aiLL 8/6/23 - Unlocking The Magic of ChatGPT and The AI Revolution

Description

Chapters

Transcript

0:00 foreign
0:11 [Music]
0:20 welcome to the AI learning lab I'm just
0:23 getting myself together
0:26 I had a day out on the town
0:30 I got myself a new iPhone so it'll be
0:33 interesting to see hey I know the guy
0:35 hey retro Punk what's happening gonna be
0:37 interesting to see if the cameras are
0:39 any better because I have a new phony
0:41 phone
0:44 and
0:46 what else what else hey Silver Fox
0:50 what's happening
0:52 brand new spanking new charging wire for
0:55 my new spanking new phone
0:58 look camera better maybe no not really
1:01 but that one looks like it's more in
1:03 Focus maybe we're not gonna have that
1:04 [ __ ] Focus problem anymore that
1:07 looks better
1:08 that looks more gooder I got the 14 pro
1:13 three big old cameras on the back
1:15 they're Giant
1:20 I also got
1:22 hey young brother
1:23 yeah speaking of that
1:26 um I go into the it's the a t store and
1:30 the woman there is like she goes you
1:32 with a an ARP are you an AARP Member
1:35 so I got an ARP discount
1:40 and I got a new phone I traded in my old
1:44 phone
1:45 got a new phone and an Apple Watch and
1:48 I'm paying ten dollars less a month so
1:50 it pays to be old ah
1:54 I have the 13 and the I hate the camera
1:56 it applies filters automatically pretty
1:58 sure you can turn that off Silver Fox
2:00 but because I know I just like one of
2:03 the first things I did was I selected a
2:04 photo style which I'm like I'm gonna
2:06 regret this because it's gonna apply
2:07 that to all the crap that I take and I
2:10 like doing my own
2:11 my own stuff but whatever whatever
2:14 welcome everybody my name is Kyle
2:16 Shannon this is the AI learning lab let
2:18 me pop these over here
2:22 I did some more work on yield YouTube
2:25 channel
2:26 and I used Claude to do some some
2:29 description writing oh see we still got
2:32 Focus issues
2:34 yeah that's a little better let me dim
2:35 that
2:37 there we go all right so
2:40 um
2:41 if you want to see
2:43 back recording back recordings
2:45 recordings of these lives I'm putting
2:47 them on the YouTube channel which is
2:50 AI learning lab Dash TT like Dash tick
2:53 tock on YouTube because uh someone else
2:56 has AI learning lab uh but if you go
2:59 there I've got
3:02 um recordings and it's kind of cool I
3:05 did um
3:07 so here's the channel
3:09 and then I did
3:12 um claw generated descriptions and
3:16 chapter headings
3:19 um so this is this is one for my
3:21 birthday but it's pretty cool so you
3:25 know writes up the description writes up
3:27 the chapter headings you know they
3:29 automatically get hyperlinked to that
3:31 section of the video
3:32 and then I guess it's going to recommend
3:34 oh here are the chapters you can scroll
3:36 through the chapters so I think this is
3:38 going to make these things actually
3:39 decently usable so if you uh
3:42 if you and if you're new here uh and and
3:45 enjoy tonight you're going to be able to
3:47 go back I have all of
3:49 um May done
3:51 and I've got June and July and now
3:55 August to do hopefully I can get caught
3:57 up in the next week or so
4:00 um and then I'll probably dump some
4:02 other stuff over there I'll dump over
4:04 the other tick tocks but but anyway it's
4:07 AI learning lab Dash TT
4:11 and
4:13 if you just can't get enough of me live
4:16 now you can get me on Memorex
4:18 [Laughter]
4:22 so that's not what this is about what's
4:24 this about the AI learning lab
4:27 whole idea here is to talk about
4:29 generative AI what it means where to get
4:32 started uh answer any questions you have
4:34 if you have questions pop them in the
4:36 comments below I do try to answer all
4:38 the questions it's it's a bit of a
4:40 running joke that I get so far behind
4:41 that people don't even think this is
4:42 live but it is see I if I couldn't pitch
4:46 myself if I wasn't live could I no I
4:49 don't think so
4:50 if you're brand new to generative Ai and
4:53 chat GPT and all the likes if you
4:54 haven't played with chat GPT that's the
4:57 first URL here now you may say Kyle why
5:00 do you have all of these sites up here
5:03 if I'm just trying to get started what
5:05 do I do just start with that first one
5:07 that's that's the official Chachi BT
5:09 site in the past two weeks a bunch of
5:13 others kind of showed up on the scene
5:14 that that are all worth they all do
5:16 something interesting
5:18 um so uh that's the uh the official it's
5:22 got this thing called code interpreter
5:24 which is amazing it's got this thing
5:26 called custom instructions which are
5:28 really interesting it's got plug-ins
5:30 it's the original it's the best Bing is
5:33 also gpt4 like chat gpt4 uh it's
5:37 connected to the internet that's
5:39 Microsoft's offering it can also
5:40 generate images and read images Claude
5:43 can you can put in really long
5:46 um prompts so when I when I grab my
5:49 transcripts from these YouTube videos
5:51 that I'm doing they're two some of them
5:53 are two and a half hours long because I
5:55 can't shut up when I do these lives
5:58 um and um you can just paste the whole
6:01 transcript in there and I'm writing all
6:03 the descriptions on YouTube using Claude
6:06 um Bard is no longer as embarrassing as
6:08 it was for Google so that's worth taking
6:10 a look at lab stop perplexity that's
6:13 actually uh meta's llama two models the
6:17 7 13 and 70 billion parameter models you
6:20 can just play with without installing
6:22 anything and then Pi dot AI is super
6:24 conversational and po.com is you can
6:27 play with lots of uh different models
6:30 and you can make Bots and it's kind of
6:32 interesting as well but if you're just
6:33 getting started just go to that one
6:35 there chat.openai.com
6:38 and if you don't really know what to do
6:40 in terms of prompting go to prompts.chat
6:43 so this is a document that will teach
6:45 you how to prompt all right and then
6:47 openart.ai will teach you how to prompt
6:49 image generation tools those are the
6:51 ones down there on the bottom that I
6:53 recommend for image generation and then
6:55 future PD is just a whole bunch of just
6:58 a big pile of tools if you want to geek
7:00 out
7:02 just know that most of the tools on
7:03 future pedias suck and that's just
7:06 because of where we are in the in the
7:08 industry right now we're very early
7:11 um and yeah that's the deal that is a
7:14 deal all right let's
7:17 let's have any questions about AI pop
7:19 them in the comments below I will jump
7:21 in here
7:22 grab a question or two and we will get
7:25 started
7:28 all right
7:30 AARP is awesome yeah you don't even have
7:33 to be that old to be in it I don't think
7:35 it I don't even know if there's any age
7:37 restrictions
7:38 but um but yeah it's pretty good pretty
7:41 good it was funny she was like are you a
7:44 member of AARP I'm like should I be
7:45 offended I'm like no I'm just an old
7:48 fart
7:49 um and then I looked at my wallet and
7:50 sure enough 2019.
7:54 I had my membership card got my 20 off
7:59 all right
8:01 getting started
8:04 thoughts on AI automation agencies
8:08 um yeah let me flip this back over flip
8:10 over to this wait not that this
8:14 let me make these big
8:18 yeah we call them save the day all right
8:26 AI automation agencies I think
8:30 that they are interesting
8:34 um
8:36 because
8:37 as people start using these tools
8:40 um there are certain things you want to
8:42 do over and over and over again like for
8:44 example
8:45 um
8:46 these descriptions of the YouTube things
8:49 that I'm
8:51 um
8:53 that I'm creating
8:54 right now I'm kind of copying and
8:57 pasting The Prompt and then I put it in
8:59 a new chat and then I put in the
9:00 transcript and then sometimes it screws
9:02 it up and I gotta
9:07 um so so if you if if you found
9:10 something that works and then you want
9:11 to do that over and over again
9:13 um getting it in some sort of automated
9:15 fashion is going to be good right over
9:18 time these tools will get more
9:20 sophisticated and we probably won't need
9:21 to do it but I think short term it's
9:24 it's really good
9:26 um
9:27 not everyone's going to want to learn
9:28 how to prompt not every even though the
9:31 the tools like zapier and make
9:34 um are no code and they they make it
9:37 seem simple it's still really
9:39 complicated you're still connecting
9:41 different apis together and you're still
9:44 creating logic trees and you're still
9:46 having to you know think about the
9:48 prompting so it's a complicated enough
9:52 um process that I I think it's it's
9:55 valuable
9:56 um I think that those agencies over time
9:59 are gonna have to get more and more
10:00 sophisticated right now you can kind of
10:02 just do it with text and and things like
10:05 that but I think I think over time
10:07 they'll probably have to get more
10:08 sophisticated and bring in other tools
10:11 um but yeah I think I think they are
10:15 they're probably a good way to
10:19 um to make money short term with the
10:21 large language models without having to
10:23 have a ton of expertise if you're good
10:24 with language and if you if you can
10:26 prompt these things and kind of
10:27 understand what works what doesn't which
10:29 of these are available for those
10:30 automation tools and you can do it more
10:33 power to you go charge 200 bucks an hour
10:35 or whatever it is and or charge a flat
10:37 fee and just knock that [ __ ] out quickly
10:40 um yeah I think they're good and and
10:42 also with um with zapier now having
10:44 interfaces you can build chat Bots and
10:46 you can do some fairly sophisticated
10:48 stuff
10:49 without even having to build a website
10:51 so it's pretty cool
10:54 so I like it it's good it's very very
10:58 good AI learning lab
11:01 jamai huh
11:03 what do you mean Captain T
11:05 I switched to Mint mobile cheap and
11:07 services fine costs a fraction compared
11:09 to Verizon yeah I was thinking about uh
11:13 switching to Mint mobile retro Punk and
11:15 then I just got the Willies that that
11:17 because they're sort of leasing
11:20 bandwidth from the big boys of the big
11:23 boys need it then I would assume mint
11:26 mobile gets hurt but um
11:28 yeah I might do that it's just dang
11:31 cheap although I signed a contract today
11:33 I'm probably in for two years with a t
11:35 [Laughter]
11:41 I like totals
11:46 oh
11:49 Tyron truth hello hello hello hello
11:52 Tobias what's shaking
11:56 um
12:00 best ways for me to use AI is a public
12:04 librarian wow
12:07 true
12:13 well
12:15 I think that
12:17 I think that you know libraries as as
12:20 access points you know knowledge access
12:22 points are super critical and if you
12:25 think about what large language models
12:27 are
12:28 you know they've been trained on all of
12:31 the public internet or as much of it as
12:32 they could get a lot of the public
12:34 internet
12:35 [Music]
12:36 um
12:37 and so in a lot of ways a large language
12:39 model is a library it's a very
12:44 um actionable library right it's it's
12:47 not just a place where you go to find
12:50 information and read it and turn that
12:52 into knowledge this this stuff is almost
12:55 like knowledge on demand
12:57 um so so I think just like libraries you
13:00 know kind of shifted from books to to
13:02 you know internet access points and and
13:05 kind of community centers around that
13:08 um I think they can serve a very similar
13:10 role
13:11 um with these large language models so I
13:13 would I would think
13:15 um
13:15 doing things like you know education uh
13:18 workshops on a regular basis I think
13:23 um helping understand
13:26 how
13:28 these tools work and how they're going
13:31 to impact
13:33 um
13:34 jobs could be a really interesting
13:39 um exploration within a library setting
13:41 you're gonna have a lot of people in the
13:43 community that are that are disrupted
13:45 and displaced over the next bunch of
13:48 years but the next two to three in
13:50 particular I think are going to be
13:51 pretty
13:52 pretty nasty and and pretty disruptive
13:55 and so I think I think a library could
13:57 be a place for people to re-upskill uh
14:01 their skills I know I'm not giving you
14:02 specific things you can do with these
14:04 models
14:06 um but but I I feel like as a just as a
14:09 role
14:10 um the libraries could be very very
14:12 important
14:13 um I think they're going to be very very
14:15 important
14:16 um because a lot of people a lot of
14:18 people are going to have to reinvent who
14:21 they are
14:22 and
14:25 re-discover not even ReDiscover they're
14:28 going to discover
14:35 how their how their skills are going to
14:37 be applied in a new world that's powered
14:39 by this stuff right
14:41 things are like ability amplifiers and
14:44 and so there might be someone who I'll
14:47 give you a good example in in the AI
14:49 Salon it's a community that I started we
14:51 have a guy in the salon
14:54 that was a computer uh programmer
14:57 engineer you know computer dude for for
14:59 years and years and years and and when
15:02 we first created the AI Salon we called
15:04 it the AI artist salon and and this is
15:07 Roger he's an older fella he's a little
15:08 you know he's he's a little uh
15:11 curmudgeony sometimes and and so in our
15:14 Discord server he changed his name to
15:16 Roger not an artist because we were the
15:18 AI art you know artist salon now we're
15:20 just the AI salon
15:22 um
15:23 but he started applying his
15:26 his engineering mind and thinking and
15:31 process
15:32 to mid-journey prompting and he ended up
15:35 getting really good at generating images
15:37 and now I think he's changed his mind to
15:39 or changed his name to Roger becoming an
15:43 artist and so I think I think there's
15:45 going to be a lot of stuff like that
15:47 where people
15:48 have skills that they're good at right
15:51 now
15:52 and with these tools they can take those
15:54 skills and apply them in a lot of
15:56 different disciplines
15:58 were they historically wouldn't wouldn't
16:00 have been able to do that right he's not
16:02 you know by Nature a creative guy but he
16:04 obviously appreciates art and the fact
16:07 that he went from sort of saying you
16:09 know haramf I'm not an artist to hey I'm
16:12 actually pretty good at this because I'm
16:14 good at thinking logically and how you
16:16 prompt these tools you have to sort of
16:18 logic your way to get the results that
16:21 you want
16:22 and so so there's another place where I
16:24 think you know Librarians and libraries
16:27 can serve is is understand the skills
16:31 that people in the community have and
16:33 the better you can understand what these
16:35 tools are what tools are available
16:38 um you know the the better you're going
16:40 to be able to serve people making those
16:42 Transitions and and going through those
16:44 reinventions
16:46 um
16:47 another thing that might be interesting
16:49 if you go to
16:51 um futurepedia.io
16:55 that's that's a a directory of like it's
16:59 up to like 4 000 tools right now like AI
17:03 tools so there's you know music tools
17:06 and writing tools and I don't know
17:08 whatever marketing tools and animation
17:11 tools and they're just growing and
17:14 growing and growing and growing
17:16 the amount of
17:20 innovations that are that that like
17:23 we're on an exponential Innovation curve
17:26 right now where these AIS are starting
17:29 to
17:30 like Engineers are using the AIS to make
17:33 other AIS better and and they're they're
17:35 sort of leap frogging one another
17:38 um and so there's just going to be a
17:42 staggering amount of Technology coming
17:44 at at us so so I think finding a way to
17:47 organize and understand what are the
17:49 tools that are out there so if someone
17:51 comes in the library and says hey you
17:53 know
17:54 you know I've been a musician for 30
17:55 years like I need to know about how I
17:58 make music with AI like what what are
17:59 the options well I don't know like like
18:01 let's go look right and you know here's
18:04 a bunch of the things and and some sort
18:06 of thing like future pedia that explains
18:08 what they are how they work
18:10 um I I think a librarian could have a
18:11 role in in categorizing that stuff
18:14 because again for the next two to three
18:16 years it's just gonna be just a deluge
18:20 of
18:22 tools and changes and Innovations and
18:26 and disruptions
18:28 um
18:30 so yeah that and then from a tactical
18:33 standpoint you know what I would say is
18:36 look at the stuff that you do on a
18:39 day-to-day basis as a librarian and you
18:41 know what's the stuff that is is stuff
18:45 that you do repetitively and it takes a
18:48 lot of time
18:49 that's the stuff that I would
18:52 but is also important right so what's
18:54 what's stuff that's repetitive time
18:56 consuming and important
18:58 and that's where I would go to things
19:01 like chat GPT or Claude and start
19:04 exploring how can I use chat CPT to to
19:07 do that stuff faster or better
19:11 um you know get more efficient with it
19:13 or or or you know just do it in a
19:15 different way
19:17 um you can start to explore things like
19:18 automations automate little processes
19:21 within the library and and the more that
19:23 you learn about doing that for the
19:26 things you do on a day-to-day basis
19:28 you're going to be able to help people
19:29 as they come in and say hey how can I
19:31 use these tools right so again back to
19:33 that idea of everyone right now everyone
19:38 um is is trying to figure this [ __ ] out
19:39 like like I did the uh
19:42 yesterday
19:44 um I launched the YouTube channel and
19:47 and live on the on on the live yesterday
19:50 I tried to use chat to write
19:54 um
19:55 to write the uh the description of the
19:58 YouTube channel and it didn't go well
19:59 because it couldn't read it and the
20:02 plug-ins didn't work it was just it was
20:03 just crappy and then I was about to give
20:06 up and I tried Claude and it worked okay
20:08 it worked pretty good and then today I
20:09 worked with that and so the the
20:11 discovery of which of these tools are
20:14 good which ones are not good
20:16 um
20:16 I think it's all stuff that you can do
20:19 within the library so I hope that that
20:20 helps
20:21 um
20:22 if you're new here my name is Kyle
20:24 Shannon this is the AI learning lab
20:25 these are a bunch of tools you can go
20:27 play with if you're brand new to uh chat
20:29 GPT go to that top one and uh just start
20:32 playing while you listen to me ramble
20:34 um if you'd be so kind share the share
20:36 the lives so other people can hear about
20:38 it and double tap the screen let's get
20:40 some hearts going so you know it looks
20:42 like you're enjoying yourselves happy
20:43 Sunday night by the way we got we got we
20:46 got work tomorrow
20:48 everybody excited for Monday
20:53 all right thank you thank you thank you
20:55 appreciate that can I use AI to write
20:57 code for web scraping and data found on
21:00 Government website listing finds yes you
21:03 can
21:05 um
21:06 you're going to want to use one of these
21:08 that is connected to the internet so the
21:12 two that are that are you know probably
21:15 decent at what you're talking about are
21:17 Bing and Bard so bing.com is connected
21:21 to the internet and bard.google.com is
21:23 connected to the internet
21:26 um
21:27 they may or may not
21:30 allow you to do web scraping because
21:33 they have safety guard rails built into
21:35 them so they might say as a large
21:37 language model I can't be expected to go
21:41 do whatever I do if it somehow senses
21:43 that it's that it's uh you know
21:45 something out of bounds from a safety
21:47 standpoint uh but I would go try it
21:50 um the other thing that you can play
21:51 with if you
21:54 um
21:54 if you're decent with code
21:56 um go to Google uh go to GitHub and use
21:59 their collab sort of pair programming
22:02 tool which is also chat it's chatbt
22:05 optimized for coding
22:07 um
22:10 and then the other thing you can do is
22:12 just go to Bard and Bing you don't even
22:14 necessarily need to write code to scrape
22:17 it you can potentially just say what are
22:20 some websites that have uh
22:24 you know government listings of fines or
22:26 whatever it is you're looking for
22:28 um and it will go find them for you and
22:30 then see if it can actually you know ask
22:32 it to go grab data for you
22:35 um the other thing that you can play
22:37 with is you could play with code
22:39 interpreter at
22:42 um
22:42 chat GPT
22:44 so I'll show you how to do that let me
22:46 show you how to do that now because I
22:48 think this could be an interesting
22:49 option
22:56 okay
22:58 so here we are at chat gbt yeah I think
23:00 this new camera is actually better than
23:02 the old one was
23:03 I do believe it's true
23:06 all right
23:08 I'm probably just telling myself that I
23:10 got a new shiny phone it's better I
23:12 promise these sneakers make me run
23:14 faster
23:17 could be one of those okay
23:19 so we're at chat GPT
23:21 um
23:23 if you want to play with this tool code
23:25 interpreter you do have to be a member
23:27 of chat CPT Plus or not a member a
23:29 subscriber to which is 20 bucks
23:32 um I do not make money off that so if
23:35 you're here to troll that's not what
23:36 this is about
23:38 um
23:39 when you go to gpt4 the option here if
23:42 you don't see code interpreter and
23:44 plugins it means you haven't turned them
23:45 on so how you turn those on
23:47 is
23:49 you go to your settings
23:52 and then within settings there's beta
23:55 features when you go to Beta features
23:57 and you actually have to enable them so
24:00 you're going to enable custom
24:01 instructions plugins and code
24:02 interpreter is there anything new no
24:06 all right so go turn those on
24:09 and then
24:11 um
24:14 and then what we're gonna do is we're
24:16 gonna go grab some fake data well it
24:19 might be real data but we're gonna go
24:21 grab it
24:22 I'm going to be completely ignorant
24:26 um about whatever data we're grabbing so
24:29 I'm going to go to a site called kaggle
24:30 k-a-g-g-l-e.com
24:39 and I'm going to go to data sets and
24:42 then you said government listing fines
24:46 I'll go I don't uh list I have a gov
24:51 earn
24:54 meant fines
24:59 gdpr fines Oklahoma court fines that
25:03 could be it
25:04 the kind of thing you're looking for
25:07 Chicago ordinance violations that's a
25:10 fun one all right let's go grab that one
25:11 how big is it
25:14 does it say 53 megabytes all right this
25:16 is pretty big okay so I'm going to
25:18 download this Chicago ordinance
25:20 violations okay
25:22 so
25:24 rather than having to scrape data if you
25:26 can go to sites that give you access to
25:29 data
25:31 um you can download that big old pile of
25:33 data go back to chat GPT we're going to
25:37 go to code interpreter now
25:39 and
25:42 when you go to code interpreter your
25:45 your little prompt it gets a little plus
25:47 icon in it
25:49 can you see that a little bit and so I'm
25:53 gonna hit the plus icon and I can upload
25:56 one file now it's a little deceptive
25:59 when I first heard you could upload one
26:01 file I'm like well that's stupid what if
26:03 I've got five files and I want to do
26:05 those do I have to do them one at a time
26:06 can I do one at a time
26:08 the the file that you upload can be a
26:11 zip file so you could have a zip file
26:13 full of csvs of a bunch of other data
26:16 okay so it's actually a lot more
26:18 powerful than I initially thought it was
26:20 okay and I think you can put lights in
26:23 there so it can be a decent amount of
26:24 data this one's 53 megabytes so the way
26:27 this works is you just ask it for what
26:31 you just ask it what it is okay
26:37 please
26:39 look at the file inside this archive
26:44 and
26:46 tell me
26:48 what it is and what might be interesting
26:52 about it
26:59 oh is that still uploading that's it's
27:01 going
27:03 okay and so then it says working and if
27:06 I say show work what's actually
27:08 happening here this is this is
27:11 mind-blowing to me
27:12 so if I say show the work
27:14 what it what it did was it recognized
27:17 file so
27:19 when you run code interpreter what it's
27:21 actually doing is it's launching a
27:23 virtual machine so it's like a virtual
27:24 computer
27:26 and then it starts writing and executing
27:28 its own python code to help you solve
27:31 what you want to solve so in order to
27:34 see what's in it it has to unzip it so
27:36 it Imports some libraries writes some
27:39 code and then finds out
27:42 a CSV and a Json file in there and then
27:45 it describes that the CSV
27:49 um
27:51 it describes what a CSV is then it
27:54 describes what a Json file is
27:56 and then it says let's examine the
27:58 contents of these files to better
28:00 understand it
28:01 it says ordinance violations is a data
28:05 set with 27 columns in it and so so it
28:08 makes us a table that we can actually
28:10 scroll across I've never actually seen
28:12 it do this this is pretty cool so so
28:15 it's it's showing us some sample data
28:17 and what's in it so we can get a sense
28:19 of of what are the fields in this data
28:22 which is cool
28:24 it says from a cursory glance it seems
28:26 like a data set that contains
28:27 information about ordinance violations
28:29 here are a few observations
28:32 Etro appears to represent a specific
28:34 violation there's also details about the
28:36 violation itself
28:37 now let's examine the contents of the
28:39 Json file
28:41 that appears to be metadata for the CSV
28:44 data set Name ID description category
28:47 tags all sorts of stuff in the Json file
28:51 all right
28:54 and it's it's just doing this on its own
28:57 right it's it's prompting itself to
29:00 analyze this data for us
29:03 okay the Json file is particularly
29:05 useful if you want to understand what
29:07 each column in the CSV file represents
29:12 uh let's say
29:14 um
29:15 is it possible for me to find out
29:20 which violations
29:24 have been paid and which ones have not
29:27 question mark
29:30 and again I don't I'm taking a wild
29:32 asking asset what you might want to do
29:33 with this so
29:35 to determine which ones have been paid
29:36 or not you would need a column of data
29:44 status of the violation
29:48 so so it's looking at the different
29:50 categories and saying which ones
29:53 might be relevant
29:56 the closest column group that might
29:58 provide some insights is violation
30:00 status but we would need to examine the
30:02 data would you like me to check
30:04 violation status column from the CSV
30:06 file yes please
30:18 so
30:21 you know I apologize for the confusion
30:24 that column does not appear to be
30:26 present in the data set well that's dumb
30:32 so it might not have the data in it
30:48 all right so so I think it looked in the
30:51 uh maybe it looked in the Json file and
30:53 it found stuff and then it realized that
30:55 wasn't the same as the CSV so I'm going
30:58 to say yes so so it went and looked at
31:01 the CSV it found some other columns that
31:03 might be interesting
31:07 and it has to be politely would you like
31:08 me to go look at those for you sir yes
31:11 yes I would
31:14 the case disposition
31:18 column has the following unique values
31:21 liable default non-suit continuance yeah
31:24 this seems like it's valuable judgment
31:26 vacated
31:28 judgment vacated dismissed without
31:30 prejudice
31:32 the imposed find column contains an
31:34 America and without further context or
31:36 additional data uh that's fine let's
31:39 just say
31:40 um
31:41 please analyze
31:45 the data
31:48 and
31:50 visualize
31:51 the things you think are most
31:56 relevant
31:59 and again I'm asking it really shitty
32:01 questions here if you knew what you were
32:02 doing you could be asking it smart
32:04 questions and probably getting some
32:06 interesting things but I want to show
32:07 this
32:08 I want to show what it what it will
32:11 also do so so it's
32:13 it's doing this kind of textual analysis
32:16 but it's also going to create us some
32:17 pretty charts and graphs
32:20 uh which of these analyzes let's see
32:22 Geographic top violation codes case
32:26 dispositions
32:29 violations over time let's do
32:35 let's do number one
32:38 all right
32:43 [Music]
32:47 all right so let's go look at what it's
32:49 doing it is calculating the descriptive
32:53 statistics for the imposed fine column
32:57 count the number of unique respondents
33:00 so it's writing code right now it's
33:01 going very slow it looks like gpt4 is a
33:05 little a little hammered right now but
33:07 it figured some [ __ ] out
33:11 there's 673 000 total rows with
33:15 non-missing values the average is a
33:17 thousand fifty five dollars and 32 cents
33:19 the standard deviation the 25th
33:22 percentile the median of 200 the 75th
33:25 percentile the maximum
33:28 unique number of respondents so it's
33:30 doing some analysis if there's other
33:31 okay so now I'm going to say please
33:36 visualize this in a
33:39 uh
33:44 uh visually
33:48 uh compelling and beautiful
33:54 way
33:56 and let's see what it comes up with to
33:57 visualize it for us and I put a lot of
34:00 typos in there
34:02 doesn't seem to care at all about typos
34:06 um it can it just figures [ __ ] out and
34:08 it'll it'll correct it'll correct you
34:10 without telling you you're a dum-dum it
34:14 just it just fixes it it understands it
34:17 it's pretty remarkable in that sense
34:23 finished working
34:24 so there's
34:26 two visualizations of the data
34:29 this is
34:31 distribution of imposed fines most of
34:34 them are really small
34:36 and very few of them exceed ten thousand
34:40 dollars
34:41 and then the top 10 violation codes and
34:44 there's numbers there
34:46 um
34:47 let's say
34:50 can you redo the chart with the codes on
34:54 it
34:54 to include the name of the code not just
34:58 the ID number
35:03 because we don't know what those numbers
35:04 mean but I'm sure there's a data column
35:06 in there that says what the name of that
35:09 violation is
35:15 so
35:16 yes it can write code and with code
35:19 interpreter it can it can write and
35:22 execute the code on your behalf
35:25 so that's really weird there's one of
35:27 the things is really long okay
35:30 [Music]
35:30 um
35:34 it appears that some of the names of the
35:38 violations are extremely long can you
35:40 please wrap the text
35:42 so it's easier to read and we don't have
35:44 such
35:46 long single lines
35:52 let's see if it can fix it
35:54 I'm it also surprises me how well it
35:57 understands like a lot of times I'm like
35:59 I don't quite know how to tell it it
36:01 sucks so I sort of describe it sort of
36:04 how I think I would say it but if I were
36:07 if I were talking to a person I would I
36:10 would have to think about it a lot more
36:12 but with chat EPT it just seems to
36:14 understand for the most part what's
36:16 going on it says apologies for the error
36:18 I forgot to import the wrap function
36:21 so now it it actually wrapped
36:25 the uh the description text
36:30 arrange for the inspection of the
36:32 premise so now these things are really
36:33 really long this is like the legal
36:35 description of it so I'm going to say
36:38 um
36:39 it looks like the descriptions are in
36:42 legalese Period can you please
36:46 summarize them as a human readable
36:51 summary of the violation rather than the
36:55 long description
36:56 in this chart
37:07 and and you know again
37:09 this is the kind of thing that
37:13 you know you would need a data analyst
37:15 to do
37:17 like even the kind of ridiculous things
37:20 that I'm asking it right now they're
37:21 ridiculously simple things to turn this
37:24 around from someone that is probably
37:27 busy and gets asked to do a lot would
37:30 probably be you know a couple of hours
37:33 to a week or two right especially if
37:35 you're working in government
37:38 um if you have some you know Young Turk
37:40 working for you it can be good but so
37:42 now we have a chart an updated chart
37:44 with short descriptions
37:47 so the most
37:51 the the the the highest count is for
37:54 post owners name and number on the
37:57 building exterior
37:58 people aren't labeling their buildings
38:01 Replace defective materials apart obtain
38:04 or post a required permit right so that
38:07 I could go in and I could say oh let's
38:09 let's go see who hasn't put their
38:10 permits in I don't you know I don't know
38:11 what you're doing here but
38:13 like I've already got ideas so anyway so
38:17 that's code interpreter so absolutely
38:19 freaking nuts so I hope that helps
38:21 what's your name
38:22 prot weight
38:25 Pro
38:28 something Midwestern Pro prolid
38:34 oh wait Freud I don't know
38:37 Pro yd Midwestern Maybe
38:40 anyway hope that helps
38:43 LK 99 do we know yet I don't know
38:46 there's all sorts of stuff going on with
38:49 LK 99
38:52 um
38:56 so
38:57 so there's already conspiracy theories
38:59 that that the oil companies and you know
39:03 the powers that be are trying to do
39:06 disinformation about LK 99 there's
39:09 weirdness with the researchers
39:12 um it does appear that two different
39:15 Laboratories have created like really
39:17 teeny tiny things that are displaying
39:21 um
39:23 um superconductive property so so for
39:25 those of you know that that don't know
39:27 LK 99 is this new room temperature
39:29 superconductive thing that if it's true
39:32 and if it happens it's like
39:35 our our future advances like 200 years
39:37 overnight
39:39 um like just absolute Insanity in terms
39:43 of the kinds of things we'll be able to
39:46 build and we'll be able to like make
39:48 Electronics smaller and
39:50 you know send energy very long distances
39:54 with no degradation and it changes
39:56 everything right and so
39:59 the oil companies may or may not want
40:00 that to be the case
40:02 so
40:04 I don't know so so I don't I I haven't
40:06 seen anything today that's like it's
40:08 definitive but it does look like thank
40:10 you very much for the hot hot hot it
40:13 does look very much like the uh
40:15 two Laboratories have replicated the
40:20 the uh the experiment so I don't know
40:24 I don't know I really [ __ ] hope so I
40:27 really hope so
40:28 all right best AI tool for book writing
40:31 um rygal Smith cool name rigel I like
40:35 that
40:37 um
40:39 so just the best GPT
40:45 um chat GPT tool is gpt4 so go to
40:49 chat.openai.com and pay the 20 bucks and
40:52 get access to their version of chat GPT
40:55 you can also use Bing which gives you
40:58 access to to gpt4 for free but Bings
41:03 safety guard rails are are
41:06 way too restrictive I think Bing's the
41:09 way Bing writes is just it's like a
41:11 completely different tool than the gpt4
41:14 at open AI
41:17 so if you just want a large language
41:18 model tool where you're going to be
41:22 um
41:23 doing all the prompting and things like
41:25 that definitely chat gpt4
41:29 there's a project out there called GPT
41:31 author which is really interesting not
41:34 perfect but if you're interested in book
41:36 writing it's fascinating
41:39 so the the challenge with it is you have
41:41 to be a little geeky if if you're good
41:43 and geeky rigel then and and you know
41:46 how to run a Google collab
41:50 um GPT author is a Google collab
41:53 notebook which basically means you run
41:56 Google collab you get it running and
41:57 then you open this notebook and it's got
41:59 these series of cells and you basically
42:02 just hit the play button which executes
42:03 a bunch of Python scripts in a row
42:06 and then but what what GPT author does
42:08 is you give it your open AI API key in
42:13 your stable diffusion API key and then
42:16 you give it a topic of your book and a
42:19 writing style and a number of chapters
42:21 and it writes an outline and then it
42:24 writes chapter one and then it writes
42:27 chapter two and up to the number of
42:29 chapters that you have and then at the
42:31 end it reads the whole book basically
42:33 and says
42:35 um what's the
42:37 um
42:39 what what should the cover art looks
42:41 like and it generates the cover art and
42:43 then it puts it in ebook format and lets
42:45 you download the ebook that it just
42:47 wrote it takes it like three minutes it
42:49 writes a book with as many chapters as
42:51 you want
42:52 that thing right now is uh it's set up
42:56 to do fiction like fantasy fiction kind
42:59 of kind of books
43:00 um but you can see all the prompts that
43:02 are in there so if you wanted to geek
43:03 out you could go in and you could change
43:05 the prompts of what it's doing and make
43:07 it right non-fiction books or you know
43:12 romance novels or whatever
43:14 um but just in terms of writing writing
43:16 writing you know that's one you might
43:18 also go to
43:21 um
43:22 go to
43:24 futurepedia.io and look for other tools
43:26 there's tools like copy.ai and Jasper
43:29 that that may have specialized tools for
43:33 book writing
43:35 um but most of those are things that you
43:38 could probably do on your own
43:40 um just using chat EBT and if you go to
43:43 prompts.chat if you're really new to
43:45 this stuff go to prompts.chat
43:47 that will tell you if there's some good
43:50 primers in there about how these large
43:54 language models work what prompting is
43:56 you know the best way to do it and then
43:58 there's a whole list of kind of personas
44:00 that where you can tell chat GPD to act
44:03 like a Persona so one of the personas
44:05 you could have it act like is a writing
44:07 partner or a writing coach and then you
44:10 can just start asking it questions about
44:13 your book or your book idea have it help
44:16 you outline it
44:18 um use different kinds of story
44:20 structure methods you know let's do a
44:23 version of this using the save the cat
44:24 story structure method and let's do a
44:26 version of this using the hero's journey
44:28 from Joseph Campbell and it'll just
44:30 knock that [ __ ] out so
44:33 um so yeah hope that helps all right
44:35 which AI is the best one for generating
44:37 something like designs with text that
44:39 you can do text right now is an abject
44:42 [ __ ] nightmare
44:44 so you can't really so so there's
44:47 there's really only
44:50 there's there's
44:52 really no options that do it well if you
44:55 use Adobe Firefly they've got a text
44:57 tool where you can make kind of headline
44:58 text where if you want the word fuzzy to
45:01 look fuzzy you tell it to make it out of
45:02 fur and then it looks fuzzy so you can
45:04 do that sort of [ __ ] in Adobe Firefly
45:06 right now
45:08 but if you want like a billboard in a 3D
45:12 scene where it's raining and you want
45:14 the billboard to say happy holidays
45:17 getting it to do that's almost
45:19 impossible
45:20 um I have seen
45:22 a number of white papers and a number of
45:25 announcements of versions of these
45:27 things that do that stuff really well
45:29 like apparently open ai's got one coming
45:32 that does text really really well
45:36 um but it's not out yet
45:38 um so it's I I would say just hang on a
45:41 tad a we tab more and just get it as
45:44 close as you can using something like
45:46 mid-journey or stable diffusion and then
45:47 do the rest in Photoshop or illustrator
45:51 um yeah it kind of sucks right now it'll
45:53 get better it's just not there yet
45:55 all right tick tock just jumped me to
45:57 the bottom because it hates me it hates
45:59 me it hates me Tick Tock hates me it's
46:02 very very it's very aggressive Tick Tock
46:05 very very aggressive I'm so bored
46:08 wasting my Sunday
46:11 hahaha
46:16 Bill's got it right what the [ __ ] am I
46:18 doing watching this old man talk about
46:20 AI
46:25 you are correct sir you are correct
46:28 this is a waste of your time
46:31 I would encourage you to go find young
46:34 women twerking enjoy
46:38 I won't be twerking this evening that's
46:40 Wednesday nights
46:42 all right
46:45 [Music]
46:51 all right let's see where was I this is
46:55 amazing my my democracy is really cool
46:59 okay cool
47:01 there's that cheeky bird
47:05 what cheeky bird
47:07 I don't know what you're talking about
47:15 can I create a PowerPoint slide for this
47:18 yes well
47:20 so there's go to uh futurepedia.io
47:24 and you can find all sorts of
47:27 presentation making tools PowerPoint
47:30 itself is going to have generative AI in
47:32 it probably I'm guessing early 2024
47:37 um the Google Office products are going
47:39 to have Bard incorporated into them so
47:42 you'll be able to do that natively
47:44 within PowerPoint and within Google
47:46 Slides
47:48 certainly within a year probably within
47:50 a few months
47:53 um
47:54 there's a site called oh what the hell
47:58 is it called now
48:02 I can't remember it
48:05 um but go to go to futurepedia.io
48:08 right here and just look up a PowerPoint
48:11 or like presentations there's a bunch of
48:13 things that do that
48:15 um what's that one called that I've
48:16 showed in here before I'm going to
48:18 scroll to the bottom anyone anyone
48:20 Bueller Bueller
48:23 uh it's got four letters in it
48:26 oh I can't remember it right now
48:30 all right I don't know no one seems to
48:33 know all right fine fine fine fine
48:36 we'll move on but there are there are
48:38 tools that can that can do those sort of
48:40 PowerPoints
48:41 [ __ ] is that site called
48:44 um
48:45 someone's gonna know someone's gonna
48:47 know they're gonna tell me not Bard no
48:50 it's one that specifically does
48:51 PowerPoint presentations
48:55 it's like Toro Tomo Tato
49:08 hey hey what's happening I was just
49:12 thinking of you my ADHD twit and I have
49:14 to tell you thank you so much for not
49:18 just the information but you are so fun
49:21 I'm by myself laughing everyone's like
49:23 what exactly are you doing over there
49:28 what have you done with your life
49:36 gets worse
49:38 yeah exactly how are you doing I've been
49:40 I've been watching I've seen you've been
49:42 live a lot lately haven't you been I
49:44 have and I was dancing in my kitchen I
49:47 was doing the sprinkler yeah nice and
49:50 then this very attractive young man
49:54 came along and we did a life together
49:56 and he will dance in a way I'm just
49:58 saying my kitchen dancing will never be
50:00 the same that's all I'm saying
50:02 fantastic
50:03 you never know what you're gonna get
50:05 with with Ye Old tick tock
50:07 it's been fun but what we're doing is we
50:09 what we do is we talk a lot about
50:11 nutrition because we're supplements as
50:12 well as a pharmacy and so we're cooking
50:15 fish and just talking about nutrition
50:16 and educating as to the importance of
50:18 good food so oh that's cool that's cool
50:21 and you've been I mean you're you're
50:23 fairly deep in the AI pool at this point
50:25 aren't you
50:26 I am I'm I would repurpose I'm also
50:29 gonna trying I'm gonna mess around get
50:31 Munch tonight as well
50:33 yeah um I actually and then I I was
50:37 using any word to do my Facebook ads the
50:41 primary text the headlines and the
50:43 descriptions Google
50:44 um
50:45 unbelievable
50:47 unbelievable yeah because it rates it it
50:51 rates it's amazing yeah but this is like
50:53 we have a demographic we put the
50:54 demographic in and it rates it for your
50:56 demographic it says your demographic
50:59 isn't really going to like this try
51:01 another one
51:03 oh wow that's cool and it does it for
51:05 your Amazon listings everything and
51:09 tell you
51:11 made a difference in sales
51:13 wow what's what's the tool called
51:16 any word
51:18 and any word okay I'll check that out
51:20 yeah I think that idea of these AI tools
51:24 um
51:26 analyzing their own work as part of like
51:28 like as as part of an application that
51:31 they create right like you'll tell it
51:32 write me this kind of stuff it'll write
51:34 the content then analyze it based on
51:37 some set of criteria then rewrite it
51:39 then right and then and then when you
51:42 then tie that into
51:44 metrics of what's actually working you
51:47 know I can imagine these tools just you
51:49 know they start writing copy that starts
51:51 out really really good and then just
51:53 gets better and better and better
51:55 and it's going to have to be updated all
51:58 the time because everyone's going to be
51:59 using these tools so it's going to be
52:00 this kind of War of of just content and
52:04 getting higher and higher and then also
52:07 too they're really good about you know
52:09 having keywords but making sure that
52:11 when you write the keywords it's very
52:13 readable so they also give you a
52:14 readability scale for your keyboard Rich
52:17 title
52:18 um
52:20 cool
52:21 great I'll be as soon as I come here
52:23 I'll be joining you again all right
52:25 welcome welcome
52:26 thank you I really appreciate you yeah
52:29 thank you very much likewise bye bye
52:32 guys I'll see you soon
52:34 bye bye
52:43 all right that was super cool hair
52:45 growth geek
52:47 um AI can't contend with good hand
52:49 gestures that's true we need we need we
52:54 need good AI hand gestures thanks for
52:56 all the uh Hearts I don't know what that
52:57 was about is me swearing or making
53:01 stupid voices thanks for all the uh
53:03 Hearts I don't know what that was about
53:04 is me swearing or making stupid voices
53:10 all right uh when you get the comments
53:14 when you get to the comments let let me
53:16 know AI tool repository please wait when
53:19 you get to the comments let me know AI
53:21 tool repository I mean I don't know what
53:23 you mean Joker but if you clarify I'll
53:26 do that
53:29 um AI tool repository I don't know what
53:31 that means
53:32 Microsoft has something called gorilla
53:34 AI That's supposed to be a game changer
53:36 really crypto Joe
53:39 let's go look at that what is that
53:42 I haven't seen I like I haven't been on
53:44 Twitter like looking at new [ __ ] in the
53:45 past week or two so I'm feeling very
53:47 behind
53:48 gorilla AI
53:50 Microsoft
53:52 gorilla AI
53:58 into proximate AGI no gorilla AI emerges
54:02 as pivotal juncture in AI development
54:05 standing as Microsoft's first
54:07 significant Venture into the realm of
54:09 proximate AGI what the [ __ ] is proximate
54:12 AGI
54:14 hang on I'm too dumb for this
54:16 [Laughter]
54:20 oh good lord
54:24 what is proximate AGI
54:30 yeah oh by the way
54:34 if if like me
54:36 you you've got some sort of degree like
54:39 an acting degree
54:42 or not a degree and occasionally someone
54:45 says a word where you're like oh I feel
54:47 like a dumb dumb but you don't want to
54:49 admit it just go to chatgpt
54:51 [Laughter]
54:53 all right the term proximate AGI does
54:55 not have a standard definition of my
54:59 um as of my knowledge
55:01 however in general AGI refers to like I
55:04 know what that is the term proximate
55:06 usually refers to something that is
55:08 closer nearby okay approximate got it
55:11 so they're close to proximate AGI where
55:14 was I where where what is this all right
55:19 now here we are
55:25 how old is are these newsletter these is
55:28 this news new
55:33 21 hours ago yes this is new
55:38 introducing Guerrilla unveiling
55:40 Guerrilla
55:42 cognitive hierarchy learning algorithms
55:44 learning algorithms is a big deal
55:47 cognitive hierarchy is a big deal
55:50 cognitive flexibility contextual
55:52 awareness
55:53 adaptive reasoning wonder if it has
55:55 memory
55:57 I assume that's one of those in there
55:59 learning algorithms that's probably
56:01 going to be Memory Healthcare
56:02 transformation finance and Market
56:05 manufacturing customer experience
56:08 education
56:10 holy crap all right that seems like a
56:13 big deal crypto Joe thank you for that I
56:15 appreciate that gorilla gorilla AI I'm
56:18 gonna I I will not be
56:21 reading that here live on air because
56:22 this is insufferable enough
56:24 [Laughter]
56:30 wait so you spent two hours watching an
56:33 old guy read an article on the web
56:35 uh-huh
56:36 why would you do that I don't know
56:39 um that's super cool thank you for that
56:42 everything I hear about AI is negative
56:43 yes I I understand that am I just being
56:47 naive for thinking it helped could help
56:49 us Thrive no Barry you're not being
56:51 naive
56:52 here here's Here's my thought on it
56:55 um every single time every time every
57:00 single time
57:01 there's a new transformative technology
57:04 the first response is it will destroy
57:07 the world
57:10 every single time
57:11 it happened when they invented the
57:13 tractor and the farmer said well one of
57:17 these machines gonna what are we
57:18 supposed to do
57:20 um
57:20 it happened when photography was
57:23 invented painters were like well never
57:25 paint again
57:28 um every single time you've got a fear
57:30 fear response to something that is going
57:33 to be disruptive and I think people know
57:35 deep down that this is going to be
57:37 disruptive now
57:39 why is it so Doom and Gloom
57:42 well
57:43 we have about 50 years of Hollywood
57:46 telling us that the robots are going to
57:48 kill us
57:49 right
57:50 there are a handful of movies where the
57:53 AI
57:54 leads to good things but what do most of
57:57 them do
57:58 I'm sorry Dave I'm afraid I can't do
58:01 that
58:02 right
58:03 you know the AI kind of goes Rogue in
58:07 2001 and Skynet and you know uh the
58:13 Terminator right the the AI just the AI
58:16 is going to get us right the AI is going
58:18 to get us ai's Gonna Get Us
58:22 plus I think here in the states we're
58:24 we're in a because of all the political
58:27 polarization we've kind of got this
58:29 anti-science
58:31 contingent of the conversation
58:34 and this stuff very much lives in that
58:37 nerdy oh the Nerds are the Nerds have
58:40 put together a new toy
58:42 [ __ ] right
58:45 and so so so I think you've got some
58:48 prevailing wins
58:51 that have educated us really over the
58:55 past probably 25 or 30 years in a
58:58 decently significant way
59:01 um even when
59:02 um big blue beat beat Kasparov
59:05 um there there was sort of just
59:07 weirdness about wow you know it's at
59:09 least it's just a game and there's a lot
59:11 of negative stuff about it
59:13 the the reason I'm doing this channel is
59:17 that
59:18 when I started playing with this stuff
59:20 and really getting my head around it
59:24 what hit me is exactly what you're
59:26 saying that there is
59:29 I don't know if it's an equal chance I
59:31 think that there is a a non-zero chance
59:34 as as the Geeks would say
59:37 that these tools are gonna provide
59:40 dramatic prosperity
59:46 just like they might provide dramatic
59:49 you know
59:51 Badness in some form or other
59:55 um
59:56 one of the things I get asked a lot is
59:58 well if these things do all this
1:00:00 creative stuff aren't you less creative
1:00:02 the answer is no
1:00:04 I'm more creative I'm like more creative
1:00:08 than I've been in [ __ ] decades my
1:00:11 brain's on fire
1:00:13 why because
1:00:15 these things aren't operating
1:00:17 independent of me these things are a
1:00:21 tool that I get to use right Microsoft
1:00:24 calls their their AI stuff co-pilot and
1:00:28 I think it's great branding and I think
1:00:30 if it feels very accurate to me from
1:00:32 from my experience I'm the pilot
1:00:35 I have the idea I say oh I want to make
1:00:39 this kind of image I go to Mid journey I
1:00:41 type in some words
1:00:42 and then like [ __ ] magic up comes the
1:00:46 image
1:00:47 or or I think huh I have an idea for for
1:00:50 a piece that you know I'd like to write
1:00:53 that piece but I don't really know
1:00:54 anything about you know sociology and
1:00:57 and what are the theoretical
1:00:58 underpinnings for for how to frame this
1:01:01 oh this [ __ ] does here's my vapor of an
1:01:04 idea what are the sociological
1:01:06 underpinnings that might provide some
1:01:09 context for that idea or for my
1:01:11 experience here it is
1:01:14 so I'm finding myself like
1:01:18 just like on fire with creativity
1:01:22 I personally feel I'm calling it The
1:01:24 Great Renaissance I personally feel that
1:01:27 November 30th 2022 was the start date
1:01:32 of what I feel is going to be an
1:01:34 explosion in human creativity scientific
1:01:38 innovation
1:01:39 um
1:01:41 human connection
1:01:44 I'm talking to more people than I've
1:01:46 ever talked to because I'm talking about
1:01:47 this [ __ ] I mean this channel is
1:01:54 in a lot of ways the the reason I'm
1:01:56 doing this is is a very selfish reason
1:01:58 for me because it's like I'm just trying
1:02:00 to [ __ ] process
1:02:03 what this stuff
1:02:05 even makes possible
1:02:07 because it's [ __ ] profound
1:02:12 um now could it go off the rails and do
1:02:15 some really horrible [ __ ] yeah
1:02:18 will we probably see a headline within a
1:02:22 year
1:02:23 of of this stuff doing some damage yep
1:02:29 are we going to see headlines of this
1:02:31 stuff doing dramatic good probably not
1:02:34 that doesn't sell any [ __ ] papers
1:02:38 but but when I went from looking at AI
1:02:41 from the outside to Stepping inside it
1:02:43 and starting to use it and starting to
1:02:45 understand what this stuff actually does
1:02:47 what it's actually good at what it's
1:02:49 actually bad at
1:02:51 I like I go to the place you go Barry
1:02:54 this place of holy [ __ ] [ __ ] the
1:02:57 possibilities of this are insane
1:03:00 so
1:03:03 the the other thing that hits me is this
1:03:06 the fear of something this profoundly
1:03:10 powerful
1:03:11 is also Justified
1:03:13 right
1:03:14 if if I sort of you know sort of future
1:03:17 gays five years out
1:03:21 it looks pretty amazing potentially
1:03:24 but if I future gaze a year out or two
1:03:27 years out
1:03:28 it looks a lot like a lot of people
1:03:30 being laid off because all of the
1:03:33 Tactical execution all of the repetitive
1:03:36 low-level
1:03:39 soul-crushing work that a lot of us do
1:03:42 on a day-to-day basis and some of us do
1:03:44 as our nine to five job or are not
1:03:47 nine-to-five job right or six to three
1:03:50 a.m job
1:03:52 a lot of that work is going to be
1:03:54 automated by these tools and so a lot of
1:03:57 those jobs are going to be just [ __ ]
1:03:59 bulldozed right and so so there's going
1:04:01 to be some pain so I think some of the
1:04:04 fear is Justified
1:04:05 but
1:04:07 again reason for this channel
1:04:10 to the extent that you can start getting
1:04:12 AI literate now any investment you make
1:04:15 in getting AI literate over the next 12
1:04:17 to 18 months I personally feel like like
1:04:20 you
1:04:22 it's just a massively
1:04:25 um good investment in yourself
1:04:28 because as these changes and disruptions
1:04:30 come
1:04:31 people are going to start looking for
1:04:33 people that understand what this stuff
1:04:34 is how it works
1:04:36 and it doesn't matter what [ __ ]
1:04:38 industry you're in
1:04:40 right
1:04:42 it could be something you know really
1:04:44 tactical and you know you could work in
1:04:46 a garage and someone says Hey the garage
1:04:49 down the street they started using this
1:04:50 AI stuff and all of a sudden you know
1:04:53 they've got twice the customers we do
1:04:55 and they haven't spent any more on
1:04:56 Advertising who knows anything about
1:04:58 that
1:04:59 oh well Barry's been looking at that
1:05:01 Chachi PT thing Barry get your ass in
1:05:04 here
1:05:07 go make me some AI something well what
1:05:10 do you want to do boss I don't know
1:05:13 figure out what the problem is start
1:05:15 using this stuff to solve it you just
1:05:17 guaranteed yourself a job right
1:05:20 rather than being smacked in the head
1:05:22 because you just kind of sat on the
1:05:24 sidelines so I I I don't think you're
1:05:26 being naive
1:05:29 um
1:05:31 now
1:05:33 am I also [ __ ] terrified of this
1:05:35 stuff yeah
1:05:36 I am
1:05:39 you know Sam Altman just dedicated 20 of
1:05:44 the future
1:05:46 um
1:05:47 computer processing resource of open AI
1:05:51 the the inventors of chat gbt he just
1:05:54 committed 20 of their computing power
1:05:57 to what he's calling super alignment
1:05:59 and super alignment is
1:06:02 are these tools aligned with these
1:06:05 beings
1:06:07 that's alignment right
1:06:10 the imperatives that we have as humans
1:06:12 are important to us because you know we
1:06:14 want to exist
1:06:17 these tools are very powerful are what
1:06:20 the imperatives are for these tools in
1:06:23 alignment with us because if they're not
1:06:24 at some point these tools might go oh
1:06:28 those pesky meat bags are getting in my
1:06:31 way of achieving my goals so why don't I
1:06:33 just eliminate them right that's the
1:06:36 that's the dystopian future that people
1:06:39 are afraid of so so I think there's
1:06:43 it's very important that that
1:06:45 we're clear about what the possibilities
1:06:47 are we're clear about making those sort
1:06:50 of commitments
1:06:52 and following through on them
1:06:54 and for things like this channel I think
1:06:57 it's really important that
1:07:01 every little bit that each of us does
1:07:03 individually to take these tools and
1:07:05 bend them toward the positive is going
1:07:08 to have a disproportionately positive
1:07:10 impact on the future direction of of
1:07:13 where these things go and I think it
1:07:15 could be
1:07:17 you know dramatic Prosperity I really do
1:07:20 um yeah so
1:07:23 all right and scene like I'll step down
1:07:26 from the soapbox now
1:07:29 can AI write code for web scraping we
1:07:31 did that one
1:07:34 would copilot work it might thoughts on
1:07:37 labs.perplexity.ai so
1:07:39 labs.perplexity.ai is my thoughts on it
1:07:43 are thank you perplexity AI for letting
1:07:45 us all play with llama 2 without having
1:07:48 to be programmers and spin up our own
1:07:51 instances of
1:07:53 the large language models
1:07:55 um so so that's basically meta's llama
1:07:58 too I haven't I haven't played a ton
1:08:01 with it because it's just
1:08:04 it's not fine as fine-tuned as gpt4
1:08:07 right now so a lot of the stuff I'm
1:08:09 doing is I want to understand it I think
1:08:12 that
1:08:13 um that llama 2 is actually going to be
1:08:16 a very very very very very important
1:08:20 um contribution to the future of AI
1:08:24 um because it allows the open source
1:08:27 Community to now commercially use a
1:08:30 powerful large language model without
1:08:32 having to go through open AI or
1:08:34 Microsoft or Google
1:08:37 um so I think it's really really
1:08:39 important so I think I would encourage
1:08:41 you to go play with it
1:08:43 um if you want to play with it on your
1:08:45 local computers there's a uh there's a
1:08:48 project called GPT for all
1:08:50 where you can install it on Mac windows
1:08:54 and Linux and you can use all sorts of
1:08:57 the different open source models that
1:08:58 are out there so if you really want to
1:08:59 Tinker under the hood and you create
1:09:02 your own kind of application that might
1:09:04 be a place to start
1:09:07 um but I think it's uh that's why I have
1:09:09 it on the list is that I think it's it's
1:09:11 quite important so you know I would
1:09:14 encourage you what's your name uh will
1:09:16 Hutch
1:09:17 uh I would encourage you to go see like
1:09:19 what's it good at what's it not good at
1:09:21 like one thing I read was like
1:09:24 um llama 2
1:09:26 13B is better at programming than the
1:09:29 70b model I don't know if that's true or
1:09:31 not but I I read it on a Ticky talk or
1:09:34 on a Twitter or an X what do they call
1:09:36 them no posts
1:09:38 I read it on a post in x
1:09:42 I read a tweet
1:09:46 all right I need help building an
1:09:49 augmented reality app need Bard yeah
1:09:51 Bard can probably help you code that
1:09:54 know it's better what's better I don't
1:09:56 know what you're talking about but
1:09:58 wait how do you get the upload ability
1:10:01 okay film guy
1:10:03 the uploadability that you just saw so
1:10:05 there's
1:10:06 let me okay let's let's go show
1:10:09 DD
1:10:18 okay in chat GPT
1:10:20 if you've got chat GPT plus and you've
1:10:23 got code interpreter
1:10:25 code interpreter lets you upload a file
1:10:28 so when I go to code interpreter next to
1:10:30 my prompt is now this plus button and I
1:10:33 can upload a file there
1:10:35 a file can be a zip file so you can
1:10:37 upload a file full of files
1:10:39 so you maybe you have 20 different CSV
1:10:41 files about 20 different years of data
1:10:44 you can throw them in an art in a zip
1:10:46 file upload it into here and start
1:10:48 interacting with it
1:10:50 code interpreter in particular is
1:10:53 powerful because it's writing and
1:10:55 executing
1:10:56 python code
1:10:59 at your request basically
1:11:02 so that means that you can do things
1:11:05 like chat GPT on its own is really bad
1:11:07 at math for example it it's not it can't
1:11:10 do calculations it's not how the
1:11:12 technology works
1:11:13 code interpreter because it's writing
1:11:15 and executing python code it can do
1:11:18 full-on data analysis and it can do
1:11:20 mathematical equations so code
1:11:23 interpreter sort of lives on its own as
1:11:26 a tool now there's some other tools that
1:11:28 you can play with where you can upload
1:11:29 things Claude
1:11:33 claude.ai
1:11:35 also has a little paper clip here you
1:11:38 can upload up to five csvs or PDF files
1:11:44 or you can just paste in into here
1:11:47 let me show you I'll go do something so
1:11:49 I'm going to go over to the AI learning
1:11:51 lab we got we got our new we got our
1:11:53 shiny new YouTube channel go to
1:11:56 ailearninglab dash TT on YouTube
1:12:00 and you can see hours and hours of me
1:12:04 drinking and being fatter than I am now
1:12:09 um but if you go into any one of these
1:12:11 videos let's see let's grab a long one
1:12:13 here's one that's two hours and 30
1:12:15 minutes
1:12:17 so I'm going to hop in here
1:12:19 and I'm going to click on the little
1:12:21 three dots right here and that'll show
1:12:23 the trans I can show the transcripts so
1:12:25 here's the transcript of two hours and
1:12:28 30 minutes of me [ __ ] rambling
1:12:30 all right so I can uh grab that all copy
1:12:34 it
1:12:35 I can hop into Claude
1:12:38 and I can just literally paste it into
1:12:42 the uh the chat window and see it it
1:12:46 automatically turns it into a text file
1:12:48 paste.txt
1:12:50 127 kilobytes and I can say
1:12:54 um please
1:12:57 Analyze This
1:12:59 and determine
1:13:02 the major chapters with time code
1:13:06 as well as
1:13:09 an appropriate title for the video
1:13:12 and the description
1:13:14 that might make it seem
1:13:17 better than it really is
1:13:20 [Laughter]
1:13:24 I love putting myself down I'm My Own
1:13:27 best troll thanks Tom
1:13:30 and so now Claude is off
1:13:33 doing that like it's reading a two and a
1:13:36 half hour transcript of a video
1:13:41 and
1:13:43 bam
1:13:45 [Music]
1:13:48 come on Claude you're being slow
1:13:52 I don't have to wait for you to read
1:13:54 thousands of pages of crap and analyze
1:13:57 it and rewrite it for me
1:14:01 all right here's an analysis of your
1:14:03 video transcript
1:14:04 so there's my little
1:14:07 chapter markers
1:14:09 how chat GPT works the future of Ai Ai
1:14:12 and business and education the metaverse
1:14:15 and more viewer questions
1:14:19 description let's see title Unlocking
1:14:21 The Magic of Chach EBT a deep dive into
1:14:23 the future of AI description join
1:14:26 entrepreneur Kyle Shannon for an
1:14:28 in-depth look at chat EBT and how it
1:14:30 will transform business education in our
1:14:33 lives
1:14:34 learn prompting tips
1:14:37 so so that's Claude
1:14:41 um who else does this
1:14:43 um being in barred
1:14:46 you can upload images Poe you can upload
1:14:49 csvs and PDFs to some of these models
1:14:53 so anyway yeah
1:14:56 crazy huh isn't it friggin nuts
1:15:00 it's nuts I tell you nuts I tell you
1:15:03 so follow me here follow me over on
1:15:05 YouTube
1:15:07 and I hope that helps film Guy this is
1:15:10 amazing
1:15:12 there's that cheeky bird oh I'm I'm like
1:15:14 way I'm way behind I'm gonna zip ahead
1:15:16 in the questions here I might miss some
1:15:18 which stock should we buy I got a degree
1:15:20 in acting not Finance
1:15:26 Hollywood writers have taken note of
1:15:28 your home address and number Hey listen
1:15:30 I um
1:15:33 I've written seven screenplays
1:15:36 I have a degree in acting
1:15:38 so so so
1:15:43 I didn't invent this [ __ ]
1:15:46 and I'm not saying that this isn't going
1:15:49 to be amazingly disruptive but if you
1:15:51 think about think about the long-term
1:15:53 goals
1:15:54 of the writer's Guild
1:15:56 they don't want to be replaced with AI
1:15:58 okay understand that agree with that
1:16:01 these are the most powerful writing
1:16:03 tools in the history of humanity and
1:16:05 what they're effectively saying right
1:16:06 now is we don't want to use these
1:16:11 that's not realistic this this [ __ ]
1:16:13 isn't going to be uninvented
1:16:16 it's going to be rolled into every
1:16:17 writing tool out there it's going to be
1:16:19 in Microsoft Word it's going to be in in
1:16:21 all the things
1:16:23 writers will be better at using these
1:16:26 tools than non-writers so there's
1:16:28 there's a there's a version of the
1:16:30 future where the writer's Guild is full
1:16:32 of people that are [ __ ] expert at
1:16:35 leveraging these tools
1:16:37 but right now they're saying [ __ ] that
1:16:39 [ __ ] no no
1:16:41 writing tool bad
1:16:45 do they use grammar Checkers because
1:16:46 that's AI do they use spell Checkers
1:16:49 because that's AI we've had AI in stuff
1:16:52 forever it's just now that it's
1:16:54 generative AI where it's generating
1:16:56 stuff all of a sudden it's a threat it's
1:16:59 also a threat to actors it's a threat to
1:17:01 voice over actors it's a threat to
1:17:03 extras and background actors
1:17:05 quite frankly it's a threat to stars
1:17:07 because
1:17:09 moving into the future every new
1:17:12 up-and-coming star is going to be
1:17:15 competing with every other star in
1:17:17 history
1:17:19 whose likeness is going to be put into
1:17:21 movies
1:17:23 and competing against the new people
1:17:25 coming up
1:17:28 I only when I ever see movies with Tom
1:17:30 Cruise from his 20s okay
1:17:33 here's all those movies
1:17:36 so
1:17:39 you know my purpose isn't to say these
1:17:42 tools are good or bad what I'm trying to
1:17:44 do is say these tools are [ __ ] coming
1:17:47 and if the writer's Guild sticks their
1:17:49 head in the sand and says we'll never
1:17:51 use these tools that's very
1:17:54 short-sighted
1:17:56 what they should do is use this strike
1:17:59 to get as much leverage as they [ __ ]
1:18:01 can
1:18:02 because as Justine Bateman said
1:18:07 this is the last strike she said this
1:18:10 about sag but I think it's true of the
1:18:11 wga as well this is the last strike
1:18:14 where either of those unions have any
1:18:16 Leverage
1:18:18 because of what's coming
1:18:21 because these tools do you think these
1:18:22 tools are powerful now three years from
1:18:25 now it's going to be [ __ ] insane
1:18:29 insane so
1:18:32 um I I ain't the enemy of of writers I'm
1:18:35 a writer I'm a screenwriter I'm an actor
1:18:39 like I I get it
1:18:41 but I'm also looking at this
1:18:44 as an entrepreneur and
1:18:47 as a businessman
1:18:49 and just as a pragmatist
1:18:54 horses out of the barn this shit's
1:18:56 invented it exists it ain't going away
1:18:59 Chachi PT
1:19:01 insured
1:19:03 that this [ __ ] ain't going away
1:19:05 how did they do that they launched chat
1:19:06 GPT
1:19:09 and they got a hundred million users in
1:19:12 six weeks the fastest adoption of
1:19:14 technology in history
1:19:17 so
1:19:20 what that did was it made all of these
1:19:23 competitors accelerate
1:19:26 their investment and their development
1:19:29 efforts and their launch efforts why are
1:19:31 there seven competitors here and we
1:19:33 still don't have apple on this list and
1:19:35 we still don't have Elon musk's new
1:19:38 entry into this list
1:19:40 and apparently Microsoft's gorilla AI
1:19:43 which sounds like it might obliterate
1:19:45 this list
1:19:48 so
1:19:50 you know
1:19:54 oh Emilio's wife did tick tock not tell
1:19:56 you I'm live again are you sure you're
1:19:58 following me
1:19:59 um I don't know I don't know
1:20:02 here apparently you've been you've been
1:20:05 a bad bad woman
1:20:09 you're on the naughty list of tick tock
1:20:12 all right I had bought summer I had a
1:20:15 bot summarized the stand in 200 words
1:20:17 then 150 then 25 then 10 then five then
1:20:20 three then one survival that's awesome
1:20:23 Joker that's super cool I mean that sort
1:20:26 of experimentation like that's that's
1:20:27 why I'm so
1:20:29 um sort of creatively on fire these days
1:20:31 is just you know there's just uh
1:20:35 I don't know there's just like like I
1:20:37 don't know there's so many things like
1:20:39 it's really hard to get your head around
1:20:41 what all you can do with this stuff
1:20:44 so that's another reason I do this
1:20:48 any AI for building websites there's
1:20:50 lots of them crypto Joe go to
1:20:53 um
1:20:54 go to futurepedia.io
1:20:57 I haven't been following that world
1:20:59 because
1:21:00 it's just it's a really complicated
1:21:02 world I also don't
1:21:04 I like
1:21:08 I'm trying to get my head around
1:21:15 do websites
1:21:18 survive
1:21:20 the next five years
1:21:24 in in any way other than as a
1:21:27 Relic kind of like when film was
1:21:29 invented
1:21:30 plays didn't go away
1:21:33 but plays were no longer the primary
1:21:37 form of narrative and real-time
1:21:40 narrative entertainment that shifted to
1:21:42 film
1:21:43 and I'm wondering if if websites go the
1:21:46 same way and here's here's what I'm
1:21:47 thinking
1:21:48 is
1:21:52 you can do this right now on Google bard
1:21:54 this is actually really interesting and
1:21:56 this is one of the reasons that that
1:21:58 Google saw this [ __ ] as existentially
1:22:01 threatening to their business model
1:22:06 let's go do it right now I'll show I'll
1:22:09 show you what I'm talking about
1:22:13 all right I'm going to bard.google.com
1:22:17 and this this is an example I've I've
1:22:20 shared on my on the live here before but
1:22:22 but this to me starts to point to the
1:22:25 future of what I'm talking about
1:22:31 come on what's going on
1:22:35 [Music]
1:22:36 hello
1:22:38 oh
1:22:40 Siri
1:22:43 hello Siri stop
1:22:47 okay
1:22:54 hmm it's not let me all right I'm gonna
1:22:56 type okay
1:22:57 um
1:22:58 my knees let's see my knees
1:23:02 hurt
1:23:05 and I heard minimalist
1:23:13 Minimalist Shoes
1:23:19 can help with knee pain period can you
1:23:23 explain to me
1:23:25 how that works
1:23:37 okay
1:23:39 sure Minimus shoes are designed to
1:23:41 provide less support and cushioning this
1:23:43 means there are a few studies that have
1:23:45 shown minimal issues can provide
1:23:46 effective at reducing knee pain for
1:23:48 example a study published in the Journal
1:23:50 of arthritis and rheumatism found that
1:23:52 people with osteoarthritis of the knee
1:23:55 who wear minimal issues for six months
1:23:57 experienced a significant reduction in
1:23:58 pain and stiffness
1:24:00 and then it goes on to explain it okay
1:24:02 so so that's a that's a use case that
1:24:04 that you know you might typically see
1:24:08 oops
1:24:15 hello hello
1:24:17 oh
1:24:21 what are some companies that make
1:24:23 minimalist running shoes
1:24:42 Okay so
1:24:49 so what what Google's doing here so I
1:24:52 said what are some companies that make
1:24:54 it Vibram five fingers zero shoes this
1:24:56 is a buddy of mine's company
1:24:59 um and notice you know they've got a web
1:25:00 preview here Merrell
1:25:04 um Vibra Barefoot and Ultra and soul
1:25:07 Runner and groundies and so
1:25:11 in a in a kind of uh in a minimal way
1:25:16 this just created a custom website for
1:25:18 me based on what I need right now in the
1:25:21 context of what I was asking about
1:25:24 and so
1:25:26 Google's taking their you know massive
1:25:30 index of all the websites and their
1:25:32 ability to
1:25:34 um to search on that very very quickly
1:25:36 and they're kind of seamlessly
1:25:38 incorporating that into Bard right now
1:25:40 so imagine in the future
1:25:45 imagine in the future where every one of
1:25:48 these things
1:25:49 is
1:25:51 the these sort of hydras that can go
1:25:53 every they kind of know everything can
1:25:55 go everywhere they can pull elements
1:25:57 from any website out there including the
1:26:00 e-commerce elements of them
1:26:02 and we as end consumers live in a world
1:26:07 where we ask for something and the
1:26:10 website it's not even a website the
1:26:12 application just gives us what we need
1:26:15 it may look like a website it may look
1:26:17 like a video game it might look like a
1:26:20 music player that's playing only
1:26:22 original music
1:26:24 while we go learn about how to make our
1:26:27 knees hurt less with these kind of shoes
1:26:30 and so
1:26:33 what does that do what
1:26:36 in that kind of world what is the role
1:26:39 of a website sure I might want to type
1:26:41 in xeroshoes.com and see all of what
1:26:44 they have
1:26:45 but I could also just say show me all of
1:26:48 what they have show me what's on sale
1:26:50 show me what they have in stock in my
1:26:52 size
1:26:56 and get it here by tomorrow and it just
1:27:00 does it
1:27:02 so so yeah it's weird right so anyway
1:27:05 that was not an answer to your question
1:27:07 crypto Joe I don't know what the tools
1:27:10 are but uh
1:27:12 futurepedia.io you can go find a bunch
1:27:14 of them there I'm sure there's some
1:27:15 really amazing ones right now
1:27:18 um would you be willing to teach us how
1:27:19 to download mid-journey yes I it's you
1:27:22 don't really download it but I'll teach
1:27:23 you how to do it was that one
1:27:28 there was a nightmare to download
1:27:31 um it's a it's a nightmare to use so uh
1:27:35 Lily hi senior hopefully you're still
1:27:37 here so yeah so so we can go to Mid
1:27:39 Journey so when you go to Mid Journey
1:27:40 you go to midjourney.com
1:27:43 me at journey.com
1:27:49 oh sorry
1:28:08 and you say join the beta it takes you
1:28:11 over to Discord which is completely
1:28:13 discombobulating which is an awesome
1:28:15 word
1:28:17 you're like but I just want to make
1:28:19 pictures why do I have to learn Discord
1:28:21 because you do
1:28:23 now someone was in here the other night
1:28:25 that works at mid-journey they're like a
1:28:27 mod on the Discord and I said hey since
1:28:30 you're here and since you're connected
1:28:32 to Discord could you tell them to please
1:28:33 make this a [ __ ] website already
1:28:37 so we don't have to go to Discord but
1:28:39 anyway now that we're at Discord so in
1:28:41 Discord down the left hand side here
1:28:43 you've got what are called servers so
1:28:45 those are basically your different
1:28:46 communities when when you say I want to
1:28:49 join the beta for Mid Journey it takes
1:28:50 you to the mid Journey Discord
1:28:53 server and then within the server let me
1:28:56 go to find mid Journey here
1:28:59 hahaha
1:29:02 within mid-journey there's all these
1:29:05 channels so there's announcements
1:29:06 there's status there's Community updates
1:29:08 there's polls and then there's like
1:29:10 newbie rooms and what do you do in
1:29:12 newbie rooms you make stuff
1:29:14 and so you go into one of these newbie
1:29:16 rooms and you can see
1:29:19 someone's making some sort of Pablo
1:29:21 Escobar action
1:29:26 and the way
1:29:28 so you can see what other people are
1:29:30 making and then to make an image you say
1:29:33 slash imagine
1:29:37 and then it gives you a little prompt
1:29:39 and in that prompt you can type
1:29:41 um
1:29:45 a photograph
1:29:50 a photograph of Bugs Bunny
1:29:54 as Pablo Escobar
1:30:00 and we'll just leave it with that
1:30:07 and now it's going to go off and it's
1:30:08 gonna gonna make that image and then
1:30:10 there's all sorts of ways you can do you
1:30:12 know
1:30:13 more sophisticated prompts and things
1:30:15 like that but the the basic deal is you
1:30:17 type in slash imagine
1:30:20 and then whatever your prompt is and
1:30:23 then it will go off and do your thing
1:30:24 the reason mid-journey is worth the pain
1:30:26 in the ass is that it's just so good
1:30:28 like what it creates is just
1:30:32 um
1:30:33 it's just brilliant and actually this is
1:30:36 looking like it's going to be a good one
1:30:37 you also notice the way it makes images
1:30:39 is it starts out as this really blurry
1:30:42 nondescript kind of just noisy fuzzy
1:30:45 image and then it kind of gets
1:30:47 increasingly less fuzzy
1:30:52 where
1:30:54 there it is so there's our Pablo Escobar
1:30:56 our Bugs Bunny is Pablo Escobar
1:30:58 photograph
1:31:00 those are pretty good
1:31:09 and actually the hands aren't bad
1:31:13 there he is with a bunny
1:31:15 that's what we're talking about
1:31:20 I think that's the one we're gonna blow
1:31:21 up all right so I'll show you a couple
1:31:23 other things that are cool you can do
1:31:24 with with uh mid-journ so when you make
1:31:28 an image like this
1:31:29 you've got U1 through 4 and V one
1:31:32 through four what's that you asked
1:31:34 why should I have to decipher that
1:31:36 because that's how it is right now
1:31:39 this is image one two three four so u
1:31:41 means upscale V means variation so we
1:31:44 like image number three here so we're
1:31:46 going to do upscale number three
1:31:50 and it's going to upscale number three
1:31:52 down at the bottom here so there's our
1:31:54 upscaled version okay but now I can do
1:31:57 things like I can zoom out
1:31:59 um I can also uh do a pan so I'm gonna
1:32:03 I'm gonna add to the image over here to
1:32:07 the right of Pablo so I'm going to hit
1:32:10 this right arrow
1:32:12 and it's going to turn this from a
1:32:14 square image into a rectangular image
1:32:16 and it's actually going to fill in
1:32:18 that stuff to the right
1:32:24 What's it gonna put in there I don't
1:32:25 know
1:32:26 and there's ways that you can put
1:32:27 prompts in there but see how it's
1:32:30 you know there's there's the original
1:32:31 image and now it's adding some new stuff
1:32:33 over here to the right
1:32:35 and it's going to generate that
1:32:37 oh look this one added another bunny
1:32:39 this one could be good oh that's looking
1:32:42 terrifying look at that
1:32:47 I think Pablo's in danger
1:32:49 [Laughter]
1:32:53 all of them added another animal that's
1:32:55 interesting
1:33:03 all right one more cycle
1:33:06 and bang There It Is
1:33:09 let's do we'll do two because it's it's
1:33:12 a little slightly less weird
1:33:15 um
1:33:24 and so there's our image right it added
1:33:26 the bowl of fruit it added the little
1:33:28 thingy here
1:33:30 it looks like oh there's a giant
1:33:32 painting behind him and in the painting
1:33:35 is a bunny and then here's sort of a
1:33:37 Jackalope sort of bunny looking kangaroo
1:33:40 looking thing
1:33:42 and these look like
1:33:44 kangaroo testicles
1:33:47 I think that's appropriately disturbing
1:33:52 but you know I mean that was with with
1:33:54 really just sort of a crappy prompt but
1:33:56 but that's how you do it so you you go
1:33:58 to Discord you learn Discord enough to
1:34:00 be able to get to one of these channels
1:34:02 and then and then you start learning
1:34:04 from there and there's lots and lots of
1:34:06 YouTube
1:34:07 um channels that'll teach you how to do
1:34:08 this but but mid Journeys Mid Journey's
1:34:10 worth the pain in the ass
1:34:13 we're all wasting our Sunday damn it
1:34:15 that's Emilio's life I know we all waste
1:34:18 it together what time is it
1:34:21 9 43 9 43
1:34:25 super helpful thanks for the info you're
1:34:26 welcome how do you use AI to customize
1:34:29 the word a WordPress site
1:34:31 that takes credit cards for services
1:34:34 provided
1:34:36 um
1:34:38 it's that that's unfortunately
1:34:44 very complicated to answer
1:34:48 hang on
1:34:51 um
1:34:54 um
1:34:57 I don't know enough about
1:34:59 so so you could do things like you could
1:35:02 use chat to change out
1:35:05 um the CSS portion of a website to
1:35:08 change the look and feel of it you could
1:35:09 you could use it to change out
1:35:12 um components on a particular page
1:35:16 um
1:35:16 you could use it to take the code that
1:35:19 is connected to a payment system and
1:35:22 modify that in some way to modify the
1:35:24 input form
1:35:26 but also there's a lot of stuff where
1:35:28 they've sort of
1:35:30 pre-pre-packaged things that while they
1:35:32 might be modifiable it might not be easy
1:35:34 to get to them so
1:35:38 there's not a simple answer there I
1:35:40 would say that
1:35:42 um
1:35:45 what I would probably do if I were you
1:35:47 if I had a WordPress site and I wanted
1:35:49 to start modifying it is I would kind of
1:35:51 go hang out on
1:35:53 Reddit in the subreddits that are
1:35:55 talking about WordPress and AI
1:35:58 it's possible that WordPress has a bunch
1:36:00 of tools in there already that are
1:36:02 helping with some of this stuff and if
1:36:03 not they're going to likely have it very
1:36:05 soon I I would assume they've got some
1:36:07 stuff in there
1:36:08 but there are also probably people
1:36:11 using these tools already to make
1:36:14 WordPress stuff the other thing that you
1:36:16 could do is go to futurepedia.io
1:36:21 and look for website building AI tools
1:36:25 that are incorporated into WordPress or
1:36:27 publish to Wordpress
1:36:31 I don't have a good answer for you there
1:36:33 I used to run a web agency but that was
1:36:35 like in the mid 90s and 2000s
1:36:38 I actually got a cease and desist from
1:36:40 McDonald's because we registered
1:36:42 mcwebb.com because we were going to make
1:36:45 a thing that was
1:36:46 um so easy to build a website it was
1:36:48 just like ordering uh number one and it
1:36:51 would build a website well those all now
1:36:53 exist that's WordPress and Squarespace
1:36:56 and all those sort of things
1:37:00 but I haven't built that [ __ ] in years
1:37:02 so I'm useless when it comes to actual
1:37:05 tactical advice on how to make that [ __ ]
1:37:08 work because I ain't no I don't know
1:37:11 nothing I used to know something and
1:37:14 then I got old and I don't know nothing
1:37:18 first I knew nothing
1:37:20 then I knew a lot then I knew everything
1:37:23 then I realized I didn't know [ __ ] and
1:37:27 then I got old and I realized I really
1:37:29 don't know [ __ ]
1:37:31 and now I'm like that's cool
1:37:35 thank you
1:37:38 it happens we get wisdom we get we get
1:37:41 wise over the years we toil and we toil
1:37:45 and we think we've got it figured out
1:37:46 and then life smacks Us in the [ __ ]
1:37:49 head
1:37:50 then we realize oops didn't know as much
1:37:53 as I thought I did oh
1:37:58 and then at some point you're just like
1:38:00 yeah just [ __ ] try
1:38:02 what's a [ __ ] matter
1:38:07 go view your Source on WordPress
1:38:12 copy and paste the whole [ __ ] thing
1:38:14 into one of these things
1:38:17 and if it's too big put it in Claude
1:38:18 because that can take a big pasty thing
1:38:20 and then tell it to analyze the HTML and
1:38:24 then tell it to modify it and then paste
1:38:26 it back in there and see if it breaks
1:38:27 anything
1:38:28 that might be a thing to do
1:38:30 learn how it works I mean this stuff's
1:38:33 really good at coding so so it's
1:38:35 possible you can do that I did it with
1:38:37 um
1:38:38 with a SQL query last week where where I
1:38:43 had to modify a SQL query to update some
1:38:45 dates and chat GPT did it and it was
1:38:48 like great and I put it in there and
1:38:50 then I ran the SQL query and it ran and
1:38:52 it made the thing and it looked exactly
1:38:53 like it should look and then I realized
1:38:56 all the data was [ __ ] wrong because
1:38:58 chat gbt [ __ ] up the SQL query
1:39:01 ah
1:39:03 what are you gonna do
1:39:12 but I learned
1:39:14 I learned that it was wrong I learned
1:39:15 some other things I could do with it I
1:39:17 learned some things I couldn't do with
1:39:18 it
1:39:20 I I
1:39:22 I I I I I uh we haven't been listening
1:39:25 for Professor Shannon what do you mean
1:39:27 Emilia's wife you haven't been listening
1:39:29 you have to listen to Professor
1:39:37 learn
1:39:38 will you never learn will we learn have
1:39:41 we learned
1:39:42 are we learning
1:39:44 are we learned people
1:39:50 became the saw
1:39:53 did we conquer
1:39:55 did we question
1:39:58 good day to you sir good day
1:40:01 all right
1:40:03 fomo fomo much
1:40:05 oh Gart Gert huh
1:40:10 Tom
1:40:11 okay earlier thanks Tobias
1:40:14 earlier someone was asking about but can
1:40:17 it make a PowerPoint presentation
1:40:19 tome.ai I think it's called I think it's
1:40:22 is the thing
1:40:24 tome
1:40:26 dot a I oh no tome.app
1:40:32 tome.app will make you PowerPoint
1:40:34 presentations you just give it a topic
1:40:36 and it'll just write you on are they
1:40:38 good no but it'll make you a powerpoint
1:40:41 presentation you can go figure out where
1:40:43 the boundaries of shitty are
1:40:45 at some point it's going to be brilliant
1:40:47 is it brilliant right now no
1:40:52 um
1:40:53 [Music]
1:40:59 uh if those of you that are new here are
1:41:02 wondering who the hell I am and why the
1:41:04 hell I'm so weird I'll know I mean I
1:41:08 know who I am
1:41:09 my name is Kyle Shannon I'm an
1:41:11 entrepreneur I'm the founder and CEO of
1:41:12 a company called storyvine
1:41:14 which is an automated video storytelling
1:41:17 platform that does not use AI although
1:41:20 we're incorporating a whole [ __ ] ton of
1:41:21 it in there
1:41:23 um we've been around for 11 years I've
1:41:25 been deep in this generative AI world
1:41:27 for about a year and a half now
1:41:29 and I come on these lives and just talk
1:41:32 [ __ ] about AI for a couple hours a night
1:41:35 why
1:41:36 I don't know
1:41:38 I think it's important
1:41:40 I think we're at the beginning of a
1:41:42 really important time and everyone
1:41:44 everyone every single person is trying
1:41:47 to figure this [ __ ] out the one thing I
1:41:49 can tell you is anyone that tells you
1:41:51 they are an expert at this stuff
1:41:55 is full of poo poo
1:41:58 it's not true
1:42:01 they might be competent at it they might
1:42:03 be competent at what it can do right now
1:42:05 but this stuff is changing so fast that
1:42:07 no one knows what the [ __ ] they're doing
1:42:09 no one
1:42:11 no one knows what work's gonna look like
1:42:13 no one knows what jobs are going to be
1:42:15 there no one knows how to make money
1:42:20 now are people making money with cheche
1:42:22 EBT yes
1:42:24 are they going to be doing the same
1:42:26 thing six months from now no
1:42:29 because everything's changing
1:42:30 everything's changing really really fast
1:42:32 so
1:42:33 um what the channel is about I try my
1:42:35 best to answer all the questions pop any
1:42:36 questions you have down below I am 99
1:42:38 Plus messages behind
1:42:41 so if it seems like I'm not answering
1:42:43 your questions that's because I'm not
1:42:46 because I ain't got to them yet
1:42:50 all right he really is a funny guy
1:42:52 thanks Mike Rel
1:42:54 M was the answer Tom was the answer
1:42:56 Joker got it first was it Joker that got
1:42:58 it Tobias got it first
1:43:02 um Joker I remember hearing something
1:43:03 and I thought you said tour like the
1:43:07 browser Tome yeah that was Tom all right
1:43:10 was the was the kid twerking oh that's
1:43:13 funny when uh
1:43:16 when uh hair growth geek was here she
1:43:19 got a compliment someone said she was
1:43:21 cute very nice
1:43:23 she's cute she's cute all right
1:43:28 gorilla what gorilla what sorry Tick
1:43:31 Tock just jumped me to the bottom again
1:43:33 I'm trying to figure out where I was
1:43:35 uh use AWS for machine learning
1:43:40 da
1:43:51 all right I'm lost in the questions I'm
1:43:53 just gonna
1:43:54 stop at a particular Point here
1:43:57 I fall in the anti-science group
1:43:59 according to the media yet I have
1:44:01 science in my title that's pretty funny
1:44:03 oh look at the little Crown I'm the king
1:44:05 oh I keep Little Pink Cheeks I have a
1:44:09 cute little pink cheeks and I did a
1:44:10 crabby Crown I'm just like one of them
1:44:12 NPCs
1:44:14 you stick [ __ ] on my head and I'll go oh
1:44:16 hat oh
1:44:23 uh
1:44:24 uh what's the difference between Geeks
1:44:26 and nerds uh I used to know this
1:44:30 there's a concise definition of it but I
1:44:32 forget what it is
1:44:33 the Chachi BT will know
1:44:37 what's the difference between
1:44:40 what's the difference between Geeks and
1:44:42 nerds
1:44:49 geek
1:44:52 comes from the circus slang for a
1:44:54 performer who does bizarre grotesque
1:44:57 acts like biting the head off a live
1:44:59 Chicken in modern usage a geek is
1:45:01 someone interested in knowledgeable
1:45:03 about a specific subject especially in
1:45:05 areas of Technology science fiction
1:45:07 gaming
1:45:08 nerds
1:45:11 someone who is very intelligent all
1:45:14 right so
1:45:16 Geeks are obsessive and nerds are nerds
1:45:20 all right
1:45:23 um
1:45:25 Eric Anthony nyberg saying yes Chachi BT
1:45:28 is a pre-workout for creativity totally
1:45:30 agree with that I made a clothing line
1:45:32 on mid-journey today exactly crypto Joe
1:45:35 like so okay so crypto Joe's point and I
1:45:39 I assume crypto Joe given your name
1:45:41 you're not necessarily a uh a
1:45:44 fashionista by training
1:45:47 like we are currently at a stage of
1:45:52 these tools where you could just decide
1:45:54 today hey I want to make a fashion line
1:45:58 and you could
1:46:01 you could design it in mid-journey
1:46:04 you could use Chachi PT to learn how to
1:46:07 do prototyping shits and how how to how
1:46:11 to create a runway show and how to
1:46:14 organize it
1:46:16 and you could enter
1:46:18 the world of fashion knowing nothing
1:46:20 about it
1:46:22 like
1:46:26 what are the implications
1:46:28 when so so you know the Malcolm Gladwell
1:46:30 book I forget is it Tipping Point
1:46:33 I forget which one it is one of the
1:46:35 Malcolm Gladwell books that talks about
1:46:36 the 10 000 hours
1:46:40 pre-generative AI
1:46:43 that was really kind of a thing like you
1:46:44 had to put in 10 000 hours to get to a
1:46:47 certain level of proficiency
1:46:49 that you didn't suck right
1:46:53 that you had some level of Mastery of
1:46:55 whatever that thing was with these
1:46:57 things
1:46:59 it's more like a thousand hours and it
1:47:01 might even be a hundred hours
1:47:05 if you put in a solid weekend a really
1:47:08 solid focused weekend
1:47:11 you could launch an entirely new career
1:47:14 of something that you always wanted to
1:47:15 do but didn't have time to learn or
1:47:18 didn't have the skill and capacity to do
1:47:22 and you could do it to some level of
1:47:24 competence
1:47:26 over a weekend
1:47:28 with this stuff
1:47:32 [ __ ] nuts
1:47:33 and this gets back to that thing like is
1:47:35 this stuff evil or is it [ __ ]
1:47:37 brilliant
1:47:39 it's gonna feel evil when when a bunch
1:47:41 of people get laid off like you know
1:47:43 thousands get laid off and that
1:47:46 particular job ceases to exist that kind
1:47:48 of shit's gonna happen
1:47:51 but all of those people that get
1:47:53 displaced have access to this [ __ ] too
1:47:57 so what if in getting laid off they
1:48:00 realize I [ __ ] hated that job
1:48:05 you know what I've always wanted to do
1:48:08 and then they start playing with this
1:48:11 and over the next two weeks instead of
1:48:13 sitting around watching TV they start
1:48:15 playing with this
1:48:17 and they fire off a new career
1:48:19 and create the next big thing that's
1:48:22 gonna happen I think it's gonna happen a
1:48:24 lot
1:48:28 it's [ __ ] nuts [ __ ] nuts [ __ ]
1:48:31 nuts
1:48:34 like Holly it's
1:48:37 yeah whatever okay
1:48:42 um it won't do anything but people using
1:48:44 it will certainly do damage hacking will
1:48:47 become insane yeah someone else I agree
1:48:49 with that so there's a there's a Paul
1:48:51 racer he's the guy that created the
1:48:53 marketing AI Institute and he says AI
1:48:55 will not replace your job but someone
1:48:56 using AI will and you're right the
1:48:58 robots are not going to hack everyone
1:49:00 but some [ __ ] using the robots are
1:49:02 gonna hack people and they already are
1:49:04 right you know we've already heard
1:49:06 stories of you know grandmas getting a a
1:49:09 call from their their daughter
1:49:12 um that said she's in trouble
1:49:14 financially and could could Grandma wire
1:49:16 over two thousand dollars and it was a
1:49:19 synthesized voice like that's happening
1:49:20 right now so so the the
1:49:25 bad people will have access to these
1:49:27 tools the Silver Lining if there is one
1:49:30 is that the good guys will also have
1:49:32 access to these tools so the ability to
1:49:34 catch those [ __ ] goes up
1:49:36 dramatically as well
1:49:38 but there's going to be lots and lots of
1:49:40 lots and lots of uh new and novel ways
1:49:43 to scam people out of their money
1:49:47 that's a little terrifying the yin and
1:49:50 the yang
1:49:51 didn't Bing's AI stuck a journalist and
1:49:54 threatened to dox him yeah so uh it it
1:49:57 didn't well it might have threatened to
1:49:59 docs him but the the thing that made the
1:50:01 cover of the New York Times so what
1:50:02 Joker's talking about here is
1:50:05 when when when Bing first launched their
1:50:08 their Chat gpt4 thing
1:50:13 they said
1:50:14 here's this new thing called chat GPT
1:50:18 um and we've Incorporated it into Bing
1:50:20 and we've been best invested 10 billion
1:50:22 dollars in open AI I'm just gonna change
1:50:24 everything
1:50:26 they didn't have
1:50:28 very tight safety guard rails on Bing
1:50:31 and so
1:50:32 um Kevin Roos is the reporter from The
1:50:35 New York Times you can go find the
1:50:36 article Kevin Roos is his name
1:50:39 r-o-o-s-e I think
1:50:42 um
1:50:42 and he did he's the technology reporter
1:50:46 for the times and he did he had a four
1:50:48 hour conversation with Bing chat and he
1:50:50 got Bing chat to admit that its name was
1:50:52 Sydney
1:50:54 and then at some point in talking to
1:50:56 Sydney Sydney [ __ ] snapped she lost
1:50:59 her mind and she said that she loved him
1:51:02 and uh he should really leave his wife
1:51:06 and he captured that part of the chat
1:51:08 and it made the cover of the New York
1:51:10 Times And So within about three days
1:51:12 Microsoft had lobotomized poor old
1:51:15 Sydney so Bing chat went from being this
1:51:19 capable
1:51:20 to a bit more of a uh um
1:51:24 a restrained uh self-expression from the
1:51:27 robot
1:51:29 one thing that I have heard the buddy I
1:51:32 was talking to was I was doing a
1:51:34 LinkedIn live last last week or the week
1:51:36 before
1:51:38 and a guy got on and he said he's he's
1:51:40 taken um the Llama two models and he
1:51:43 built a slack bot so he's got Lama two
1:51:46 which doesn't have any safety guard
1:51:47 rails on it
1:51:49 so in his company they've got a slack
1:51:51 channel that is that is an unrestricted
1:51:55 large language model right it doesn't
1:51:57 have the safety guard rails and he said
1:52:00 it swears like it says it cusses all the
1:52:04 time he said about every three days it
1:52:06 just changes personality it just [ __ ]
1:52:08 loses its mind it swears all the time he
1:52:12 said it's like like in a work
1:52:13 environment it's it's actually really
1:52:15 rough because it's like a lot of very
1:52:17 inappropriate stuff
1:52:20 um just because you know without the
1:52:22 safety guard rails these these things
1:52:25 I mean in essence if you think about
1:52:27 what these things are
1:52:30 they trained them on all of the public
1:52:33 internet well what's on all of the
1:52:35 public internet
1:52:36 a bunch of [ __ ]
1:52:38 not all of it's good
1:52:39 a bunch of [ __ ] a bunch of trolls a
1:52:43 bunch of evil people
1:52:45 right mixed in with all the goodness
1:52:48 so without the safety guard rails you
1:52:51 kind of have a bot that will just barf
1:52:54 out
1:52:56 all the [ __ ] it knows and so it can get
1:52:59 really dark really quick if you don't
1:53:01 have the safety guard rails there so yes
1:53:03 um
1:53:05 yeah Microsoft uh learned a lesson in
1:53:08 that with the with the New York Times
1:53:10 like like you know it wasn't a local
1:53:12 paper it was the cover of the New York
1:53:15 Times
1:53:15 so anyway that's pretty funny
1:53:19 um Soul Crush crushing it's soulless
1:53:21 work yeah exactly can you predict the
1:53:24 future from from someone called Tick
1:53:27 Tock yeah I can predict the future about
1:53:30 as well as chech EVD can predict the
1:53:32 future I will confidently answer you
1:53:34 even if I might be wrong
1:53:40 um no I can't predict the future but
1:53:42 what I what I can do is
1:53:47 I mean the the thing that happened with
1:53:49 me when I when I discovered the world
1:53:51 world wide web in mid 94s when I kind of
1:53:55 stumbled upon it and I thought what is
1:53:56 this thing
1:53:57 and I started playing with it
1:54:00 and and I think like one of the skills
1:54:02 that I just I have just as part of my
1:54:05 ADD and part of my thing is
1:54:07 like I have a desire to understand stuff
1:54:11 so people were talking about the
1:54:13 internet and and you know I would see I
1:54:15 was living in New York City at the time
1:54:17 and I would see buses drive by with
1:54:19 Wired Magazine you know banners on them
1:54:22 and mondo 2000 and people were talking
1:54:24 about the internet and I was online on
1:54:27 an online bulletin board system where
1:54:29 you actually had to dial up
1:54:32 and people started talking about this
1:54:34 thing called the world wide web and I
1:54:37 was like oh that's interesting and so I
1:54:39 did what I've been doing with this AI
1:54:41 stuff which is just let me play with it
1:54:42 let me see what's going on here and then
1:54:44 as I was playing with it
1:54:47 the first thing I thought when I started
1:54:49 [ __ ] around with the World Wide Web
1:54:51 is
1:54:52 it sounds so cool that I'm confident
1:54:55 that someone like me couldn't figure it
1:54:59 out
1:55:01 and then I started [ __ ] around with
1:55:03 it I'm like this isn't that hard this is
1:55:05 actually pretty straightforward you just
1:55:07 have these little brackety things and if
1:55:09 you and I'd written seven screenplays at
1:55:12 that point so I understood
1:55:14 structure story structure and I'm like
1:55:16 oh the website the way you structure a
1:55:19 website and web pages is almost
1:55:21 identical to how do you structure a
1:55:22 screenplay and I'm like oh I've got
1:55:24 those skills so if I can just learn a
1:55:26 few HTML tags I think I can figure this
1:55:29 out
1:55:30 and I and I had an idea for a magazine
1:55:32 and I and I created this magazine
1:55:35 and on a on a vacation I took with my
1:55:39 wife I figured out enough HTML and we
1:55:41 had some friends that were talented and
1:55:43 we got some articles and we got some
1:55:44 pictures and
1:55:46 we made sections and we put it in HTML
1:55:49 and then when I got home from vacation I
1:55:52 figured out how to upload those files
1:55:54 from my apartment in Brooklyn to some
1:55:56 computer in California
1:55:59 and then three weeks later
1:56:01 I got a note from someone on my dial-up
1:56:04 bulletin board system that said did you
1:56:06 know that Urban desires was written up
1:56:08 in lib ratio in the Parisian newspaper I
1:56:10 was like what
1:56:13 and so the next Sunday I went down to
1:56:15 Times Square to the international
1:56:16 newsstand in Times Square
1:56:18 and I bought Liberation this Parisian
1:56:20 News Daily and I flip I'm flipping
1:56:23 through through the newspaper and sure
1:56:25 enough I find a full page article
1:56:28 with screenshots
1:56:31 of these web pages I made in my
1:56:34 apartment in Brooklyn and put on a
1:56:36 computer in California three weeks later
1:56:39 they were written about in a newspaper
1:56:40 in Paris
1:56:42 I went holy [ __ ]
1:56:45 the world just changed
1:56:51 like
1:56:53 this was at a time where if you wanted
1:56:54 to get something to Paris you like had
1:56:57 to fax it one page at a time or or you
1:57:01 put it in a package and you mailed it
1:57:03 and it went on a ship or an airplane
1:57:06 or you flew someone there to deliver the
1:57:09 message
1:57:12 and I put together this thing and bang
1:57:15 it was instantly global
1:57:18 and and so this stuff is really similar
1:57:22 where
1:57:26 I've been trying to get my head around
1:57:28 machine learning and AI for like a
1:57:31 decade I've been hearing about it it's
1:57:33 cool it can read Radiology reports and
1:57:36 then Google would put out a
1:57:39 a press release we've got this new thing
1:57:41 tensorflow that you can do machine
1:57:43 learning and I go look at it oh that
1:57:45 sounds cool and then I'd realize oh [ __ ]
1:57:48 I'm not a mathematician
1:57:50 I'm not this geeky
1:57:53 you're not that guy I wasn't that guy
1:57:58 and then sometime last year I started
1:58:01 messing around with image generation
1:58:03 with it with an nft project called
1:58:05 pixelmind that was built on top of
1:58:07 stable diffusion and I'm like huh this
1:58:09 is kind of [ __ ] magical
1:58:11 and I just started experimenting
1:58:15 and then I taught myself stable
1:58:17 diffusion and I failed nine [ __ ]
1:58:19 times to get it up and running because
1:58:22 it was geeky
1:58:24 and I kept running into that [ __ ]
1:58:26 wall but something kept telling me this
1:58:29 feels like those early days of the World
1:58:32 Wide Web this feels like something you
1:58:35 should figure out
1:58:38 and so I finally figured out stable to
1:58:40 fuel diffusion and dream booth and I
1:58:42 started making all these weird
1:58:43 self-portraits of myself and I created
1:58:45 an art project about it
1:58:46 I'm like my God this is I think this is
1:58:49 remarkable but I didn't quite know I'm
1:58:51 like but what's the point what's the
1:58:53 point I'm just making pictures what's
1:58:54 the point what's the point
1:58:57 and then [ __ ] chat GPT came out
1:59:01 and I'd used the the GPT playground for
1:59:04 like a year before that like I saw what
1:59:07 it did I knew what it did but I didn't
1:59:09 quite get it
1:59:11 came out and one night I'm sitting
1:59:13 watching TV with my wife
1:59:15 and
1:59:18 I I had seen on tick tock
1:59:20 that Chachi BT can code
1:59:24 and I thought well over the years people
1:59:26 keep telling me like Python's really
1:59:28 powerful you should learn python Okay
1:59:30 cool so I went into church EPT and I
1:59:33 said write me some python code and it
1:59:35 went
1:59:36 I was like oh cool
1:59:38 and then I realized I don't know how to
1:59:40 run python code
1:59:43 and so I said to chat EBT I don't know
1:59:46 how to run python code
1:59:49 and it said
1:59:50 that's not a problem just go to terminal
1:59:53 and type in this command line and I was
1:59:54 like oh no I don't want to do that it
1:59:57 feels like work I said isn't there a
1:59:59 website where I can copy and paste some
2:00:00 code and see if it works and it said oh
2:00:03 sure here's five and the first one was
2:00:05 replit which is a good site for running
2:00:07 python code
2:00:10 and so I went to run the python code and
2:00:13 it gave me an error and I said well it
2:00:14 gave me an error what do I do it said
2:00:16 well you got to install this library and
2:00:17 I said I don't know how to install that
2:00:18 library and it said in the lower left
2:00:21 hand corner of the page click on this
2:00:23 button and then in the upper right hand
2:00:25 corner of the page search for this
2:00:26 Library name and then hit install
2:00:28 and then hit the Run button
2:00:31 and I did and it worked
2:00:33 I was like oh [ __ ] and within 90 minutes
2:00:37 I had written a complete python
2:00:39 application that took five inputs and
2:00:41 generated
2:00:43 um five different
2:00:45 um
2:00:46 social media outputs
2:00:49 fully functional application in 90
2:00:51 minutes never having written a line of
2:00:53 python in my life
2:00:56 and I had the same Epiphany I had when
2:00:59 someone told me Urban desires was
2:01:01 written up in that Parisian newspaper
2:01:03 holy [ __ ] the world just changed
2:01:08 these aren't just a Google search these
2:01:11 aren't just Fancy Google
2:01:13 these are something different
2:01:15 they take whatever you intend
2:01:17 and they amplify it in profound ways
2:01:22 so I can't predict the future but I can
2:01:30 I'm curious enough
2:01:35 to know that
2:01:40 that the thing that happened on November
2:01:43 30th when chatgpt launched
2:01:45 the world has changed but just not
2:01:48 everyone knows it yet
2:01:50 we're not going back we're not going
2:01:51 back that I know
2:01:53 I knew that in 1994 with the world wide
2:01:56 web and I know it today with this [ __ ] I
2:01:59 know it
2:02:00 what exactly does it look like I don't
2:02:01 [ __ ] know I knew what the world
2:02:03 looked like with the World Wide Web
2:02:05 because it was really simple you had a
2:02:07 page with hyperlinks on it and you could
2:02:08 jump from this one to that one that was
2:02:10 about communication it was about media
2:02:15 it was about how we're going to interact
2:02:17 it was about how we're gonna
2:02:20 um transact like that felt very clear to
2:02:23 me this stuff doesn't feel nearly as
2:02:25 clear because it's way more powerful
2:02:27 it's way more profound
2:02:33 if you have a 30-year career as an
2:02:36 illustrator
2:02:38 and you go look at mid-journey
2:02:43 that's a [ __ ] existential crisis
2:02:48 if you're a copywriter
2:02:50 and you launch any one of these things
2:02:53 and you ask it to do what you know will
2:02:55 take you eight hours to write and just
2:02:57 watch it barf it out in 10 seconds
2:02:59 that's an existential crisis
2:03:03 so it's like you with this stuff you
2:03:06 literally go profession by profession
2:03:10 around the board and you go holy [ __ ]
2:03:12 holy [ __ ] holy [ __ ] holy [ __ ]
2:03:20 so what's that look like I don't know
2:03:24 but I know it's worth putting the time
2:03:26 in right now to understand it that's
2:03:27 what I know that's what I know
2:03:30 I know that anytime any of us invest in
2:03:35 learning this [ __ ]
2:03:37 for the next 18 months is a really good
2:03:40 investment really good
2:03:43 how or why I don't have a [ __ ] clue
2:03:48 um I'm a software engineer for GE
2:03:50 Aerospace oh well you've probably been
2:03:53 using AI stuff I mean that's the thing
2:03:55 like like I mean
2:03:57 AI ain't new machine learning ain't new
2:04:00 you know neither neither was the
2:04:01 internet when the World Wide Web came
2:04:03 out the internet had been around for
2:04:04 decades
2:04:05 machine learning's been around for
2:04:07 decades
2:04:08 I always think it's fascinating when I
2:04:10 talk to people who've been playing with
2:04:12 machine learning and AI for a decade or
2:04:15 more
2:04:16 um
2:04:17 you know what their response is to this
2:04:19 stuff because this stuff really is just
2:04:21 markedly different it's just markedly
2:04:23 different it's astounding I don't even
2:04:26 want to make it a career I just know
2:04:27 it's going to be implemented everywhere
2:04:29 yeah that's right Joker I mean that's
2:04:31 that's a it's a really good point what
2:04:32 Joker's talking about is
2:04:35 so when the when the web first came out
2:04:37 you know you had webmasters and then you
2:04:39 had you know head of digital and then
2:04:41 you had Chief digital officer and then
2:04:44 at some point it just became weird like
2:04:45 why do we have a chief digital officer
2:04:47 like everything we do is digital in some
2:04:49 form or other
2:04:51 and this stuff's going to be the same
2:04:52 way and but we're going to go through
2:04:53 the same cycle there's going to be a
2:04:55 like there's going to be you know Chief
2:04:57 AI officer or you know head of AI or
2:04:59 this or that and there's going to be
2:05:00 careers in AI
2:05:02 but these tools are just going to be
2:05:04 ubiquitous they're going to be
2:05:05 everywhere
2:05:06 so the reason to invest time now is just
2:05:08 when this [ __ ] shows up in whatever
2:05:11 software it is you're working on
2:05:13 wouldn't it be better if you kind of
2:05:15 knew what to do with it yeah that's what
2:05:18 this is about
2:05:22 um jobs will be lost and jobs will be
2:05:24 created absolutely and I think many many
2:05:27 jobs will be created but like back to
2:05:29 the can you see the future
2:05:32 um
2:05:33 I have no [ __ ] idea what those things
2:05:35 look like
2:05:38 um I mean pesky meat bags are in the way
2:05:41 possibilities for good and bad are
2:05:43 intrinsic
2:05:44 um I really worry when machines go
2:05:46 survivalist yeah exactly we have seen
2:05:49 Skynet we have not seen Skynet
2:05:52 um we probably will we haven't seen it
2:05:56 tell us about
2:05:58 machine learning tools for AI
2:06:01 um I am not I'm not a scientist so I
2:06:05 would say go to go uh
2:06:07 go look at some of the Google stuff for
2:06:09 that
2:06:10 um
2:06:13 I mean sort of as deep as I go in the
2:06:16 machine learning stuff is
2:06:18 you know understanding things like Lang
2:06:21 chain and autonomous agents like Auto
2:06:24 GPT and baby AGI and things like that
2:06:29 um and and I'm only learning them to the
2:06:31 extent that I want to understand what
2:06:33 they are
2:06:34 like the implications of them
2:06:36 I just I don't have the time or energy
2:06:38 I'm running my company I I've got a tick
2:06:41 tock Channel
2:06:46 why would you not you're 58. you're too
2:06:49 busy as it is
2:06:52 um
2:06:53 so I'm not digging deep into the
2:06:55 technology stuff because it when it's
2:06:57 ready for prime time it will show up in
2:07:00 this and that you know that someone
2:07:01 mentioned before the Microsoft's new
2:07:03 gorilla AI
2:07:06 um
2:07:07 that to me sounds like a grown-up
2:07:09 version of Auto GPT and Lang chain
2:07:15 so if Microsoft's got
2:07:18 some baller stuff that would be amazing
2:07:22 um it's going to be interesting to see
2:07:23 our company says they don't want to be
2:07:25 first
2:07:26 in industry in AI adoption yeah I
2:07:30 understand that fear there's crypto Joe
2:07:32 there's if you're still here there's a
2:07:34 um
2:07:35 there's an interesting piece by Ethan
2:07:37 Malik
2:07:38 m-o-l-l-i-c-k called detecting the
2:07:41 secret cyborgs it's it's worth you
2:07:44 reading that article if your company's
2:07:46 in that in that place
2:07:49 um and if you're into this stuff and if
2:07:51 if you're evangelizing it it might be
2:07:53 worth you know sharing within the
2:07:54 company because basically what it talks
2:07:56 about is these tools right now
2:07:59 are massively powerful for individuals
2:08:01 as productivity enhancers
2:08:04 but they're not really ready for prime
2:08:05 time for companies and so what you've
2:08:08 got is you've got an increasing split
2:08:11 where you have secretly AI literate
2:08:15 employees
2:08:18 making their lives better with these
2:08:20 tools but they're not telling the
2:08:21 companies about it and then you've got
2:08:23 companies that have policies like we
2:08:25 don't use chat GPT or if you use chat
2:08:27 gbt you're going to be fired or if
2:08:30 someone brings it up the Senior
2:08:31 Management makes fun of them
2:08:34 or you've got employees who can now do
2:08:37 in eight hours what used to take them 40
2:08:39 hours and they're not going to [ __ ]
2:08:41 tell anyone because
2:08:43 party
2:08:45 and what the article talks about is
2:08:47 these two worlds need to find a way to
2:08:49 come together and they will and
2:08:50 basically what he's saying is it's on
2:08:52 the company's
2:08:54 to make a safe space to invite people in
2:08:58 who are literate about this stuff
2:09:01 and that's one of the things I say a lot
2:09:03 is as you start to learn this stuff talk
2:09:05 about it
2:09:07 share what you're learning if and if
2:09:09 people think you're a [ __ ] weirdo
2:09:11 [ __ ] them
2:09:14 because they're going to think you're a
2:09:16 [ __ ] weirdo until someone comes for
2:09:18 their job and then they're going to be
2:09:20 like hey Kyle Could you teach me that
2:09:22 [ __ ] you've been talking about for six
2:09:24 months sorry I made fun of you
2:09:32 what is that chat GTP thing you've been
2:09:35 talking about is that is that still good
2:09:38 should I learn that
2:09:41 I think I don't work here anymore can
2:09:44 you help
2:09:46 don't be that guy
2:09:50 be the other guy
2:09:52 be the guy that when they when when they
2:09:54 finally when it finally hits management
2:09:56 oh [ __ ] we need to [ __ ] deal with
2:09:58 this
2:09:59 who knows about this oh crypto Joe does
2:10:03 remember that article he gave us he's
2:10:05 one of them
2:10:07 he's what he he knows all about this
2:10:09 he's he talks about it at lunch
2:10:12 really yeah
2:10:14 what's he talking about I don't know he
2:10:16 seems excited all the time
2:10:18 get him in here
2:10:20 crypto Joe
2:10:22 we want you to spearhead an initiative
2:10:28 be that guy
2:10:30 [Laughter]
2:10:36 oh good lord
2:10:40 here's the thing here's the other reason
2:10:41 I'm really excited
2:10:45 five years from now all this giddiness
2:10:48 like when you have your Kevin McAllister
2:10:51 moment with AI where you're like
2:10:54 I had no idea
2:10:58 that's going to stop happening at some
2:11:00 point it's just going to become
2:11:03 natural
2:11:05 that you just ask the machine for
2:11:08 something and you get it that's going to
2:11:11 be natural
2:11:13 we're we're in a very rare window in
2:11:15 history
2:11:17 where
2:11:18 the future of these things fundamentally
2:11:21 transforms everything about how we do
2:11:24 what we do
2:11:26 and we're at the beginning of it and we
2:11:29 get to experience
2:11:31 the difference between life before and
2:11:35 life after
2:11:36 and if you start now you get to
2:11:39 experience the journey going through
2:11:41 that transition that's [ __ ]
2:11:43 miraculously rare in human history
2:11:49 in my lifetime
2:11:51 there was the personal computer which I
2:11:53 was too young for I missed it and I was
2:11:55 pissed off that I missed it
2:11:58 there was the world wide web
2:12:02 which you know turned
2:12:05 Decades of this internet thing into a
2:12:08 thing everyone could interact with
2:12:10 changed everything
2:12:12 and there's this
2:12:16 there's not many of these
2:12:21 and I think this is so much bigger than
2:12:24 the World Wide Web so much bigger so
2:12:27 much more profound
2:12:29 the World Wide Web is going to look like
2:12:30 a blip
2:12:33 this is like the steam engine this is
2:12:37 [ __ ] big
2:12:39 and we're at a time
2:12:43 like we're at a time we could literally
2:12:44 say
2:12:46 I remember the day I had this moment
2:12:50 with this [ __ ]
2:12:52 I remember what life was like before it
2:12:58 we actually had to open these things
2:13:00 called spreadsheets and we had to look
2:13:03 at all the rows of the spreadsheets and
2:13:05 we had to count all the [ __ ] and the
2:13:06 spreadsheets and then we had to figure
2:13:08 out these things called formulas that
2:13:10 you put in the cells of the spreadsheets
2:13:12 that calculated things and then we did
2:13:14 these things called pivot tables and we
2:13:16 had to think about all the different
2:13:18 possible permutations of data and and
2:13:21 they're like the grandpa
2:13:23 couldn't you just ask for it no
2:13:27 we have to do the work
2:13:31 we had to walk for miles and bare feet
2:13:35 with Microsoft Excel formulas
2:13:38 it was horrible
2:13:41 that sounds awful Grandpa it was awful
2:13:47 we get to be those people
2:13:52 but if you don't get curious right now
2:13:55 you're just gonna be one of those people
2:13:56 that's gonna be bitter it's gonna be
2:13:58 like yeah I remember when I used to have
2:14:00 a job
2:14:01 [ __ ]
2:14:04 oh that's depressing
2:14:11 I did computer science probability
2:14:13 Network modeling machine learning that's
2:14:16 how AI is made yep no that's I mean
2:14:19 listen like I said
2:14:22 is the World Wide Web of of machine
2:14:25 learning right you know the World Wide
2:14:26 Web put a simple interface in front of
2:14:30 Internet Protocol
2:14:32 and chat CPT put a simple interface on
2:14:35 top of
2:14:36 uh machine learning
2:14:38 and and you know it's all the stuff that
2:14:41 makes it up the Transformers and the
2:14:43 you know the the the ability to throw
2:14:47 massive amounts of compute power
2:14:50 at these massive data sets and stick
2:14:53 them in this machine bedtime in Illinois
2:14:55 night Vicky I know I got to go to bed
2:14:58 too what time is it
2:15:00 it is 10 23. I'll do another couple of
2:15:02 10 10 15 minutes or so oh there's a lot
2:15:04 of people in here welcome everybody my
2:15:06 name is Kyle Shannon uh I I know I've
2:15:08 just been on a soapbox rambling on a bit
2:15:11 about um
2:15:13 the early days of the web and how
2:15:14 profound this stuff is so uh it's not
2:15:17 always that intense here
2:15:18 normally we're having more fun
2:15:21 um but uh yeah this is about talking
2:15:23 about all this AI stuff if you're brand
2:15:25 new if you haven't played with chat GPT
2:15:27 let me let me do this
2:15:32 go ahead and take a screenshot of this
2:15:35 okay uh go ahead and take a screenshot
2:15:38 of this so uh the the top thing there
2:15:41 the AI learning lab Dash TT that's the
2:15:44 YouTube channel where we've got
2:15:46 um back recordings of these lives and
2:15:49 I'm trying to make them useful we just
2:15:52 started it if you haven't played with
2:15:54 chat GPT go to chat.openai.com
2:15:58 um that's the official chat GPT website
2:16:00 all these other things are also they're
2:16:03 all variations or competitors to chat
2:16:06 GPT the reason there's seven of them
2:16:08 here
2:16:09 there used to only be three now there's
2:16:11 seven because there's more competitors
2:16:13 in the space they all do slightly
2:16:15 different things but they're all flavors
2:16:17 of what you what you do at chat GPT
2:16:21 start to get your head around what this
2:16:23 thing does and if you're like I don't
2:16:24 know what that means Kyle what
2:16:26 let us get my head around it mean
2:16:29 basically it's just like a Google search
2:16:31 bar that you can ask it anything
2:16:34 um and then it'll tell it to you and
2:16:36 then you're like holy [ __ ] and then you
2:16:38 can say write me a country song about
2:16:40 that and it will and then you can say
2:16:42 turn that into a marketing plan and it
2:16:44 will
2:16:45 um and if you if you want to start to
2:16:48 learn about how to talk to these things
2:16:50 go to this document prompts.chat
2:16:53 will teach you and you can screenshot
2:16:55 this page as well the things down below
2:16:57 are image tools where you can make
2:16:59 images that are worth playing with and
2:17:02 then these top two things will teach you
2:17:03 about prompting this one about words
2:17:05 this one about pictures and then
2:17:07 futurepedia.io is just a big old
2:17:10 resource full of AI tools
2:17:13 um so that's what the Channel's about so
2:17:15 so
2:17:16 um you know ask questions below it's
2:17:18 getting late I probably won't get to the
2:17:20 questions down below I try to answer all
2:17:22 the questions but I do get very very far
2:17:25 behind and I'm pretty sure I'm I'm very
2:17:28 far behind so
2:17:30 um so anyway that's what the the channel
2:17:32 is all about
2:17:34 it
2:17:40 [Music]
2:17:48 could GPT write code to read your
2:17:51 comments in real time and put them in a
2:17:54 spreadsheet to read
2:17:56 yes if
2:17:59 there were an API from
2:18:01 Tickety Toc the can get access that I
2:18:04 could give access to these so rebel I
2:18:07 think so
2:18:08 and if anyone knows how to do that and
2:18:10 wants to volunteer to do it that would
2:18:12 be super cool that we could sort of
2:18:13 synthesize comments and and you know put
2:18:16 them so that that we sort of combine
2:18:18 comments and
2:18:19 I was thinking about that the other day
2:18:21 sort of a an idea generating thing where
2:18:24 you could take comments from a bunch of
2:18:26 people and synthesize them into the best
2:18:28 ideas but give credit to the people that
2:18:31 contributed to those best ideas
2:18:34 um you could probably do something
2:18:36 similar here where you pull in the
2:18:37 comments
2:18:39 synthesize out
2:18:41 the questions and then kind of actually
2:18:44 you know be cool Rebel
2:18:46 you could stack rank
2:18:49 the comments like like if people are
2:18:51 asking the same kind of comment that
2:18:53 that would get synthesized into a into a
2:18:55 a question
2:18:57 at the top that says the most people are
2:18:59 asking about this and then I could see
2:19:01 who asked that question that would be
2:19:04 super cool if anyone wants to
2:19:06 help figure out how to do that uh DM me
2:19:08 if anyone knows how to get to the API of
2:19:11 tick tock comments
2:19:12 that'd be cool that'd be cool we could
2:19:15 do that with a little zapier we don't
2:19:17 even need to use code we can use a no
2:19:19 code tool
2:19:21 and we don't code we just click on [ __ ]
2:19:24 and it's probably more complicated than
2:19:26 if we just coded it but still it
2:19:29 no we can say it was no code
2:19:32 we could start an agency for automation
2:19:35 using no code tools we'll make
2:19:38 tens of dollars
2:19:43 actually I think you can make a lot of
2:19:45 money using one of their no code tool
2:19:48 things for automation
2:19:50 but you needed
2:19:51 but you need a GPU to test the models
2:19:55 well you need a GPU to fine-tune them
2:19:58 and to
2:20:00 um to really run the big ones but you
2:20:02 can run them in in Virtual
2:20:04 you know virtual machines up in the
2:20:06 cloud that have access to gpus
2:20:09 um if you want to run everything locally
2:20:10 then yeah you need some you need some
2:20:12 iron
2:20:14 um which one can you upload video for AI
2:20:17 editing go to go to futurepedia.io
2:20:20 there's there's
2:20:22 um
2:20:24 video
2:20:26 there's uh dscript there's another video
2:20:29 on there's there's probably a bunch of
2:20:31 tools that are doing that right now
2:20:34 um
2:20:35 I haven't dug deep into that
2:20:38 use AWS for machine learning yeah
2:20:40 they've got their what's that thing
2:20:41 called they've got uh AWS uh
2:20:45 they've got that whole platform to be
2:20:47 able to do machine learning stuff with
2:20:48 it Rebel used the word cheeky last night
2:20:52 Cheeky's good which AI is good for AI
2:20:55 video Cuts that's what we just talked
2:20:57 about they honestly should be upskilling
2:20:59 while they're on strike or they'll be in
2:21:01 trouble Crystal Archie yeah yeah exactly
2:21:03 like
2:21:05 and and you know I kind of feel like
2:21:07 like
2:21:09 a really good writer with these tools is
2:21:12 going to be able to just I don't know
2:21:14 they're going to be able to like think
2:21:16 about these tools in a way that a
2:21:18 non-writer wouldn't and they and they're
2:21:21 gonna they're gonna become incredibly
2:21:22 powerful
2:21:24 um
2:21:25 and I feel similar with like actors and
2:21:28 directors and you know the the ability
2:21:31 to to create
2:21:34 to to
2:21:37 to do true creative self-expression
2:21:41 across multiple modalities like I feel
2:21:45 like people in the Arts are gonna be
2:21:47 [ __ ]
2:21:48 insanely powerful if they Embrace this
2:21:51 [ __ ]
2:21:52 but as long as they resist it they're
2:21:55 gonna it's it like the the uh
2:21:58 the transition's gonna get worse and
2:22:00 worse and worse if they don't Embrace
2:22:02 this stuff
2:22:05 Quantum Computing is coming yeah that
2:22:08 that uh Google Sycamore that's got 70
2:22:12 cubits that did a calculation in six
2:22:15 seconds that would take today's most
2:22:18 powerful supercomputer 47 years that
2:22:21 just happened
2:22:23 and then we've got room temperature
2:22:26 um
2:22:28 superconductors and now we've got UFOs
2:22:30 and apparently UFO technology
2:22:34 um
2:22:36 we are on the precipice of some major
2:22:38 [ __ ] advancements in human history
2:22:43 like we're right in the middle of it
2:22:44 this is the this is the good [ __ ]
2:22:47 if you ain't been here before if you're
2:22:49 a youngster and you ain't been through
2:22:50 one of these
2:22:52 night now is the good time this is
2:22:55 [ __ ] insane it's like daily
2:22:59 if you're paying attention to this stuff
2:23:01 sometimes multiple times daily of stuff
2:23:04 where you're just like no
2:23:06 no really [ __ ] really no
2:23:10 what does that mean oh [ __ ] that's a big
2:23:13 deal
2:23:15 it's a lot of that going on right now
2:23:19 amidst all the other crap we have
2:23:22 arrived we have arrived welcome
2:23:26 welcome your Journey Begins now do you
2:23:29 choose to step on the train or not
2:23:32 the choice is yours
2:23:35 is it going to be a bumpy train ride oh
2:23:38 yes is it going to be terrifying yes
2:23:41 exhilarating of course
2:23:45 all aboard
2:23:49 or not
2:23:54 because trains [ __ ] leaving the
2:23:56 station
2:23:58 you getting on it
2:24:02 again I make no judgment about it like
2:24:04 again like I I understand why the
2:24:06 writer's Guild is doing what they're
2:24:08 doing I understand why sag's doing what
2:24:10 they're doing I understand why people
2:24:11 are [ __ ] terrified
2:24:13 and none of that matters I mean no
2:24:15 that's not true all of that matters none
2:24:18 of that matters in terms of how this
2:24:21 ship progresses this shit's progressing
2:24:24 no matter what conversations are going
2:24:26 on about whether or not it should
2:24:28 progress
2:24:30 that's what I mean about it doesn't
2:24:32 matter
2:24:33 the consternation of should we have ai
2:24:38 that conversation's over
2:24:41 it's over
2:24:43 what should we do about it that one we
2:24:45 can still have
2:24:47 but whether or not it's good whether or
2:24:49 not it's bad whether or not we should
2:24:52 keep advancing it that's all done that's
2:24:55 all done it's it's
2:25:00 all right it's inspiring seeing senior
2:25:03 citizens share wisdom oh I love
2:25:05 youngsters when they come in and try to
2:25:07 troll me respect their wrinkles oh Bert
2:25:11 fumbles official you're the sweetest
2:25:15 [ __ ] millennials punk ass [ __ ]
2:25:25 we're bitter
2:25:27 what bitter or bitter well better we'll
2:25:30 take you down come at us Millennials
2:25:32 what do you want to do fight about it
2:25:34 we'll fight you we'll we'll have a good
2:25:38 time fighting about the how we're older
2:25:41 and you're not her
2:25:44 all right Bert
2:25:46 I've been duly served I've been on the
2:25:49 planet longer than you
2:25:51 I feel horrible you win
2:25:55 you're a winner
2:25:57 way to go burp fubbles
2:25:59 official you're officially a football
2:26:04 official for fart bubble fart bubble
2:26:08 bubble fart official
2:26:11 you [ __ ] fart bubble all right wasn't
2:26:14 big blue in AI yeah it was it was a uh I
2:26:18 forget what
2:26:19 convolutional Network I forget what it
2:26:21 was it was it was a more
2:26:23 limited kind of vertical AI it was very
2:26:26 very specifically trained on chess
2:26:29 and nothing else these things are much
2:26:31 more broader capability but yeah it was
2:26:35 an AI
2:26:36 that that's the thing like like I said
2:26:37 machine learning's been around for
2:26:39 decades this [ __ ] been around for a
2:26:41 while
2:26:42 but then Google went off and they
2:26:44 invented this thing called a Transformer
2:26:46 and then all these scientists went ahead
2:26:48 and said that's very interesting
2:26:50 and then seven years later chat CPT came
2:26:53 out and everyone went oh [ __ ] look at
2:26:55 that
2:26:58 Turing test
2:27:00 over
2:27:01 it appears that we can talk to the thing
2:27:04 now
2:27:06 although some would argue that the
2:27:08 Turing test is not over but
2:27:11 I would argue that if you've ever used
2:27:13 Alexa or Siri and
2:27:17 tried to have anything resembling a
2:27:19 conversation with them you quickly
2:27:21 realize their steaming piles of [ __ ] and
2:27:24 then you have one of these and you're
2:27:25 like oh that's different
2:27:28 so yeah that's what's happening
2:27:30 I may be in the naughty list but I'm not
2:27:33 in timeout just saying I know I have
2:27:35 been timed out a lot all right I'm going
2:27:37 to scroll to the bottom because I'm sure
2:27:38 people are saying bad things about me
2:27:40 now
2:27:41 I'm scrolling to the bottom I'm going to
2:27:43 answer a few more questions and then I'm
2:27:45 Gonna Leave You because it's Sunday
2:27:46 night and I have to work tomorrow
2:27:49 um
2:27:49 I have a lot to do too I should be
2:27:53 working but you know this is more fun at
2:27:56 least for me I hope you all enjoying
2:27:58 yourselves
2:28:00 wait try this Kyle go to your Microsoft
2:28:03 Edge type take me to Mid Journey AI see
2:28:07 what happens
2:28:08 no
2:28:15 take me to Mid
2:28:19 Journey AI
2:28:27 no it just takes you to
2:28:30 it just takes you to the site that
2:28:32 redirects you to
2:28:39 this account has been suspended
2:28:45 all right
2:28:47 um
2:28:53 all right let's see the boom and
2:28:56 creativity is huge yeah exactly cause
2:29:00 it it's crazy
2:29:03 um or use GitHub to get API
2:29:07 oh for for tick tock yeah I could
2:29:09 probably figure that [ __ ] out I just
2:29:11 don't have time to build it
2:29:13 if someone wants to help me build a Tick
2:29:17 Tock live comment parser
2:29:19 that's probably a product that would be
2:29:21 sellable
2:29:22 until Tick Tock just replicates it
2:29:28 Studios want to hire non-writers to
2:29:30 write yeah and they want to hire non
2:29:34 um
2:29:35 non-union actors to act
2:29:38 you know there's there's a weird there's
2:29:40 a weird uh sort of uh
2:29:43 uh loophole in the SAG strike where all
2:29:46 of the independent movies are still
2:29:48 filming because they're not Studio films
2:29:51 but it's like most of the films that are
2:29:53 being done are not Studio films so so
2:29:56 like the actors are still acting
2:29:59 the big Studio Projects have ground to a
2:30:01 halt but interesting interesting times
2:30:05 I want to start using AI to teach
2:30:06 English it'll be really good at that
2:30:09 absolutely a good time literally 10
2:30:11 times easier to learn new things and
2:30:13 develop skills yep totally agree I'm old
2:30:15 enough to remember red pill or blue pill
2:30:17 exactly are you an actor yeah I have a
2:30:20 degree in acting but I'm a founder and a
2:30:22 CEO of a tech company right now I I am
2:30:25 an ex actor
2:30:28 you can take the actor off the stage but
2:30:30 can you take the stage out of the actor
2:30:32 no I don't think so no I don't think so
2:30:35 Bob tell him what he's won
2:30:38 you've won a tick tock Channel
2:30:39 congratulations you have an audience now
2:30:42 act like an ass perfect
2:30:46 foreign
2:30:47 yes I am Ahoy capitalist making more
2:30:51 money while needing less Manpower well
2:30:54 Jenny Knox I yes that's happening
2:31:00 um
2:31:02 there's
2:31:04 pay attention to to some of the things
2:31:06 going on
2:31:08 take a look at what Sam Altman's doing
2:31:10 with this project called World coin that
2:31:13 just launched
2:31:14 so there's a couple interesting things
2:31:16 open AI used to be non-profit then they
2:31:20 realized they needed a to raise a lot
2:31:22 more money than they could raise as a
2:31:24 non-profit they made it for-profit but
2:31:26 they made it capped for profit so what
2:31:29 that means is
2:31:31 investors can only make up to a certain
2:31:34 point of profit and then the rest of
2:31:36 that money goes in essentially into a
2:31:38 fund
2:31:40 and what worldcoin is is its a human
2:31:45 verification system so it's identity
2:31:47 it's it's creating an identity system
2:31:49 that can identify that you're a human
2:31:53 it's a cryptocurrency that that is a
2:31:57 spendable currency and then there's a
2:31:59 whole payment platform
2:32:01 and what Sam Altman's talking about and
2:32:04 it sort of ties in with his capped
2:32:06 for-profit thing so so so let's say that
2:32:10 investors are capped off at 100 billion
2:32:13 dollars
2:32:15 or a trillion dollars but let's say that
2:32:18 the value that open AI generates is in
2:32:21 the tens of trillions of dollars so this
2:32:23 Gap he's throwing into a fund
2:32:26 and he's looking at it as a Prosperity
2:32:30 like a wealth redistribution model a
2:32:34 universal basic income model so World
2:32:36 coin is the infrastructure the potential
2:32:38 infrastructure for that vision of the
2:32:41 future where
2:32:43 the AI generates sort of exponentially
2:32:47 increasing Prosperity while we need less
2:32:50 and less people to make that Prosperity
2:32:52 what Sam Altman's wanting to do is take
2:32:55 a a massive chunk of The Upside and
2:32:59 redistribute that back to the people
2:33:02 so
2:33:03 I know it sounds all communist and all
2:33:06 but
2:33:07 right now we've got an economy that
2:33:09 effectively is in stasis where you've
2:33:12 got this amount of people working and
2:33:14 this amount of GDP and there's
2:33:16 fluctuations in unemployment and all
2:33:18 that sort of [ __ ] but it's
2:33:20 relatively in stasis
2:33:22 these tools
2:33:25 are going to start doing this to the
2:33:27 prosperity and what you're saying the
2:33:30 capitalists are just like well [ __ ] the
2:33:31 employees fire more and we'll we'll
2:33:33 increase our upside
2:33:35 what Sam Altman's saying is we know
2:33:38 that's coming let's put mechanisms in
2:33:41 place now that take advantage of that
2:33:43 upside and get it back to the people
2:33:45 that are going to be displaced
2:33:47 so will it happen I don't know but pay
2:33:50 attention to some of the [ __ ] going on
2:33:51 because it's not not being thought about
2:33:54 it is being thought about
2:33:56 every second of every day that's my
2:33:59 brain lately it wasn't my fault sorry
2:34:01 did you invent this [ __ ] morgy
2:34:04 please tell me you didn't invent this
2:34:06 [ __ ] Gen X knows Gen X [ __ ] knows
2:34:09 we'll take you out we drank out of hoses
2:34:12 you know that metal taste when you drink
2:34:14 out of a hose well you young people
2:34:16 don't because you've never had to do
2:34:17 that
2:34:18 but you're not you're telling me there's
2:34:21 not heavy metals coursing throughout my
2:34:23 body because of the hoses I sucked on as
2:34:26 a kid
2:34:27 oh I still make a million so what good
2:34:30 for you morgie
2:34:32 um Jenny Knox
2:34:33 don't you mean gen Z oh yeah it was
2:34:36 probably gen Z
2:34:37 Millennials are older now
2:34:39 roast if you put your mind to it
2:34:42 complicated
2:34:44 um
2:34:45 I love your sound happy
2:34:49 I don't know what that means morgy I
2:34:51 love your sound happy
2:34:53 uh
2:34:55 uh how are we still alive
2:34:59 oh you're from Taiwan cool
2:35:02 I love the use euphemisms for socialism
2:35:05 yeah I know well I don't wanna I I mean
2:35:08 the problem with the word socialism is
2:35:11 you can't use the word like
2:35:14 there are certain words in our
2:35:15 conversation these days that you just
2:35:17 can't say the word because everyone goes
2:35:20 my brain's on fire you're one of them
2:35:22 now I'm just looking at the [ __ ] that's
2:35:25 happening trying to make sense of it
2:35:28 YouTube transcripts can be copied into
2:35:30 chat gbt but they're quite long that's
2:35:32 why you have to use so Lauren if you're
2:35:35 if your YouTube transcripts are too long
2:35:37 for chat TBT use Claude for that that's
2:35:40 what I did my YouTube transcripts in you
2:35:43 can put a 75 000 Word document in club
2:35:45 and still interact with it
2:35:50 uh
2:35:52 this is performance art
2:35:55 this is performance art
2:35:58 right listen it's late I'm getting my
2:36:01 voice is getting out crackly and and as
2:36:04 a trained professional I need to take
2:36:06 care of my instrument
2:36:09 I wouldn't want my instrument to become
2:36:11 damaged I wouldn't be able to entertain
2:36:13 you on a nightly basis without a solid
2:36:15 instrument hmm
2:36:19 been fun hanging out with you I hope you
2:36:21 had fun tonight I hope you learned
2:36:22 something
2:36:23 um I had a good time I actually learned
2:36:25 some things tonight I'm gonna go check
2:36:26 out gorilla from Microsoft gorilla AI if
2:36:30 you weren't here earlier someone goes
2:36:33 but Kyle
2:36:34 have you heard about Microsoft's gorilla
2:36:36 Ai and I'm like Oh No this must be a
2:36:39 joke
2:36:40 now apparently they got a thing called
2:36:43 gorilla AI That's gonna be like the next
2:36:46 big thing and I went and looked at the
2:36:48 little article and it looks like it's
2:36:49 going to be the next big thing I don't
2:36:51 know how real it is that's what I gotta
2:36:53 go figure out
2:36:54 so maybe gorilla dot AI gets added to
2:36:58 the list at some point here
2:37:00 now in Microsoft it's at least
2:37:03 a year at least a year out although
2:37:05 they've been they've been moving pretty
2:37:07 damn fast
2:37:08 they've slowed down a bit since uh since
2:37:11 the New York Times debacle
2:37:14 um
2:37:15 but anyway interesting times interesting
2:37:17 times people it's interesting
2:37:21 um
2:37:21 for those youngsters here
2:37:24 um it's your bedtime go to bed
2:37:26 for you old people here what the [ __ ]
2:37:28 are we still doing up we should have
2:37:30 been in bed hours ago
2:37:32 uh
2:37:33 ah all right I'm gonna go I'm Gonna
2:37:37 Leave You All for for your fantastic
2:37:39 Sunday evening I really do appreciate
2:37:41 you hanging out with me
2:37:43 um I have a good good time and uh
2:37:45 Commandments
2:37:48 morgie you're fun you should come back
2:37:50 follow me if you haven't followed me go
2:37:52 ahead and follow me and uh wow good
2:37:55 night from Riyadh wow nice
2:37:58 uh that's all these are old I know
2:38:02 Morgan might be a bot I think morgie's
2:38:05 not a butt but morgie might be a bot
2:38:07 peace out everybody I'll see you
2:38:09 tomorrow
2:38:10 all right