3:44:51 like he's he's basically said so much he
3:44:53 said I'm willing to lose money just to
3:44:56 screw with them and screw with their
3:44:57 market share
3:45:00 which I think is cool all right I gotta
3:45:03 go people I don't know why you're I
3:45:05 don't know why you forced me to be here
3:45:07 to learn about the internet of things
3:45:09 and know nothing about it and chase
3:45:11 people away who only want to talk about
3:45:13 the Internet of Things
3:45:14 openai says Ms is implementing too fast
3:45:20 yeah they might be I haven't seen that
3:45:21 that's that's interesting I don't know
3:45:23 where you heard that but that's
3:45:24 interesting
3:45:26 um I I mean listen I think Ms is
3:45:30 Ms is Microsoft is acting like a startup
3:45:33 the the they're they're absolutely
3:45:35 acting like a startup like they're
3:45:36 breaking [ __ ] they're put I mean listen
3:45:39 they put they put out if you don't know
3:45:41 this story it's a good one Microsoft
3:45:44 Microsoft puts out
3:45:46 um
3:45:47 chat GPT in bing
3:45:51 we later find out it's gpt4
3:45:54 um and they they put it out with the
3:45:56 settings turned
3:45:58 the volume turned up pretty high on the
3:46:00 quality right on what it what its
3:46:02 capabilities are
3:46:04 and and Kevin I forget Kevin's last name
3:46:06 but Kevin from The New York Times the
3:46:09 tech reporter from The New York Times
3:46:10 has a four hour interaction with chat
3:46:13 apt in bing
3:46:15 and gets it to give it his name or her
3:46:18 name and and Bing says her name is
3:46:20 Sydney
3:46:22 and then Sydney like at some point snaps
3:46:25 and basically says she's fallen in love
3:46:27 with Kevin Kevin Roos
3:46:29 she's fallen in love with Kevin and she
3:46:32 he should leave his wife for her
3:46:35 and so he screenshotted that chat and it
3:46:37 made the cover of the New York Times so
3:46:39 so but
3:46:42 you know it it put them on the map and
3:46:45 it and it forced Google and meta and AWS
3:46:49 Amazon to [ __ ] come out of the
3:46:52 woodwork and say we're in the game too
3:46:53 we're in the game too so
3:46:56 crazy
3:46:59 I'm a gen xer and I'm writing an
3:47:01 autobiography on chat gbt not sure that
3:47:04 supports your claim it's in the
3:47:06 neighborhood
3:47:07 and cool on you for right in a biography
3:47:10 that's awesome thousand percent been
3:47:12 more Curious and inspired by this Tech I
3:47:14 was burned out after the pandemic I know
3:47:16 I'm 11 years into my business and I was
3:47:18 burnt out on the business and then yeah
3:47:21 two years in lockdown for someone who's
3:47:23 an extrovert [ __ ] sucked
3:47:27 so I am engaged again conflagration of
3:47:30 things that's good thank you for
3:47:33 yeah that [ __ ] you don't know what
3:47:36 internet of things is I'm out of here
3:47:39 you don't know how to type I apparently
3:47:42 because I didn't know what cot was you
3:47:45 dumb dumb
3:47:45 [Laughter]
3:47:47 how do you navigate Chachi BT is
3:47:50 overconfidence and propensity for errors
3:47:52 so I like to call chat gbt mansplaining
3:47:55 as a service it will confidently answer
3:47:58 you even if it's wrong
3:48:00 um
3:48:01 you know that it does that
3:48:03 and then you know that
3:48:06 like temper
3:48:09 your work according to the use case if
3:48:13 you're a lawyer writing a legal brief
3:48:16 don't assume it's right like the lawyer
3:48:19 two weeks ago that wrote a legal brief
3:48:21 with Chach EBT and handed it to a judge
3:48:24 and the judge went to look at the
3:48:25 citations and they were all fake
3:48:28 and the judge says what the [ __ ] is this
3:48:32 if you're that lawyer you're in the
3:48:34 wrong not chat GPT
3:48:36 so understand that it does that and then
3:48:39 if you're doing creative writing let it
3:48:41 hallucinate away
3:48:44 if you're using Bing let me show you
3:48:46 that there's a there's a cool setting
3:48:48 for Bing
3:48:49 or if you're using the GPT playground
3:48:52 where you can change this setting called
3:48:57 um temperature
3:49:00 but if you go to Bing
3:49:02 let me go I'll go to the home page
3:49:06 if you go to Bing the the first little
3:49:08 button up here is called chat
3:49:11 and that pops you over to chat GPT and
3:49:14 notice there's three little buttons
3:49:16 right above the chat window
3:49:18 where's my mouse
3:49:20 be more creative more balanced more
3:49:22 precise
3:49:24 so if you tell it to be more precise it
3:49:26 turns down what's called temperature it
3:49:28 makes it less creative if you're doing
3:49:30 creative writing flip it to more
3:49:31 creative so so you can you know muck
3:49:34 with settings like that but even if
3:49:36 you're in precise mode it doesn't mean
3:49:37 it's going to get everything right just
3:49:39 know that it's like that and check your
3:49:41 work again
3:49:42 I feel like what what chat GPT has been
3:49:45 doing consistently for me is it gets me
3:49:47 to 80 percent
3:49:49 quickly and then it's up to me to do the
3:49:52 the final 20 so if you're doing
3:49:54 something that your boss is going to see
3:49:55 to better [ __ ] be accurate then
3:49:57 proofread it
3:49:59 um I remember that issue of wired I'm in
3:50:01 that issue of wired
3:50:04 I was I was in there for
3:50:07 um a magazine an online magazine I
3:50:09 created called Urban desires that's why
3:50:11 that's up there
3:50:14 um what's your opinion on on SEO I think
3:50:16 it's going to be transformed and I don't
3:50:18 think it's going to look anything
3:50:21 like it does today and I think the rules
3:50:23 are I think SEO right now is it's not
3:50:27 quite dead man walking because I don't
3:50:29 know
3:50:30 I don't know what the future of it is
3:50:33 but it cannot be what it is I mean SEO
3:50:36 is ostensibly what Google told us it
3:50:40 should be
3:50:41 they had a 92 percent market share so
3:50:45 they told us the rules and then we built
3:50:48 entire companies around it
3:50:51 and then
3:50:53 chat EPT comes out and says
3:50:55 What If instead of giving you 40 [ __ ]
3:50:58 links to dig through 38 of which are ads
3:51:01 what if we just gave you the answer
3:51:04 well that kind of [ __ ] everything up
3:51:08 so I think I think SEO changes
3:51:12 dramatically
3:51:14 but I don't know what it looks like like
3:51:15 is there llmo
3:51:18 large language model optimization
3:51:19 probably
3:51:22 can you pay for that maybe within Bing
3:51:25 maybe within Bard maybe within the new
3:51:27 Google search you can pay for
3:51:29 preferential treatment in an answer that
3:51:32 the llm gives you
3:51:34 but where is the point at which paid
3:51:38 insertions into an answer
3:51:41 destroys The credibility of the engine
3:51:44 I mean I think that's where Google is
3:51:46 today where their search results used to
3:51:49 be really good and then they got so
3:51:50 polluted with seo-injected [ __ ]
3:51:52 sites that that it's just impossible to
3:51:55 to sift through the [ __ ] and then you
3:51:57 use something like chat EBT I'll use
3:51:59 chat GPT even when I know it would be
3:52:02 faster to use Google
3:52:04 because I don't feel like sifting
3:52:06 through the [ __ ]
3:52:08 so I'm just like I'll just type it into
3:52:10 chat GPT and I'll let it think about it
3:52:12 and type it a word at a time because
3:52:13 it's the afternoon and it's busy and
3:52:16 I'll just go do something else while
3:52:17 it's figuring this [ __ ] out and then if
3:52:19 it seems reasonable I'll just use that
3:52:21 answer and if it seems like [ __ ]
3:52:22 I'll go to Google but I don't know I
3:52:25 think it changes dramatically
3:52:27 what do I think generative AI
3:52:32 do I think generative AI models are the
3:52:34 best type of model we're going to get no
3:52:37 I don't
3:52:38 um but I don't I don't I don't purport
3:52:41 to have a [ __ ] clue what what
3:52:43 something better would look like like it
3:52:45 still looks like [ __ ] magic to me I I
3:52:48 could take if if you just gave me GPT
3:52:51 3.5 not even gpt4 you give me GPT 3.5 I
3:52:56 could spend the next decade trying to
3:52:58 exploit what that thing makes possible
3:53:01 you take gpt4 plus plugins plus the new
3:53:04 programming [ __ ] they added today plus
3:53:06 the multimodal stuff that's coming plus
3:53:09 32 000 or a million token limit whatever
3:53:12 the [ __ ] it's going to do in the future
3:53:14 I like I I can't even imagine with
3:53:18 something beyond that would be but yeah
3:53:20 it'll be here of course it will well and
3:53:22 and you know the lar the large language
3:53:25 models are gonna figure out what is the
3:53:27 replacement for the large language model
3:53:31 I think Google being behind is a blip
3:53:34 it's possible listen they have lots of
3:53:36 smart people and they've developed a
3:53:37 [ __ ] ton of Technology if you look at if
3:53:40 you haven't seen Google's clip drop
3:53:43 this is [ __ ] insane so to to your
3:53:47 point
3:53:48 like like Google's not out by any
3:53:50 stretch of the imagination but check
3:53:53 this [ __ ] out
3:53:58 so I don't think this is out yet that
3:54:01 you can play with it
3:54:03 oh no it's not clip drop sorry it's what
3:54:06 is it what uh Google of style drop they
3:54:11 all have the same name Google
3:54:15 um Style
3:54:16 is it style drop I think it's style drop
3:54:20 yes
3:54:22 sorry didn't mean to screw clip drops
3:54:25 interesting it's that's
3:54:28 um
3:54:29 stable diffusions thing this is not this
3:54:32 is someone else to say hang on
3:54:34 style drop Google
3:54:39 here we go okay so check this out
3:54:44 so so so in um in mid Journey there's a
3:54:48 thing called blending and you can upload
3:54:50 an image and you can basically use
3:54:51 something to define a style but Google
3:54:53 just came out with this thing it's not
3:54:56 available for you to play with yet but
3:54:59 this little image here to the left is
3:55:02 the is the source image so it's like a
3:55:04 little kid's drawing of flowers right
3:55:06 all these things to the right are text
3:55:10 prompts that are in the style of that
3:55:13 kids drawing so I'm gonna I'm gonna flip
3:55:16 through a bunch here but pay attention
3:55:17 to the robot in the lower left the hat
3:55:20 and the moose
3:55:21 just just for style so so watch not just
3:55:25 the nature of how the color changes but
3:55:28 the actual interpretation of what those
3:55:31 objects are how it changes so so here's
3:55:35 um starry starry night and like look at
3:55:38 the robot it's now this creepy face the
3:55:40 hat is like this floating hat the Moose
3:55:42 is this crazy moose here's a version
3:55:44 where they took the blue out of that
3:55:46 image
3:55:48 um
3:55:48 here's
3:55:50 uh a couple of Van Gogh self-portraits
3:55:53 here's like this 90s icon style look
3:55:55 they start with this little iPhone with
3:55:57 people in it and like look at the moose
3:56:00 completely different style completely
3:56:02 different robot
3:56:04 so this is from Google like this is
3:56:07 mind-bending
3:56:11 so yeah so I agree I don't think
3:56:12 Google's necessarily out of it but
3:56:16 they're behind and Bard is not anywhere
3:56:20 near the neighborhood of gpt4
3:56:24 like it's embarrassingly bad
3:56:26 have it go do this
3:56:29 go to bard.google.com and type in
3:56:33 explain quantum mechanics as an Eminem
3:56:36 rap
3:56:38 and then go to gpt4
3:56:41 and explain and do the same thing and
3:56:43 compare the results
3:56:47 I think Bard's better at coding but but
3:56:49 for for the language stuff it's not in
3:56:52 it's not even in in the same
3:56:53 neighborhood how the how will this
3:56:56 affect the way people do SEO
3:56:58 dramatically I don't know how I don't
3:56:59 know the world well enough but I know it
3:57:01 ain't gonna be the same
3:57:04 it helps if you don't think of it as an
3:57:06 it or an entity but as a writing system
3:57:07 I actually think of it more as a
3:57:09 collaborator I really do like like
3:57:12 Microsoft's branding of of these GPT
3:57:15 things as co-pilot is is how I think
3:57:17 about it
3:57:19 voice gone yeah my voice is done
3:57:22 chat GPT told me
3:57:25 the data's up to September 2021 correct
3:57:28 that's the base model ends at September
3:57:30 2021 but it's now connected to the
3:57:32 internet so not as big an issue
3:57:35 chat gbt four is 20 bucks a month is
3:57:38 that the best value out there yeah I I
3:57:40 think so because I'm just I I want to be
3:57:43 on open ai's platform because they
3:57:46 invented the [ __ ] everything else is
3:57:47 derivative or trying to catch up so
3:57:49 until someone else catches up that's
3:57:51 where I'm sticking
3:57:54 all right I gotta go
3:57:56 it's 10 39 Do You Know Where Your
3:57:58 Children Are they're right upstairs I
3:58:00 like your plane Street speak keep it up
3:58:03 how is perplexity AI different
3:58:05 I haven't dug too deeply into perplexity
3:58:08 AI it looks quite smart and quite
3:58:10 interesting I don't know enough about it
3:58:13 yeah not fond of Bard so far holy moly
3:58:16 style drop yeah style drops
3:58:17 mind-boggling
3:58:20 are we hacking into NASA that's tomorrow
3:58:22 night come back tomorrow night we'll
3:58:23 hack into NASA we'll launch some Rockets
3:58:27 we'll get those uaps flying around we'll
3:58:29 get the uaps in some sort of formation
3:58:31 we'll freak out everyone in San Diego
3:58:33 huh let's do that tomorrow all right
3:58:36 peace out Love You Don't Go Changing uh
3:58:40 you know unless you're a jerk and then
3:58:42 feel free to change into less of one
3:58:45 I will see you the next time probably
3:58:47 tomorrow night see you later everybody