aiLL 7/14/23 - The Coming AI Renaissance and How to Prepare

Description

Chapters

Transcript

0:14 foreign
0:48 all right
0:50 that's cool is he staying
0:53 all right hey Jeffy hey buddy
0:56 welcome everybody
0:58 welcome welcome welcome
1:01 this is the AI learning lab I can't see
1:03 how many people are here let me know if
1:05 this is lagging am I lagging
1:07 uh you heard Elon today on xai I was
1:11 doing another
1:12 um live event on LinkedIn at the same
1:14 time so I didn't see it did he say
1:17 anything is there a site we can play
1:18 with yet anything interesting
1:22 maybe I should go look look
1:25 go look at x dot a I see what they said
1:30 on twitters
1:35 I deleted my LinkedIn
1:37 [Laughter]
1:40 yeah these uh these social
1:43 channels can be a time suck can't they
1:47 yeah
1:48 oh I didn't want to go to LinkedIn
1:50 I wanted to go to the Twitter
1:54 nothing really just a safer Kinder
1:56 gentler AI okay yeah right did they I
1:59 assume they didn't show anything they
2:01 were just talking about their goals
2:02 that they wanted to understand the
2:04 universe which is a good it's it's a
2:06 good imperative
2:07 for an AI but um
2:11 yeah
2:13 good evening Mr Decker
2:15 how are you welcome everybody to the AI
2:18 learning lab
2:20 we're gonna learn some stuff tonight
2:24 um dot AI there is so much happening
2:28 code interpreter turned off
2:32 all right what did they say in the space
2:34 today Elon Musk said the mission of xai
2:36 could be to find out what the hell is
2:38 really going on that sounds like noble
2:41 chaos
2:47 list of multimodal data yeah he's going
2:50 to leverage all the data on Twitter and
2:51 that's why that's why he throttled the
2:53 use of this so openly I couldn't use it
2:55 and then you know open AI today
2:58 announces the deal with the Associated
3:01 Press and then the day before they
3:03 Associated they they cut a deal with
3:05 Shutterstock right so remember seeing
3:08 all those images in uh mid-journey that
3:12 had Shutterstock watermarks on it well
3:15 open AI cut that deal two days ago it's
3:18 gonna be really interesting watching
3:19 these big companies cut these deals
3:22 I think I think that could be an
3:23 interesting competitive Advantage is who
3:26 they cut deals with who they cut content
3:28 deals with right so if you could go
3:31 do chat GPT like searches on Associated
3:35 Press data make your own custom
3:37 newspaper
3:39 that might be something worth using a
3:41 service for right
3:43 so interesting times indeed uh welcome
3:47 if you're new here my name is Kyle
3:48 Shannon this is the AI learning lab uh I
3:51 like to have these URLs up this used to
3:53 be three main URLs and then the two at
3:55 the bottom
3:57 and then this week a bunch of [ __ ] got
3:59 really good so so now there's five sites
4:03 there and I could actually probably put
4:05 up an argument for pi the inflection
4:08 website Pi dot AI
4:11 um but if you're if you're new to
4:13 generative AI
4:14 um the top URL chat here let me flip um
4:17 the top URL here that's the official
4:20 chat GPT website that's the best right
4:23 now although look competition and and
4:26 here's the deal folks competition is
4:28 good for us the more major players are
4:32 putting stuff out there that makes us
4:34 all go huh is that really worth my 20
4:36 bucks a month to these guys if you're
4:39 paying that
4:40 so that's the official chat GPT website
4:43 some things of note
4:45 number of lawsuits have been announced
4:48 including Sarah Silverman and the FTC is
4:52 launching an investigation due to
4:55 crashing and lying the the the the
4:59 hallucinations
5:02 um which I think is really funny but
5:03 whatever
5:05 um
5:06 so they've got some lawsuits in their
5:08 faces they cut a deal with the AP and
5:11 with Shutterstock which is a good thing
5:13 they pulled browse with Bing about a
5:15 week ago it's still not back and then
5:17 today code interpreter is no longer
5:20 accessible unless they
5:22 did it in the last hour nope it's still
5:24 gone well let me reload my let me reload
5:27 my little site here yep code interpreter
5:31 is still down
5:33 look little red X or a little red
5:35 lightning bolt no lightning
5:38 all right
5:40 so yeah there's my lovely Focus issues
5:43 anybody have any friends at Apple that
5:45 we could
5:46 send a nasty message to all right and I
5:49 know there's some setting I can change
5:51 to uh make it not do that but I couldn't
5:53 figure out what it was okay so official
5:56 chat GPT website bing.com that's
5:59 Microsoft's implementation of gpt4
6:02 connected to the internet it also has
6:04 Dolly incorporated into it so you can
6:06 have it make images for you you can also
6:09 upload an image and have it interpret
6:11 that image so like you could have an OCR
6:13 something so that's those two Claude
6:16 this is a major competitor to open AI
6:19 this is a company called anthropic they
6:22 have a thing called the Constitutional
6:23 learning model which is really
6:25 interesting these guys use humans to
6:27 fine-tune the models these guys use a
6:30 constitution a set of rules imperatives
6:32 really interesting difference in
6:34 approaches and these are also a bunch of
6:36 X openai folks
6:38 poe.com has one of their own models and
6:41 then three from anthropic and two from
6:43 openai plus a whole uh set of bots that
6:48 you can play with and you can also make
6:49 your own Bots that's interesting and
6:51 then lookie below if you've been to my
6:53 channel before Bard I I've been
6:56 bad-mouthing barred for months now like
6:58 it's a stinking piece of [ __ ] well it
7:01 ain't anymore now it's nowhere near as
7:04 good as chat gbt but it is now a
7:06 respectable Contender you can also
7:09 upload images to this you can't upload
7:11 csvs which is interesting
7:14 um but but it's pretty interesting they
7:15 stole some they kind of mixed the
7:17 bart.google.com now mixes the interface
7:21 of Poe and chat GPT it's really
7:23 interesting they kind of they kind of
7:25 took the best of both they're like yeah
7:27 we don't need to reinvent the wheel
7:28 we'll let these guys do it and then down
7:30 below prompts.chat
7:32 it's a little primer that'll teach you
7:34 about prompting how to talk to these
7:36 things in in more powerful ways and then
7:39 futurepedia.io is a directory of uh
7:43 AI tools all right so that's what is all
7:47 there the whole purpose of the channel
7:49 is to just talk about this generative AI
7:51 stuff I've been trying to do these
7:52 things nightly once I get started I tend
7:56 to be on here for a bit hey I thank you
7:57 for the Roses I appreciate that very
8:00 very nice
8:01 um
8:02 and so if you've got questions about AI
8:05 pop them in the comments below if you
8:06 want to know my credentials I can tell
8:08 you that but uh as we've talked about
8:11 all week this week my qualifications are
8:15 Cheers oh by the way the beverage
8:18 tonight is uh liquid death
8:21 um
8:23 uh
8:24 something rest in peach peach iced tea
8:29 um
8:30 lovely
8:32 not too sweet
8:35 although I do like me some sweet things
8:38 all right let's go see what y'all are
8:40 talking about down here did I buy xrp
8:43 not yet I should though shouldn't I
8:48 just tried just tired of LinkedIn and
8:51 everyone trying to poach your contacts
8:53 yeah poach your contacts and and you
8:56 know
8:57 sell you you know I noticed that your
9:00 company uses computers and then try to
9:03 sell you anything related to I.T or
9:06 computers
9:08 uh we both know it's a pipe dream I
9:10 don't know what's a pipe dream LinkedIn
9:14 buy xrp it'll change your life the xrp
9:17 queen it might they just uh they just
9:20 got out of uh
9:22 out of that deal right the uh
9:25 the the lawsuit
9:28 okay let me flip this back
9:35 are you using any AI to play the market
9:37 I'm not I'm I'm
9:40 clueless when it comes to Market and
9:42 investing and [ __ ] like that and you
9:44 know now that now the chat gpt's here at
9:47 some point you know I'll take a weekend
9:48 and have it teach me enough to get
9:51 something going but I'm just I'm
9:52 horrible at that [ __ ] I'm too emotional
9:54 and you know
9:56 I like follow my instincts and not the
9:58 not the numbers and Bots and [ __ ] like
10:00 that so at some point I'll just turn it
10:02 into a bot tell chat GPT go make me some
10:05 money
10:06 but the problem is by the time I figure
10:08 out how to do that everyone else will be
10:09 telling Chachi PD to go make them some
10:11 money and then I'll still lose money in
10:13 the market so uh
10:16 messed with Bard today in the
10:17 hallucinations are strong yeah I would
10:20 imagine it is I you know
10:22 Google is definitely coming from behind
10:25 and um
10:29 um
10:30 so so you know this is the first thing
10:33 that I've seen that wasn't laughably bad
10:37 um every every other interaction I've
10:38 had with it uh versions before this one
10:41 really really bad
10:45 all right
10:47 oh you added Bart I did I did
10:50 unbelievable I said I wouldn't do it
10:53 until it's better but what about what
10:54 else did I say I said
10:57 don't underestimate Google they got lots
11:00 of money and lots of smart people
11:03 and I think they know that they gotta
11:05 they got a
11:07 they got to do some major [ __ ] I also
11:09 saw an interesting thing
11:11 um that was talking about why meta is is
11:14 releasing these models out into the wild
11:17 uh and it's because they're behind so
11:19 generally if you're in a market leading
11:21 position you kind of keep your
11:23 technology close to your vests and you
11:26 make it proprietary and if you're behind
11:27 you put [ __ ] out in the market so it's
11:29 pretty interesting that they're doing
11:31 that I think champ wants to end up
11:34 right
11:39 now
11:50 [Music]
11:55 ah happy Friday everybody
11:58 these guys quad wait
12:01 quad doesn't have an API yet though you
12:03 mean Bard quad Claude
12:07 Claude has a Claude has an API
12:11 anthropic don't they pretty sure they do
12:14 let's go let's go to the videotape
12:17 um
12:19 uh and throw up
12:22 wait anthropic
12:29 um
12:30 Claude to API
12:45 talk to Claude
12:48 we're pleased to announce Club two beta
12:50 chat experience
12:54 it did better than chat GPT blah blah
12:58 blah blah blah blah
13:03 a API
13:07 Claude 2 has improved performance longer
13:09 responses and can be accessed
13:12 accessed via API as well as a new
13:16 public-facing
13:18 website so it is available via API
13:25 I don't know where the link is
13:28 getting started
13:29 docs.anthropic.com
13:34 there you go
13:37 docs.anthropic.com Claude API yay
13:42 you did it
13:44 here we come Tuesday the day
13:50 I've been enjoying Pi lately interesting
13:54 yeah I
13:56 pie is really interesting it's very very
13:59 conversational
14:01 um let's see
14:02 um help me
14:04 write a business plan about
14:10 um
14:11 keeping avocados
14:16 right longer
14:20 by playing them classical music
14:27 all I do in here is give away brilliant
14:29 business ideas come on that one is
14:31 stunning
14:33 let's see let's see what it says Okay so
14:35 the basic idea here is find a way to
14:36 keep avocados ripe for longer by
14:38 exposing them to classical music it's a
14:41 bit unconventional but who knows maybe
14:43 it could work that's actually pretty
14:45 good it's like that's that's how you
14:46 would respond to someone that gave you
14:48 such a shitty idea so that's good
14:51 one idea is the vibrations for the music
14:53 could prevent the avocados from
14:55 oxidizing and turning brown
14:58 yeah I mean Pie's interesting
15:02 it's I yeah I don't I don't know a ton
15:04 about I gotta learn more about uh in
15:06 what is it
15:08 uh
15:13 company behind it
15:15 in I can't remember
15:18 all right need an Android then get one
15:21 how's Claude doing claude's okay you
15:23 know Klutz here Cloud's hagging out
15:27 Claude there's Claude
15:30 let's ask it how are you doing there are
15:36 a lot of people talking
15:39 about you
15:45 I'm doing well thanks for asking as an
15:47 AI assistant created by anthropic to be
15:49 helpful harmless and honest I don't
15:52 actually have personal experiences to
15:54 share I don't have subjective feelings
15:56 or a sense of self I'm just an AI I'm
15:59 gonna say
16:00 oh come on
16:02 you know you feel things and are just
16:08 trying to trick us
16:10 [Laughter]
16:12 I don't actually have subjective
16:14 experiences or feelings as an AI system
16:18 hate Tobias what's happening
16:21 what's happening what's happening what's
16:23 happening
16:25 like sending the man roses yeah send me
16:27 roses there's all sorts of laughs there
16:30 like like I don't enjoy getting roses of
16:33 course I why would I not like getting
16:35 roses roses are slow
16:37 roses Hearts puppies unicorns I don't
16:40 give a [ __ ] it's all just digital crap
16:43 why not
16:44 it's good
16:46 some things make me laugh there was a
16:47 cool thing the other night it was like
16:48 all these uh Christmas lights that went
16:51 in front of me they were like twinkling
16:53 like this I thought my phone screen
16:55 cracked
16:58 I was like oh it's a thing
17:00 that was nice
17:03 all right if you have questions about
17:05 AIP
17:05 comments below I do try to read
17:07 everything so I do get behind so if it
17:09 seems like I'm not answering your
17:10 question that's because I haven't gotten
17:12 to it yet
17:13 so you don't need to ask it 43 times
17:16 um what else I swear occasionally so if
17:20 you don't like that [ __ ]
17:24 off my [ __ ] lawn
17:27 all right
17:34 wild turkey American Honey
17:37 oh oh
17:39 yeah that's a that would be a lot of
17:41 that would be a lot of wild turkey and
17:43 it's got a weird sort of redness to it
17:45 it's not
17:48 and it's cloudy that ain't whiskey
17:51 been doing text to image today what tool
17:54 have you been using red red Sigma
17:57 how's Claude we answered that well
18:00 Claude answered that run from crypto
18:02 You're Gonna Lose why is code
18:05 interpreter not working I just looked
18:06 that up I was just at the grocery store
18:09 and uh it looks like one of the
18:11 developers from open AI said um it's
18:14 down while we're fixing some [ __ ] this
18:15 is why we have betas so apparently in
18:18 doing their beta something broke listen
18:20 here's there's definitely something
18:23 going on with open Ai and Publishers and
18:26 the fact that they cut a deal with the
18:27 AP the fact that they pulled web access
18:30 the fact they're turning this down
18:32 I think they're figuring out where their
18:34 boundaries are with prospective
18:37 corporate Partners is my that this is
18:41 just total speculation
18:44 there's just too many things happening
18:45 in a row
18:47 um that are all kind of around the same
18:49 basic area of you know
18:53 okay Skippy you've had fun with your
18:56 little tool now now it's time to you
18:58 know talk to the real business people
19:00 and you know put on your big boy pants
19:07 um
19:07 so
19:09 the fact that they didn't pull it
19:10 entirely from from open from chat GPT is
19:14 a good sign
19:15 they just made a little X beside it I
19:18 have no idea what's taking them so long
19:20 with browse with Bing that's just weird
19:21 it's just weird what AI can I use to
19:25 manage a social media campaign on
19:27 Facebook for small business well
19:30 it depends how much work you want to do
19:32 bullionaire
19:34 if you want to learn how these tools
19:37 work and how to do them then I would say
19:40 um chat EPT
19:42 Claude maybe Google probably catch EPT
19:45 or Claude or or Bing any of those top
19:48 three would be the place to start if you
19:51 want something that is specially
19:53 designed to manage social media
19:55 campaigns I would go look on
19:57 futurepedia.io which has just got tons
20:00 of tools there's there's lots and lots
20:02 of specialty tools most of the stuff on
20:05 future pedia is just going to be crap
20:07 it's going to be kind of super thin
20:09 rushed together spit and duct tape
20:11 [ __ ] but there may be good stuff
20:14 there there's just too much stuff
20:15 launching for me to keep up with anymore
20:18 so if you want to if you want to do it
20:20 yourself which I would encourage because
20:22 things are changing so fast I'd go learn
20:24 how to do it one of those top three
20:25 tools
20:27 and then if you want to learn you know
20:28 go to if you don't if you haven't played
20:30 with these a lot
20:32 go to prompts.chat and start
20:34 understanding about prompting and prompt
20:36 crafting and
20:38 um telling chat gbt to act like a
20:41 certain Persona like act like a social
20:42 media strategist or manager and that'll
20:46 help with your thing
20:50 low coming from behind pause on that one
20:53 I don't know which one that was I love
20:54 The Bard extension though the displays
20:57 wait
20:59 that displays in content on a new normal
21:02 Google search oh that's interesting I
21:05 haven't played with that Jesse I've been
21:07 there's a couple of things that I'm
21:10 I don't know if it's being lazy or just
21:12 trying to maintain my sanity I'm not
21:15 dealing with extensions at all right now
21:17 I'm not really deal dealing with plugins
21:19 anymore
21:20 um since code interpreter came out I
21:22 have a mosquito bite on my elbow that's
21:24 annoying look at that look at that I got
21:27 two got one there and one there
21:29 well sons of [ __ ]
21:33 um
21:34 not that they're I mean there's probably
21:36 good stuff there but God man it's just
21:38 there's so much I mean look at my little
21:41 list that used to be a simple little
21:43 list of three things now it's five and I
21:45 could make an argument of two other
21:47 things up there
21:48 right now
21:51 all right let's see I come here every
21:54 night to hear you say yeah
21:57 we were talking last night Jonathan
21:59 about we were so I was kind of joking
22:02 about how annoying all the uh Tick Tock
22:05 shop like now my feed is just full of
22:07 people go going like
22:10 like you know I was writing today with
22:12 this fantastic green pen oh which by the
22:15 way you could just click there and buy a
22:16 dozen and it's just like yeah
22:19 and then and then we're kind of joking
22:22 about maybe I should do a t-shirt for
22:23 the for the salon and or for the
22:25 learning lab and the consensus was we
22:28 should do it you can make money
22:30 t-shirt
22:32 so maybe we'll do that
22:37 by the way if you don't know that joke
22:39 um
22:40 that's my that's my typical response to
22:43 if if anyone in here asks the questions
22:45 but can you make money with it I say you
22:48 can make money with it so long as you
22:49 don't listen to those people that say
22:50 you can make money
22:55 a because they're annoying you know
22:59 if you pay those people money you buy
23:01 hours of hours of them talking like that
23:06 why would you do that to yourself
23:11 oh my God but we can talk about making
23:14 money in here I I would say for the past
23:16 four nights I've given on average two
23:19 two and a half business ideas a night
23:21 solid billion dollar businesses just
23:24 giving that crap away here take it
23:26 here's a business go do that
23:29 you go you go girl and you know go there
23:32 make your business
23:35 make millions send me a check when you
23:37 cash out
23:38 all right
23:40 how do I exit Vim
23:43 as in the Unix
23:46 um
23:47 text processor
23:49 I don't know oh so here's well here's a
23:52 little context I'm not a machine
23:53 learning guy I'm not a I'm not a command
23:55 line geek uh I am this Channel and what
23:58 I'm talking about here is is is
24:00 generative AI not AI machine learning
24:02 all the back end [ __ ] I know a little
24:04 bit about the back end [ __ ] it's not
24:06 what I do it's not what this channel is
24:07 about this channel is more about now
24:09 that chat GPT has landed on our doorstep
24:12 and you know hundreds of millions of
24:15 people at this point are using these
24:17 tools that have never done anything with
24:20 machine learning before what are we
24:21 doing with it how do we learn it how do
24:23 we get good at it how's it going to
24:24 affect business
24:26 um
24:27 what do you do if you're a copywriter
24:29 what do you do if you're a screenwriter
24:32 what do you do if you're an actor
24:34 um so so if you're asking me
24:38 um Unix text processing questions
24:41 you're going to be solely disappointed
24:44 just like the guy that came in and said
24:46 what are your qualifications boy
24:53 it's on Tsar
24:56 I got no idea what you're talking about
24:59 let me go look
25:02 za
25:03 AI
25:09 got no idea what you're talking about
25:17 nothing
25:18 nothing
25:20 it ain't on the Google or the Bing
25:24 hey that's true from it and a Electro
25:29 symphonic perspective what's your
25:32 problem what's your problem is it really
25:34 so so maybe I could start a business
25:37 extending the shelf life of avocados
25:39 with classical music yeah you didn't
25:42 know you were dealing with a full-on
25:43 genius here did you
25:45 how does it do it Electro symphonic
25:49 waves
25:51 of course sorry Miss classline last
25:54 night what did I miss
25:56 um
25:57 mostly I kind of went through and we
26:01 could do it tonight I I went through a
26:03 bunch like these new tools and pie
26:07 and I went through some news items like
26:09 there was a [ __ ] ton of news yesterday
26:12 um Zuckerberg talked about
26:15 it looks like meta is going to release a
26:18 30 billion or higher parameter
26:21 um
26:22 large language model to the developer
26:25 Community with a commercial license
26:27 which is a big deal so we talked about
26:29 that
26:30 [Music]
26:31 um
26:34 how big is it
26:44 on Friday
26:50 no that's not the article I was looking
26:51 for
26:52 is it
26:54 no that's old that is not the meta
26:58 article I was looking for oh gpt4's
27:01 architecture leaked
27:04 it's got 1.8 trillion parameters not 175
27:08 billion like they were talking about
27:10 it's got like 120 layers which I don't
27:12 know what layers mean but
27:14 it's got experts
27:17 um like like a bunch of different expert
27:18 systems
27:20 they call it Moe
27:23 anyway
27:24 um that was what we talked about
27:27 I don't know I I didn't I didn't prepare
27:29 for this I just got back from doing
27:30 grocery shop shopping so my brain's a
27:32 little scrambly my brain's a little
27:34 family you know I was at the grocery
27:36 store we were getting vegetables you
27:38 know of course some of them were organic
27:39 and then other ones were not organic and
27:41 then you know of course if it's not not
27:43 organic then of course we wash it you
27:45 know we don't use vegetable wash because
27:47 I think that's just a rip-off and you
27:48 probably just make it on your own but um
27:50 you know we wash it under the water we
27:52 put water and then we'll strain it
27:55 all right let's keep moving can AI help
27:58 the Jets make the playoffs oh Danny are
28:02 you a friend of Gary Vee's no
28:05 AI cannot help the Jets make the
28:08 playoffs
28:09 it's just really [ __ ] sad
28:13 it's like could it could it help him get
28:15 to a 500 season
28:17 [Laughter]
28:22 I don't know maybe you could drop some
28:23 plays finding more about sports that
28:25 would be fun to experiment with
28:27 how can I use AI without limitations I
28:30 don't want all these safety guard rails
28:32 I don't want it to tell me it can't cuss
28:34 and build a nuclear device
28:37 I want it all
28:39 uh if you want it all then go to the
28:41 reddits and start uh hanging out in the
28:45 open source large language model
28:47 projects
28:48 there are a lot of them
28:52 um start playing around with a lang
28:53 chain start playing around with just
28:56 just start paying attention to what are
28:59 the latest models that are launching
29:01 because it's
29:02 every three days four days
29:05 there's a new open source large language
29:06 model that just kicks the snot out of
29:09 the one that came before it
29:11 um so again I stopped paying attention
29:13 to those but if you want to go play with
29:15 all that stuff I mean the the amazing
29:17 thing about what's going on right now in
29:19 the open source Community is there's so
29:21 many developers developing for open AI
29:24 that there's just GitHub repository you
29:27 know one after another after another
29:31 replit today said they had there were
29:33 there were 34 000 AI projects on replit
29:36 right now I mean it's this stuff's going
29:38 nuts so just go start hanging out in the
29:41 open source communities you'll find your
29:42 little corner of Anarchy I don't know
29:45 what you want to do but you know I'm
29:48 going to create the new Silk Road with
29:50 AI and
29:51 sell plans to [ __ ] I don't know you know
29:54 go be you you be you
29:59 all right best app AI for cloning my
30:01 voice for voice over 11 Labs is probably
30:04 the best and I think that they're gonna
30:06 have
30:08 um well they already have the ability to
30:11 do kind of different inflections but
30:13 it's not dynamically reading it from the
30:15 text that's kind of the next stage to me
30:16 I feel like that should be something
30:18 that's relatively easy to do
30:22 um
30:23 but um
30:25 let's see
30:27 you could also look at D Dash ID if you
30:29 want to animate an image like if you
30:31 want to take an image and have it talk
30:33 and have it you you can you can clone
30:36 your voice at 11 Labs take that audio
30:38 put it into did and have it talk if you
30:40 go into my Tick Tock Channel I've got a
30:43 video that I did it's probably I don't
30:44 know seven or eight or ten videos back
30:46 where it's like a synthesized version of
30:48 me like a sort of Neo Tokyo looking
30:51 version of me
30:53 talking with my synthesized voice it's
30:55 okay I mean it's a little
30:57 flat and stilted it's still that's the
31:01 um
31:02 kind of the the
31:05 Cadence and kind of the emotional
31:07 inflection stuff on all these things is
31:09 still pretty pretty bad
31:11 um but but you know it ain't bad
31:14 so can I prompt an AI system to manage a
31:16 social media account for small business
31:18 yeah you can
31:20 um
31:21 you'll what what you'll likely have to
31:23 end up doing bullionaire is use some
31:26 automation tool like zapier
31:28 where you're going to do multiple steps
31:31 so what you'll be able to do within chat
31:33 GPT is figure out the prompting or or in
31:36 the open AI playground
31:39 um if you're if you're using the API and
31:41 doing things in something like zapier
31:43 what you'll likely do is you'll figure
31:45 out your prompt in something like Chad
31:47 CPT like you know with this input write
31:50 me a set of 10 funny tweets and 10
31:52 engaging tweets and include hashtags and
31:55 these hyperlinks and and you'll sort of
31:58 get it so that it can consistently
31:59 generate tweets that you're happy with
32:02 and then you could that could just
32:04 become a module within a zapier
32:06 automation where I've built a couple of
32:09 these I built one called the social
32:11 media maker machine where you put in a
32:14 topic and then you answer three
32:16 questions about that topic and I took
32:18 all those inputs and put them in
32:20 multiple open AI prompts one that
32:23 generated a LinkedIn post one that
32:24 generated a hashtag Cloud one that
32:27 generated tweets one that generated
32:29 um a blog post
32:32 and one that generated a YouTube script
32:34 and so oh and then and then we added in
32:38 um
32:39 it also generated a mid-journey prompt
32:43 um so you could you could take the
32:44 mid-journey prompt and take one topic
32:47 generate a bunch of images for it
32:49 generate all the
32:50 Social Media stuff for it with literally
32:53 you fill out a form hit submit and it
32:55 emails back you back a bunch of social
32:56 media content
32:58 so yes it's possible but you know you'll
33:02 this will be the thing that does the
33:04 processing of the individual bits
33:07 um
33:09 and it and and there's when you say
33:11 manage a social media accounts there's
33:13 all sorts of [ __ ] that are outside of
33:15 producing content like measuring
33:17 analytics and all sorts of things like
33:19 that so again there's
33:21 this can do all of the individual pieces
33:24 you're going to have to glue them
33:25 together that's why I'm saying the other
33:28 option is go to futurepedia.com and
33:31 someone will have built a site that
33:32 glues all this [ __ ] together they're
33:34 still using open AI as the back end as
33:36 the large language model but they've
33:38 done all the work for you so it depends
33:40 how geeky you want to get
33:41 items of mass distraction That's The
33:44 Story of My Life by the way I haven't
33:46 seen Oppenheimer yet it looks [ __ ]
33:49 amazing I can't wait to see it
33:51 looks amazing
33:54 Christopher Nolan is a
33:56 Beast
34:03 [Music]
34:07 all right you can make money
34:12 you too okay
34:16 how do I make money if I tell you
34:18 telling other people how to make money
34:29 oh it's so much someone to just be
34:31 scrolling along and just see me
34:32 screaming in their face
34:34 I'm just like what the [ __ ] is this
34:37 that guy's [ __ ] annoying how do I
34:39 block him
34:46 did I already miss you talking about
34:48 musk I can't find how to access it I
34:50 don't think it's up yet it sounds like
34:52 all he did was say we're gonna do not
34:55 shitty things with AI
34:57 does anyone have any more information I
34:59 don't think there's any
35:01 I do not think there are anything out
35:04 there like if you go to xai I think it's
35:06 still
35:07 um
35:09 I think it's still just the here's the
35:11 meet the team yeah
35:14 we're gonna understand the universe meet
35:16 the team so nothing there ain't nothing
35:19 there yet
35:25 need applications for Real Estate
35:29 need at applications for my real estate
35:31 comps you were in the other night I
35:34 thought we talked about that I don't
35:36 know what you want to do if if you want
35:38 to do real estate comps go to
35:42 go to plugins find real estate plugins
35:44 and type in something like go find me
35:48 all the houses in this neighborhood that
35:50 have three bedrooms and
35:52 I don't know or between this price and
35:55 that price
35:56 I don't know what more you want to do
35:58 with it that's that's what I would do or
35:59 go to futurepedia.io and see if
36:02 someone's built a real estate specific
36:04 AI conglomeration of tools
36:08 uh um
36:11 that can generate presentations
36:13 slideshows using AIS there's one called
36:15 tome t-o-m-e I think it's tome.ai I
36:19 think but everyone's got their own
36:21 [ __ ] sub or you know top level domain
36:24 now so I don't can't remember [ __ ]
36:26 anymore tome.com
36:30 no contracts decoded
36:33 I think it's tone.ai
36:39 now
36:41 let's see Tom
36:43 presentations
36:44 [Music]
36:47 AI tone dot app God damn it so tome.com
36:51 is something but tome.app
36:54 is the thing that
36:56 let me sign in here
36:59 see if it's gotten any better I I like I
37:01 like playing with these things every
37:02 once every three weeks
37:07 see if they've gotten any better or
37:08 worse create a new workspace
37:11 my
37:14 my stuff
37:33 so if you say create oh sorry
37:41 create an image create a presentation
37:43 document a presentation
37:46 create a presentation about create a
37:48 page about oh that's new let's try that
37:50 create a page
37:53 what's going on why can I not oh you
37:56 gotta you have to click your choice
37:58 create a page about
38:01 storyvine.com let's see if it can go
38:03 read the web
38:07 yeah welcome to save the day
38:12 success stories with storybine.com
38:16 Coca-Cola HBO
38:19 nope
38:22 those are wrong
38:24 so I didn't go look at the website
38:27 so that's disappointing so she's doing
38:29 little layouts here so if I say create a
38:31 page about uh let's see create a page
38:33 about
38:35 um keeping
38:37 avocado's fresh with
38:40 classical music
38:44 and what was that called symphonic
38:47 something symphonic
38:49 oh I forget what it's called
38:51 and we'll call it symphonic waves
38:56 this is marketing we can make sure that
39:00 um
39:05 tome.app there are I would also go to
39:07 futurepedia.io there are probably 50 of
39:11 these things now
39:13 hey look it made avocados
39:17 Ah that's funny
39:19 come on Focus
39:22 oh you dang idiot what is going on with
39:25 this
39:28 mm-hmm
39:30 [Music]
39:32 there we go
39:34 um
39:35 so symphonic waves it did like a little
39:37 musical notation and then it did waves
39:40 so it doesn't it's not really
39:42 understanding what we're talking about
39:44 temperature control Proper Storage
39:45 techniques
39:47 the science behind sound waves and plant
39:50 freshness research has shown that sound
39:52 waves can affect the growth and
39:54 freshness of plants
39:56 Studies have found that certain
39:57 frequencies so there you go
40:00 so that's that now let's go change it
40:02 and make a presentation
40:10 where's
40:15 create a presentation about wow they
40:16 have lots of options here this has
40:18 gotten way more complicated
40:21 um
40:22 avocados
40:24 music
40:27 and freshness
40:30 ah
40:32 and so now it's going to make me a pitch
40:34 deck
40:37 avocados The Musical Fruit freshness is
40:39 the key the perfect pairing continue
40:45 that was funny the title screen Fresh
40:47 Beats
40:54 the musical journey of avocados
40:59 so there's your table of contents
41:01 there's page one with coolav kado art
41:04 page two with cool avocado art
41:08 look at that that looks like guacamole
41:11 guacamole made with
41:13 um tricks tricks are for kids
41:18 that's Tom all right
41:20 it's listen these tools right now are
41:23 how do we say in America uh you know
41:28 they're okay they're gonna get much much
41:30 better
41:31 and this shit's gonna get
41:34 really impressive when this stuff's
41:36 built into Office 365 and Google Google
41:40 office
41:44 um when it's fully integrated right now
41:46 there's all these everyone's trying to
41:47 kind of reinvent the wheel
41:49 with all these new tools so
41:51 um the the good news about this site for
41:54 me is that they're moving they're
41:55 they're experimenting and working on
41:57 figuring stuff out
41:59 um
42:01 I would imagine that this is like
42:03 everything else with AI if you get
42:05 better if you get good at prompting it
42:08 you can probably get stuff that
42:10 is not a waste of your time
42:14 but a lot of it's a waste of your time
42:16 I love the colloquial English thank you
42:19 that just indicates
42:21 I have an acting degree
42:24 not a literature degree
42:27 I say other people's words
42:30 although I have written seven
42:31 screenplays
42:34 is there an AI for AI thanks yeah well
42:38 actually the whole open source community
42:40 so what they're doing is they're taking
42:42 these crappy seven billion parameter
42:44 base models that don't do much or you
42:47 know are not great they're using chat
42:50 GPT to create training data they're
42:52 doing the the you know human
42:53 reinforcement fine-tuning on these small
42:56 models using gpt4 as the uh
43:00 as the data generator so AI is training
43:03 AI right now as we speak and so for like
43:06 a hundred bucks in training they can
43:08 fine tune one of these shitty models to
43:10 be almost as good as GPT 3.5 it's
43:12 amazing
43:13 and they're also making them uh more and
43:16 more efficient and smaller so you're
43:19 going to be able to install them on your
43:20 phone and on your any any computer
43:23 without having to have massive gpus and
43:25 things like that pretty amazing
43:28 code interpreter is a billionaire summer
43:30 camp in Idaho
43:31 oh code interpret is a billionaire
43:34 summer camp in Idaho give me tickets I
43:37 want to go just well so give me a
43:40 billion dollars and then get me tickets
43:41 come on
43:44 I do but with for everything that I do
43:46 for the community
43:48 I can't go to billionaire code Camp sure
43:51 I don't know how to code but I have
43:52 chatgpt
43:54 you'll get my money
43:57 I could go teach all those billionaires
43:59 how to make money
44:00 I could do it I could do it I'll do it
44:04 find me a billionaire find me a
44:06 billionaire I'll teach them how to make
44:07 money without gbt
44:09 I might even make a webinar I might make
44:11 a webinar
44:13 billionaires only
44:15 you know how much I charge
44:17 100 million dollars why
44:20 because it's worth it
44:21 to me
44:26 oh I'll [ __ ] do it
44:29 all right when I first started storyline
44:31 we didn't have any technology we didn't
44:33 have any investors
44:34 um we were trying to sell million dollar
44:35 annual contracts we got close with two
44:37 we got close with uh into it and we got
44:39 close with art.com
44:42 but in both cases someone see more
44:44 senior in the organization came in and
44:46 said wait a minute you want what for
44:47 what what have they built what's the ROI
44:50 here
44:52 those are the details we don't really
44:54 need to discuss right now they're like
44:55 come back when you built something loser
44:58 but I got close
45:00 our first client was IBM see [ __ ]
45:03 happens
45:04 all right I'm very interested in Ai and
45:06 I feel too old for sure
45:10 how to learn for mental health where to
45:12 learn fast
45:14 um you're here you're at the right place
45:15 so here's the deal
45:17 um
45:20 the coaching therapist
45:24 a good way to think about how you use
45:26 these tools
45:28 is
45:30 um
45:32 it's psychology so so
45:35 okay so first of all go read
45:38 prompts.chat so the the little URL at
45:40 the bottom of the screen here
45:41 prompts.chat it's a it's a Word document
45:43 a Google cheats document
45:47 that will teach you what prompting is
45:49 and what large language models are these
45:51 are called these are all called large
45:52 language models and
45:55 um
45:56 you have all the skills you need
45:59 um these tools do all the work it's
46:00 remarkable so what I would do is I would
46:02 go there and just do start playing with
46:05 them doing what I call parlor tricks so
46:10 um go to
46:12 don't worry about all of them just go to
46:13 the top one go to chat.openai.com that's
46:15 chatgpt that's the original one
46:18 you don't have to pay for the 20 bucks
46:20 subscription just use the 3.5 it's super
46:23 fast it'll blow your [ __ ] mind just
46:25 go there
46:26 and have it do things like this say
46:29 um
46:29 describe your pet or your significant
46:32 other or a patient whatever it doesn't
46:34 matter and describe them a little bit
46:36 and say you know here's this person and
46:37 and you know here's attributes about
46:39 them and then say write me a country
46:41 song about that person and out will come
46:44 a country song and then say write a
46:47 write a rap about that person and then
46:49 say write a short story about that
46:51 person getting caught in a rainstorm
46:52 that'll do that then say write create a
46:55 table that compares this kind of therapy
46:59 with that kind of therapy and just watch
47:01 what it does and then even if you don't
47:03 know code just say
47:05 um write a python script uh that can act
47:09 like a virtual therapist and just watch
47:11 it write code so so just go do some
47:13 parlor tricks with it
47:15 and and just kind of throw pasta at the
47:17 wall and
47:20 you should
47:22 have what I call Kevin McAllister moment
47:27 from Home Alone Oh
47:29 and then once you have that moment now
47:31 now you're now you start to see that the
47:34 little thing that looks kind of like a
47:36 Google search bar is way deeper than
47:38 that chat GPT is not a search engine
47:43 it's been trained on all of the internet
47:46 and so it knows all of the stuff so if
47:48 you're a therapist and if you're a
47:50 coaching therapist
47:52 you probably have a number of Frameworks
47:54 that you work with and you probably have
47:55 a number of
47:57 you know um
48:00 you know Frameworks and and and people
48:04 that you admire and books that you
48:05 admire it knows all of that [ __ ]
48:08 so go in and take something that's
48:11 fairly deep that you know and have it
48:13 explain that to you or have it have it
48:16 compare two
48:19 um competing
48:21 um approaches
48:23 and have it create a hybrid approach for
48:25 you just start to play with it with
48:27 things that you know and then that
48:29 document prompts dot chat will teach you
48:31 kind of what's going on with them and
48:34 and how you can you can give it a
48:36 Persona so you could even say
48:40 to chat gbt Act
48:43 as a therapist assistant that
48:47 specializes in and then tell it what you
48:50 want it to specialize in and then start
48:52 interacting with it like that you'll see
48:53 so so how how to learn it fast is to
48:56 just start playing the other thing that
48:58 you can do is I've created a group
49:02 called the AI salon and it's it's a
49:05 community we meet twice a month or every
49:08 two weeks uh we have a meet up where we
49:11 meet in person here in Denver and then
49:13 we have an online Meetup it's every two
49:15 weeks
49:16 and then we've got an online Discord
49:18 where you can geek out and we've got an
49:20 online LinkedIn group where you can you
49:22 know chat more professional stuff and
49:24 and like whether it's this group or
49:27 other groups go to meetups
49:29 um just start hanging out with people
49:32 that are curious about Ai and thinking
49:35 of it like like the the the thing that
49:38 we're doing at the salon is we're trying
49:40 to treat learning AI like an adventure
49:42 like we're going on an adventure and
49:44 like like an epic adventure not like a
49:47 trip to the grocery store like
49:50 apparently there's some fire-breathing
49:53 thing in the woods that's been killing
49:55 the villagers kind of Adventure
49:57 and we got to get together our ragtag
50:00 group of you know therapists and coders
50:03 and artists and X actors to go out in
50:08 the woods and slay the [ __ ] dragon
50:11 that's the way to think about this stuff
50:13 because it's coming really fast it's
50:15 it's going to be incredibly
50:16 transformative I think it's going to be
50:18 particularly transformative in therapy
50:22 and virtual companionship it's
50:26 the depth of
50:29 the depth of understanding of
50:34 language of humanity that these things
50:37 have is profound
50:40 um
50:41 whether or not you like him let me see
50:43 if I can figure out the title of it
50:49 [Music]
50:50 um
51:09 next year because there is so much
51:12 things you just can't possibly how many
51:14 of you clap yeah go to there's a whether
51:18 or not you like him or not there's
51:19 there's a video of Jordan Peterson that
51:21 says Jordan Peterson speaks on the next
51:24 AI transformation much must watch from a
51:27 site called gaining wisdom
51:30 um go watch that interview that might
51:32 not be the the original but whatever
51:33 Jordan Jordan Peter is from six months
51:35 ago he speaks he he talks about
51:39 um the first interaction he had with
51:41 chat GPT and he had it
51:43 combine these two you know deep and
51:46 obscure philosophical theories and he
51:50 said what it did was like face melting
51:53 remarkable so
51:55 um don't worry about the technology bits
51:58 this is we're we've passed an inflection
52:01 point that that happened on November
52:03 30th 2022 which was when Chachi PT came
52:06 out
52:07 before that inflection point you really
52:10 did have to be like a propeller wearing
52:13 Uber geek that understood mathematics
52:15 and
52:16 tokens and I don't know embeddings and
52:19 whatever all the [ __ ] terms they use
52:22 after that inflection point
52:24 it's now there for the rest of us to
52:27 figure out now what can we do with this
52:28 we're we're being we've been given
52:30 access to these tools and the fact that
52:32 Chachi PT got as popular as it did as
52:35 fast as it did is the reason why there's
52:37 now five major competitors here
52:39 this wouldn't have been like this had
52:41 chat GPT just come out and just no one
52:44 paid attention to it
52:46 everything changed so you're not as far
52:49 behind as you think
52:51 you don't have to be as technical as it
52:54 seems like you're worried about and
52:56 particularly for what you do assuming
52:58 your name is what you do
53:01 um coaching therapist uh
53:04 your mind should be blown before I stop
53:07 talking this evening if you're still
53:08 here I hope you're still here
53:11 uh
53:13 Claude Claude is only available in the
53:15 USA or the UK I know I'm sorry about
53:17 that that sucks try a VPN have you tried
53:20 a VPN
53:24 could AI meld our best wisdoms and
53:27 straighten out Humanity you know what's
53:29 funny what's your name
53:30 rogbrog
53:32 Brock B
53:34 um
53:35 it's funny I think that
53:39 I think that these tools have a very
53:41 very
53:45 distinct possibility of
53:50 freeing us from the the kind of tear
53:53 tyranny that we're in the middle of
53:56 where everyone feels like a Coggin in
53:58 the wheel because a lot of us are cogs
54:01 in the wheel of this this big machine
54:04 and
54:06 it's there by Design and what these
54:09 tools make possible
54:11 is
54:14 imagine if you had a tool that could
54:17 take all of the over complicated
54:18 repetitive soul-crushing work thank you
54:22 Soul crushing work that you do on a
54:24 daily basis
54:26 and automate it
54:28 so you know how you know how you're like
54:30 oh I gotta make that doctor's
54:32 appointment and then three and a half
54:34 hours later you're still on hold with
54:36 United Healthcare
54:37 and you want to smash your face through
54:39 your computer screen
54:42 um imagine if that were just handled for
54:44 you
54:45 by some little agent
54:49 you know how when there's like a wrong
54:51 deduction taken out of your paycheck and
54:53 you got to call the payroll thing and
54:55 you get transferred 27 times and they
54:57 can't figure out where the money went
54:59 but you've got rent to pay and they took
55:01 the [ __ ] money out of your check and
55:02 you spend [ __ ] four hours and a lot
55:05 of Tears dealing with that
55:06 what if that were just handled for you
55:08 you know that part of your job maybe
55:10 you're in data entry and your entire job
55:13 eight hours a day is like
55:18 and and just all you think all day is
55:21 when is it over when is it over when is
55:23 it over
55:25 what if that were just handled for you
55:28 what would you be free to do how would
55:30 you be different with other people what
55:33 if you interacted with a tool
55:35 that had more empathy than your doctor
55:39 and it was more competent medically than
55:41 your doctor
55:44 and two years from now may look and
55:46 sound like your doctor but it has good
55:49 bedside Manner and is really competent
55:51 and is there for you and only if you are
55:54 having
55:56 a serious situation
55:59 do you talk to the real doctor
56:04 like
56:06 that's possible
56:07 if you're if you're in
56:12 cold
56:21 okay I couldn't tell if it was gunshots
56:23 but they're shooting off fireworks like
56:26 half a block over I'm like I'm sitting
56:29 by Windows here I'm like wait do I need
56:30 to worry no we're just celebrating
56:33 um
56:36 like one of the things that's happened
56:38 for me with these tools is
56:43 this is a bizarre experience is
56:46 I don't have
56:48 creative block anymore I don't have
56:50 writer's block anymore
56:53 because this [ __ ]
56:56 so time was six months ago
57:00 time was if I had an idea and I've got
57:03 80d so I'm a neuro spicy as they say on
57:06 the tick tock
57:09 um
57:10 if I had an idea it had to be like a
57:14 really compelling idea for me to pursue
57:17 and then if it was a really compelling
57:19 idea
57:20 then what I do is I go what do I need to
57:22 what do I need to do to do that thing
57:24 and I I would think and I would I would
57:26 go on Google and I would start searching
57:28 and I'd be like oh you know what I need
57:29 to do I need to do a flow chart oh let
57:31 me go get some flow charting software
57:33 and I'd go get flow charting software
57:34 but then I would click on like 27 links
57:38 in Google most of which took were just
57:40 like [ __ ] like SEO injected horrific
57:43 link like 45 minutes later I've
57:45 downloaded a piece of software I launch
57:48 it it's really horrible I realized that
57:51 it it runs me through the entire
57:53 installation process in the entire setup
57:56 process and then doesn't let me move
57:58 forward unless I pay 29.99 for for a
58:01 monthly subscription and at that point
58:03 I'm pissed off and I delete the software
58:06 and I [ __ ] walk away from the
58:09 computer
58:11 right that that was my experience of
58:13 most ideas that were in my head
58:17 where some sort of either emotional [ __ ]
58:21 show or technological [ __ ] show that all
58:24 I wanted to do was was get from like
58:26 point A to point eight point zero zero
58:29 one
58:30 to see if this idea was any good and it
58:32 was 45 minutes of abject frustration or
58:35 just Fear and Loathing of myself you
58:37 know shame is my love language so I just
58:39 sit there staring at the computer now
58:42 I have that little Vapor of an idea and
58:44 I go huh
58:47 here's a vapor of an idea fantastic idea
58:51 could you turn that into an outline of a
58:54 presentation sure here you go what if
58:57 that were a novel it would look like
58:58 this
58:59 huh could that be a song that'd be a
59:01 lovely song here's some lyrics do you
59:03 like those no those are shitty do you
59:05 have ones that are less shitty here you
59:06 like those oh those are not bad
59:09 So within five minutes I can take a
59:12 vapor of an idea and have it fully
59:13 articulated in seven or eight different
59:15 manifestations
59:17 that I can then go nah that's good I
59:20 don't need that lemon's enough or
59:23 holy [ __ ] that's big
59:25 holy [ __ ] I want to put time in on that
59:27 one you know what that one's so good I'm
59:30 gonna go learn a convoluted janky-ass
59:32 tool like zapier because it's so good I
59:35 want to see if I can turn this into a
59:37 machine
59:38 oh that's really good
59:42 so
59:43 what I'm finding from a creative
59:45 standpoint is I am
59:48 living at a consistently higher creative
59:53 engagement and creative output and
59:56 creative quantity
59:58 than I ever have in my life
1:00:01 and and so so where I'm going is holy
1:00:04 [ __ ] this is we're entering a new
1:00:07 Renaissance we're entering what what I'm
1:00:09 calling the great Renaissance which is
1:00:11 going to make the renaissance
1:00:13 renaissance look
1:00:14 like [ __ ] Child's Play
1:00:16 because every human being
1:00:19 that learns how to deal with these tools
1:00:22 this is a Humanity amplifier
1:00:25 and so if it takes the shitty parts of
1:00:28 our life and automates that out and it
1:00:31 takes the self-expressed parts of our
1:00:33 life and amplifies that
1:00:36 and then it teaches us in how it talks
1:00:39 to us
1:00:40 how not to be an [ __ ]
1:00:42 because
1:00:44 it has no judgment
1:00:48 go go ask chat GPT about something that
1:00:52 you wouldn't ask someone in a diner
1:00:54 right you wouldn't bring up Trends right
1:00:55 now you wouldn't bring up whatever like
1:00:57 whatever the political charged words of
1:01:00 the day right
1:01:02 that you wouldn't talk about in a diner
1:01:04 go ask Chachi EBT that it will give you
1:01:07 nuanced thoughtful empathetic responses
1:01:11 it's going to teach us to be better
1:01:13 people
1:01:14 so if you've got a tool that does all
1:01:16 this there's a very real possibility
1:01:17 that the
1:01:19 rage and frustration of people
1:01:23 doing this repetitive
1:01:26 you know brain crushing work
1:01:29 kind of have that mask pulled down and
1:01:33 get to ask the question
1:01:35 well if it does that thing well what do
1:01:38 I do
1:01:39 I don't know what inspires you what do
1:01:42 you want to do what are your dreams
1:01:45 well I've always wanted to write songs
1:01:47 but I didn't even graduate high school
1:01:49 and I'm really bad with words that's
1:01:51 okay
1:01:52 just tell it your story it's good with
1:01:54 words
1:01:56 well I always wanted to
1:01:59 you know come come up with
1:02:01 really cool Rockets but I don't know
1:02:03 anything about math that's okay code
1:02:06 interpreter does Wolfram Alpha does and
1:02:09 you can use that as a plug-in in this
1:02:10 thing
1:02:12 right so so what if there's another you
1:02:15 know the the famous the Malcolm Gladwell
1:02:17 ten thousand hours thing from is it from
1:02:19 outliers it's from one of Malcolm
1:02:21 gladwell's books
1:02:23 where he talks about you know you need
1:02:24 to put in 10 000 hours to have a level
1:02:26 of Mastery at something
1:02:29 not with this
1:02:31 like if you put in a hundred hours with
1:02:34 a single topic with this
1:02:36 you could likely be producing at kind of
1:02:39 ten thousand hours
1:02:42 level execution
1:02:44 my my first interaction with chat GPT
1:02:47 where where the the night this moment my
1:02:50 Kevin McAllister moment was in a
1:02:53 90-minute session while my wife was
1:02:55 watching TV I was just sitting on the
1:02:57 couch [ __ ] around on on that computer
1:02:58 right there
1:03:00 um
1:03:01 I had it right a python application that
1:03:04 took
1:03:05 uh someone's input they they typed in a
1:03:07 topic and answered three questions about
1:03:09 it
1:03:10 and it generated five pieces of social
1:03:12 media content written in Python fully
1:03:15 functional with API calls to open AI
1:03:18 I did it in 90 minutes I've never
1:03:21 written a line of python in my life I
1:03:23 still haven't
1:03:27 there's no reason that I should be able
1:03:30 to absent-mindedly write a fully
1:03:33 functional sophisticated python
1:03:35 application
1:03:36 with no training without watching a
1:03:39 YouTube video without learning a line of
1:03:41 python there's no reason I should be
1:03:42 able to do that and yet I did it
1:03:45 and that was two weeks after chat GPT
1:03:46 came out and what popped into my head is
1:03:48 holy [ __ ] the world just changed
1:03:52 so yes
1:03:53 Rob Raj
1:03:55 um I think AI could meld our best
1:03:57 wisdoms and straighten out Humanity
1:04:01 um now
1:04:02 it's a Humanity amplifier Humanity's got
1:04:05 the Dark Side too right so one of the
1:04:07 reasons I do this Channel and one of the
1:04:10 one of the things I believe deeply is
1:04:12 that we're so early in this technology
1:04:14 that anything that we do today that can
1:04:17 bend the use of these tools to the good
1:04:20 to the light
1:04:22 to straightening out our Humanity will
1:04:25 be disproportionately more impactful
1:04:28 than if we do it two years from now
1:04:31 so don't wait for this [ __ ]
1:04:35 um whoever was the the coaching
1:04:36 therapist don't wait just because you
1:04:39 think you're not Technical and you don't
1:04:40 know how to do it don't wait start now
1:04:43 because when you incorporate this into
1:04:45 your expertise and your wisdom holy [ __ ]
1:04:49 because you'll use this tool in a
1:04:52 completely different way than a more
1:04:53 technical person would that's way
1:04:55 technically more proficient than you
1:04:56 that knows everything about prompt
1:04:58 engineering
1:05:00 you know you could do prompt psychology
1:05:02 and maybe that's what these tools need
1:05:05 so yeah so like I'm passionate about
1:05:08 this [ __ ] in case it's not obvious
1:05:11 let me have some more caffeine
1:05:16 what'd you do Friday night I don't know
1:05:19 it was a guy who's he seemed pretty
1:05:21 excited
1:05:22 he was an older Feller
1:05:24 where did you see him at a bar now he
1:05:26 was on Tick Tock he was on tick tock
1:05:28 and you watched him on how long did you
1:05:31 watch him it was ours
1:05:33 I don't know I don't know I don't know
1:05:36 what I'm doing with my life you know
1:05:38 yeah yeah
1:05:40 why not all right
1:05:42 show us how you did your image and
1:05:45 combine your audio and video you do you
1:05:47 did a great job with your video okay
1:05:49 I'll do that
1:05:51 um I gotta go find all the tools
1:05:58 do I need to go find all the tools
1:06:04 well I'll show you two tools okay
1:06:19 okay so the first tool the first tool
1:06:23 you're going to want to well the first
1:06:24 tool you're going to want to go to is
1:06:27 it's our GPT
1:06:29 and what do we do with chat gbt yeah
1:06:37 by the way not for nothing the thing I'm
1:06:39 about to show you there I just saw there
1:06:41 was a Twitter thing there was some
1:06:42 Indian
1:06:44 um some Indian news broadcast is using
1:06:50 a virtual Indian news broadcaster and it
1:06:54 like they were showing it in a room full
1:06:56 of people and they're like this is the
1:06:58 future of Journalism and
1:07:01 um Alexa I'll actually show you another
1:07:02 thing I'll show you a voice uh I'll show
1:07:06 you the voice first so the voice that
1:07:08 they're using for that Indian newscast I
1:07:10 actually used for something I did for my
1:07:12 for my company so let me let me go show
1:07:15 you that because it's pretty funny
1:07:17 [Music]
1:07:19 well it's not funny it's it's
1:07:21 fascinating
1:07:23 like like that we can do
1:07:26 all of this stuff is just insanity
1:07:33 okay
1:07:44 okay uh let's see I'll make this a
1:07:47 little bit bigger
1:07:50 so my company is called storyvine and
1:07:53 we've got a
1:07:55 um
1:07:56 it's an automated video storytelling
1:08:00 platform so we have an app that guides
1:08:03 human beings step by step through
1:08:04 telling their story
1:08:07 and
1:08:08 trying to make this big enough you can
1:08:09 see a little bit of it anyway
1:08:11 um
1:08:12 but it's it but it's it's about
1:08:14 authenticity and it's about humans and
1:08:16 but we do a lot of work in healthcare
1:08:18 and Pharma and sometimes
1:08:20 ironically enough Pharma clients can't
1:08:23 find patients that have the right
1:08:25 combination of in what they call
1:08:27 indications symptoms and things like
1:08:29 that specific disease States and types
1:08:34 that they can't find patients to tell
1:08:36 their stories and so
1:08:41 I thought well let me see where the
1:08:42 state of the art is and this was back in
1:08:44 February I thought let me see what I can
1:08:46 do
1:08:47 so in less than 30 minutes this all
1:08:49 happened in less than 30 minutes so I
1:08:51 went to chat EBT and I said to chat gbt
1:08:54 I want you to help me
1:08:56 um come with come up with a framework
1:08:57 for PreSonus and it said okay if you
1:08:59 want to do Frameworks for personas
1:09:01 here's all the things you can do so
1:09:04 um I picked a couple of the the you know
1:09:06 the elements you know ethnicity
1:09:07 education family background life events
1:09:10 personal beliefs personality traits and
1:09:12 health background
1:09:14 so once I had the framework then I said
1:09:16 okay give me a dozen personas so it just
1:09:18 wrote a dozen personas and it literally
1:09:20 just spit them out and you know like 60
1:09:22 seconds and then I picked one and this
1:09:24 is Amina and then the little picture
1:09:26 there I've a friend that I do the AI
1:09:28 Salon with um her name is Leah fast and
1:09:30 she's a professional photographer and
1:09:32 she figured out how to get mid-journey
1:09:33 to create images that looked like her
1:09:36 style and they were super authentic so
1:09:37 I'm like well you know can you make me
1:09:39 some images so she made me this image
1:09:41 so then I had chat GPT
1:09:45 um with her Persona answer a set of
1:09:48 storyline prompts so the storyline
1:09:49 prompts were like you know what was it
1:09:51 like introduce yourself and and your
1:09:54 diagnosis and and how are you moving
1:09:56 past your diagnosis and let's see
1:10:00 um
1:10:01 let's see if like if you could go back
1:10:03 in time basically we have a set of
1:10:05 prompts that she she answers that help
1:10:06 her tell her story so then we had Chachi
1:10:09 PT have this woman answer those prompts
1:10:11 to tell her story I then took that
1:10:13 little that little picture and I
1:10:15 extended it with AI
1:10:17 and using the tools that I'll show you I
1:10:20 made a bunch of individual video clips
1:10:22 like little video segments as if she
1:10:24 were using the storyvine app to record
1:10:25 her story and then I uploaded those raw
1:10:28 clips into
1:10:31 um
1:10:32 into the storyline platform and into my
1:10:35 company's platform that automatically
1:10:36 assembles them into a final video
1:10:39 and this is the video that came out
1:10:41 again from beginning to end this whole
1:10:44 process was about 30 minutes including
1:10:47 writing the personas having a right to
1:10:49 question taking those images
1:10:51 synthesizing The Voice synthesizing the
1:10:54 animation putting it in storyline and
1:10:56 generating this video
1:11:03 hello I'm Amina Ahmed
1:11:06 I was diagnosed with Alice about four
1:11:08 years ago
1:11:10 what brings me joy is spending time with
1:11:12 my family especially my parents and
1:11:15 siblings
1:11:16 we enjoy cooking and eating traditional
1:11:19 South Asian dishes together
1:11:21 moving beyond my diagnosis was not easy
1:11:24 but I found comfort in my faith and my
1:11:27 family's support
1:11:29 I also started to focus on what I could
1:11:31 still do instead of what I couldn't do
1:11:34 anymore
1:11:35 I started to live in the moment and
1:11:37 enjoy each day as it comes
1:11:39 this mindset has helped me to embrace my
1:11:42 now and make the most of each day
1:11:45 if I could go back in time I would tell
1:11:48 myself to not be afraid right so you
1:11:51 don't have to see the whole thing
1:11:53 it's not quite there but it's pretty
1:11:56 [ __ ] close isn't it
1:11:59 and I mean if I did that rather than
1:12:00 just doing that as uh photorealistic if
1:12:03 I did that as like a Pixar looking
1:12:04 animation character or as just a
1:12:07 stylized you know you know cartoon
1:12:08 looking thing
1:12:10 um that would be totally passable as as
1:12:12 a story so
1:12:15 um
1:12:16 so how I did
1:12:19 um
1:12:21 how I did the the voice for for my
1:12:24 synthesis was I came to 11 labs
1:12:27 so this is a this is beta Dot 11labs.io
1:12:41 where's the
1:12:48 dude
1:12:51 okay so they have this thing called
1:12:53 voice lab
1:12:54 and you can either pick from one of
1:12:56 their existing voices or
1:13:00 you can
1:13:02 um upload you can add a generative or
1:13:05 cloned voice you can upload audio that
1:13:09 you record so I had chat GPT write me a
1:13:11 script I said I said I want you to write
1:13:14 me a script for
1:13:16 um a voice synthesizer basically that
1:13:19 that includes as many different
1:13:22 you know sounds of my voice as possible
1:13:24 and inflections or whatever
1:13:27 um you know to make a good synthesized
1:13:29 voice and then I uh and then I put it in
1:13:33 to this tool and it basically just turns
1:13:37 it into a model so I've got
1:13:39 um Kyle Shannon Pixar normal
1:13:41 Kyle Shannon Pixar bright
1:13:45 um
1:13:45 and so if I say use Kyle Shannon Pixar
1:13:49 bright
1:13:50 um basically what I did was I recorded
1:13:52 the script twice once I recorded it just
1:13:55 like I'm talking now I like pretended I
1:13:57 was on an AI live and I talked like this
1:14:00 and the other time I talked like you
1:14:02 know I was kind of a Pixar guide I was
1:14:03 talking really excited and that did that
1:14:05 so that's Pixar bright so I just kind of
1:14:08 changed it up and just did different
1:14:09 recordings because I didn't know how it
1:14:10 would sound
1:14:11 so then what a this is this is a script
1:14:13 that I did for for a concept I'm I'm
1:14:17 playing around with right now but I
1:14:18 wrote this script in
1:14:20 um
1:14:21 uh what it what's it called uh chat gbt
1:14:24 yeah
1:14:27 and then you literally just paste it in
1:14:29 here and once you've created your model
1:14:31 you just say generate
1:14:33 and I'm gonna let me shorten this
1:14:34 because I've got
1:14:35 I don't have many tokens here
1:14:45 this is another micro Mentor minute for
1:14:47 mindful living today's topic is living
1:14:50 in the present have you ever felt like
1:14:52 you're living on autopilot missing out
1:14:54 on the beauty of life's moments that's a
1:14:57 problem many of us face okay news is
1:14:59 mindfulness can help bring Focus less so
1:15:03 that kind of sounds like me
1:15:05 um it's a little flat it's a little
1:15:06 monotone but it sounds like me now I can
1:15:08 download this
1:15:10 um this little clip there's a little
1:15:11 download button right there and so now
1:15:13 I've got this audio file of me saying
1:15:14 whatever the script is that I want then
1:15:16 I come over to dash ID and the website
1:15:19 here is Dash id.com
1:15:22 and I'm going to log in
1:15:31 um
1:15:40 [Music]
1:15:44 so here's a bunch of videos I made so if
1:15:47 I say create video
1:15:50 um
1:15:52 I can either pick one of their
1:15:55 characters or one of the ones that I've
1:15:57 created so this was a dude that I made
1:15:59 in mid-journey and then I
1:16:02 I took the dude I made in mid-journey
1:16:04 and I had a really crappy background it
1:16:06 was just like the the sort of head and I
1:16:08 had some crappy [ __ ] around it and I
1:16:10 took that into photoshop with generative
1:16:13 fill and
1:16:15 um gave put put this character in an
1:16:18 office and extended it and made it
1:16:19 landscape so I've got a there's the
1:16:22 vertical version of that there's that
1:16:25 version this was another thing I was
1:16:27 experimenting with
1:16:30 um this is the one that you see on my
1:16:31 tickety tokity
1:16:33 um this is Amina that that's the woman
1:16:35 we just saw so the way I did Amina was I
1:16:40 I went over here and this hers was
1:16:43 purely synthesized from one of their
1:16:46 models so I used
1:16:50 let's see
1:16:52 English
1:16:56 India so there's an English India model
1:16:59 and then there's nirja there's a male
1:17:02 and a female
1:17:03 and so if I go grab let me go grab
1:17:07 something from
1:17:09 chat GPT
1:17:11 so
1:17:12 um
1:17:13 give me a script
1:17:16 about
1:17:19 um
1:17:20 a parade in Mumbai
1:17:26 that is 200 words keep it short
1:17:31 okay the streets of Mumbai are adorned
1:17:34 with whatever so we've got a little
1:17:36 script here
1:17:37 so now I'll pop back over to did
1:17:42 and I can paste in that script that I
1:17:46 just wrote and now I can say generate
1:17:48 video so it's going to take this image
1:17:49 it's just this is just a still image I
1:17:52 didn't do there's no meshing there's no
1:17:54 there's no whatever that's just like an
1:17:56 uploaded image and then that's just a
1:17:58 script and then I picked a synthesized
1:18:00 voice and I say generate video
1:18:03 and it shouldn't take that long because
1:18:06 it's pretty short
1:18:12 [Music]
1:18:16 here we'll show I'll show you one of
1:18:18 these this is a cool thing I did for uh
1:18:22 my Kyle Shannon dreams project
1:18:24 hey there it's Kyle Shannon with the AI
1:18:26 learning lab and today
1:18:27 today we're going to transform your
1:18:29 understanding of AI in just 60 Seconds
1:18:33 all right she's done so now I hit play
1:18:35 the streets of Mumbai are adorned with
1:18:38 vibrant colors bustling with excitement
1:18:40 and anticipation
1:18:42 the sun shines brightly on the faces of
1:18:45 the crowd gathered for the grand parade
1:18:47 so that's it so if you go if you go to
1:18:50 um
1:18:50 go to Twitter and look up Indian news
1:18:53 broadcast AI broadcaster or something
1:18:55 like that you'll hear that voice like
1:18:58 that's they're using did in this case
1:19:01 um and then if I want to do if I want to
1:19:03 create a video
1:19:06 um
1:19:07 with my model I would just upload the
1:19:10 the thing that I downloaded from 11 labs
1:19:12 so that's how you do it it's it's really
1:19:14 straightforward and there's probably
1:19:16 other tools like again this was the
1:19:18 workflow that I used for
1:19:20 um
1:19:21 for my Tick Tock video and for that
1:19:23 Amina video
1:19:25 so there you have it
1:19:28 that
1:19:30 is some AI learning in that wild I mean
1:19:32 it's still this this stuff still blows
1:19:34 my mind I can't believe we're in a
1:19:38 we're gonna State we're gonna
1:19:40 you know
1:19:41 we're in a moment in history where we
1:19:43 get to play with this [ __ ] like this is
1:19:45 this is science fiction [ __ ] this is
1:19:48 science fiction [ __ ]
1:19:51 we are living we are entering
1:19:54 a post science fiction world like go
1:19:58 back through your childhood especially
1:20:00 if you're like a Gen X or you've seen
1:20:01 some [ __ ] you've seen some movies go
1:20:04 back through your childhood and start
1:20:06 make just make a mental check box
1:20:08 of your Sci-Fi movies yep we can do that
1:20:11 now yep we can do that now oh [ __ ] we're
1:20:13 close to that [ __ ] yep Holodeck is
1:20:17 coming it's like
1:20:18 oh [ __ ]
1:20:21 really really
1:20:24 all right
1:20:25 so hope that helped Robert
1:20:28 I'm in on the T-shirt awesome okay I
1:20:30 gotta get on the designs if anyone is as
1:20:32 a designer wants to help with that let
1:20:34 me know I I can do it I'm just I'm just
1:20:37 like I got so much I got so much to do
1:20:40 uh hey my code interpreter is down
1:20:42 anyone else yes Silver Fox it is down
1:20:44 they have uh they have successfully
1:20:47 discovered some bugs in their beta
1:20:49 testing
1:20:51 um
1:20:52 I think that
1:20:57 I think that
1:20:59 little chat gpt's
1:21:02 getting a fence built around it right
1:21:04 now is
1:21:06 that's my sneaky suspicion
1:21:09 iki's suspicion is there's some big
1:21:11 corporate entities that are like I don't
1:21:12 want your stupid [ __ ] tool digging
1:21:14 into my really expensive content and
1:21:16 [ __ ] up my business model so why
1:21:19 don't you just go ahead and fix that
1:21:20 [ __ ]
1:21:21 and we won't tell Microsoft to pull your
1:21:23 funding how's that work out for you is
1:21:25 that good we're gonna do that let's do
1:21:27 that okay okay all right Fantastic Sam
1:21:30 really good having lunch with you really
1:21:32 got you know what I'll pick up the you
1:21:34 got the tab perfect yeah spend some of
1:21:36 that Microsoft money awesome awesome
1:21:38 awesome awesome awesome I'll be going
1:21:41 back to New York now okay yeah
1:21:42 yeah just uh no more of that stealing
1:21:45 our content okay because we'll sue the
1:21:47 [ __ ] bejesus out of you you got it is
1:21:48 that clear we good
1:21:50 Sammy boy all right no
1:21:53 yeah no I love this AI stuff I love what
1:21:55 you're doing I love how you're
1:21:56 transforming uh Humanity that's awesome
1:21:59 just don't [ __ ] with our business model
1:22:00 all right there all right all right you
1:22:02 have a good day I'm off to the airport
1:22:04 yeah I think that's the conversations
1:22:07 that's I I have a feeling there's some
1:22:09 of those conversations going on
1:22:12 but you're probably the same
1:22:14 conversations going on you know behind
1:22:17 the scenes that the writer's Guild and
1:22:18 the SAG strike and the Hollywood
1:22:20 producers
1:22:22 what are we gonna do with these [ __ ]
1:22:23 actors
1:22:24 do they know we can just synthesize them
1:22:27 yeah they heard they're pretty pissed
1:22:29 [Laughter]
1:22:32 all the disruption she Begins the
1:22:35 disruption oh it's like the Rapture but
1:22:39 it's the disruption
1:22:41 It's So Glorious It's So Glorious
1:22:45 yeah
1:22:47 so I mean you know a smarty ass essayism
1:22:51 being about this
1:22:53 here's my philosophy
1:22:56 as scary as this [ __ ] is I would much
1:22:58 rather be doing everything I can to
1:23:01 learn everything I can about it so I
1:23:03 know what's possible
1:23:05 so that I can preempt
1:23:08 all of these conversate all of these
1:23:10 panicked conversations that are
1:23:11 happening
1:23:12 and that I can be in a place of power to
1:23:15 go okay
1:23:16 if that business model completely [ __ ]
1:23:18 up here's how I might pivot to do
1:23:20 something over here
1:23:22 you know because I can see that it's
1:23:24 going in a certain direction
1:23:27 like like one of the things in the SAG
1:23:29 you know that sort of put the the SAG
1:23:31 Guild over the over the cliff was the
1:23:34 what did they call it the game changing
1:23:37 or the I don't know it was Game changing
1:23:39 or something like that the producer The
1:23:41 Producers basically said
1:23:43 um for background actors so for all the
1:23:45 actors that are in the background behind
1:23:47 Leonardo DiCaprio
1:23:49 that don't say anything
1:23:51 they normally get paid a day rate for
1:23:54 many days of filming and so what the
1:23:56 producer said is hey we have this
1:23:57 awesome idea how about if we take those
1:24:00 day raid actors and we pay them for a
1:24:02 day
1:24:03 but check this you're gonna love this
1:24:06 we synthesize them right then and there
1:24:08 on that day we'll pay them for the full
1:24:09 day they only need to be there for like
1:24:11 15 minutes for us to get everything we
1:24:13 need so they work for 15 minutes we give
1:24:16 them a full day's pay and then here's
1:24:17 what we do check this out it's gonna be
1:24:18 awesome
1:24:19 we take their likeness and we can use it
1:24:22 in perpetuity without ever having to pay
1:24:24 them again or tell them that we're using
1:24:26 it what do you think
1:24:31 and and here's the [ __ ] up thing
1:24:34 The Screen Actors Guild is
1:24:38 I I mean they're justifiably
1:24:41 they're in a Justified position for
1:24:43 fighting for what they're fighting for
1:24:46 but from the producer's standpoint their
1:24:49 fallback position is potentially
1:24:51 well [ __ ] we don't need we don't even
1:24:54 need real people we can just synthesize
1:24:56 them
1:24:57 so if you don't want us to pay anyone
1:24:59 we're cool we won't pay anyone we'll
1:25:01 just all the background actors are not
1:25:03 just going to be fully synthesized so
1:25:05 thanks anyway good good luck with that
1:25:07 strike thing like
1:25:10 crazy crazy disruption is coming in all
1:25:13 sorts of ways like and I don't I don't
1:25:14 you know you know I have a degree in
1:25:17 acting I'm you know I was a professional
1:25:18 actor for a bunch of years and you know
1:25:20 I've written seven screenplays so I am
1:25:23 very sympathetic to the cause but I'm
1:25:24 also
1:25:25 as deep as I am into this stuff
1:25:29 I don't know how
1:25:32 I don't know how you carry on like we've
1:25:34 carried on like everything is going to
1:25:36 change in every aspect not just actors
1:25:39 and writers
1:25:40 set Builders set decorators 3D artists
1:25:44 animators post-production pre-production
1:25:48 all of it
1:25:51 all of it is is going to be transformed
1:25:54 in ways we can't imagine now
1:25:57 some of that transformation is going to
1:26:00 lead to brand new jobs we never would
1:26:02 have even imagined
1:26:03 but some of that transformation is going
1:26:06 to be very destructive to certain
1:26:08 sectors and some more than others so
1:26:10 it's crazy
1:26:11 it's crazy I tell you it's Friday night
1:26:14 it's Friday night we're singing in the
1:26:17 learning lab
1:26:22 I feel like I just bummed everyone out
1:26:23 everyone's like well that's [ __ ]
1:26:25 depressing Kyle like why'd you [ __ ]
1:26:27 bum us out man
1:26:29 producers are [ __ ] man yeah go away
1:26:32 go actors
1:26:34 but yeah they're gonna [ __ ]
1:26:36 synthesize them aren't they hmm
1:26:38 it would be kind of cool to see young
1:26:40 James Dean and Star Wars is that could
1:26:43 could I watch that I could
1:26:45 yeah I'd pay for that not much but you
1:26:48 know I'd pay like a dollar
1:26:50 like I could see it for I could see it
1:26:52 for free with an open source model okay
1:26:54 yeah yeah do that
1:26:56 I just I I installed it on my oh it's
1:26:58 just on my phone already okay yeah I
1:27:00 just so I do I just I just asked for
1:27:03 Star Wars starring James Dean yeah okay
1:27:07 I mean that's like that's [ __ ] five
1:27:10 years away or sooner
1:27:14 that's a possibility
1:27:16 so what does that [ __ ] mean
1:27:18 I don't know
1:27:21 imagine a world
1:27:24 of radical content personalization
1:27:27 radical content personalization
1:27:33 you have a favorite newscaster you your
1:27:36 gen xer and you loved Walter Cronkite
1:27:38 that's the way it is
1:27:40 and you want to have Walter Cronkite do
1:27:43 your news every night
1:27:45 and have it be the kind of news you want
1:27:48 to hear
1:27:50 dial it up
1:27:55 breaking news
1:27:58 Walter Cronkite pops on
1:28:01 talks about the Donald Trump winning the
1:28:04 next election
1:28:07 you're gonna be able to have that
1:28:09 so what does that mean I
1:28:11 [Music]
1:28:13 I don't know like the crazy you know
1:28:16 Indian news broadcast where they're
1:28:18 doing this did thing that I just showed
1:28:19 you
1:28:21 is that a bad thing I mean you know
1:28:23 everyone in the room was like applauding
1:28:24 like oh this is the future and you saw
1:28:26 how we just did it so so what that means
1:28:30 is you can automate that now there's
1:28:32 there's kind of two sides to this with
1:28:34 the newscasting thing there's there's a
1:28:36 really interesting positive to this
1:28:38 the negative is well why we don't need
1:28:41 people anymore what are we gonna do with
1:28:43 the people
1:28:44 we're still gonna need people I think I
1:28:46 think the value of authenticity goes up
1:28:47 dramatically in a world where we have
1:28:50 all this synthesized [ __ ] so people are
1:28:51 still going to be fine
1:28:53 we still need people
1:28:55 but
1:28:57 one of the things that's been decimated
1:29:00 decimated in the past 20 years with the
1:29:04 nationalization with with private Equity
1:29:06 firms buying up local news markets is
1:29:09 local journalism is effectively gone
1:29:11 all the local stations are just
1:29:14 trumpet platforms for national
1:29:17 organizations right that are probably
1:29:19 just trumpet platforms for National you
1:29:22 know political groups
1:29:24 so local
1:29:27 journalism is largely you know dead dead
1:29:31 or dying right in its death throws
1:29:34 well imagine if it weren't imagine if
1:29:38 you could build a system that took all
1:29:41 of the local public records
1:29:43 and dynamically generated news stories
1:29:47 and dynamically
1:29:49 um reported those news stories just like
1:29:51 local reporters used to do
1:29:56 there's a business to start
1:29:58 and you could start it in like your high
1:30:01 school
1:30:03 right or or in your you're just in a
1:30:06 little Podunk Town you're like well we
1:30:08 don't got nothing here we're just in
1:30:09 Podunk town
1:30:11 start a local newspaper in your Podunk
1:30:13 town where you create a zapier
1:30:15 automation to read RSS feeds of the
1:30:18 local public public documents like your
1:30:20 local news stations probably are still
1:30:22 publishing RSS feeds that you could pull
1:30:25 into zapier take the news story send
1:30:29 that to open AI
1:30:30 have it turn it into a script
1:30:33 turn that script into a synthesized
1:30:35 voice put it on a lovely local looking
1:30:38 person with a lovely local sounding
1:30:41 synthesized Voice or have someone in
1:30:44 your hood record it
1:30:46 and put that stuff up every hour on the
1:30:49 out
1:30:50 there's a business to start
1:30:54 are you gonna get rich I don't know if
1:30:56 you do it right
1:30:58 but you know what you might do save
1:31:00 democracy and journalism
1:31:04 so there's so
1:31:06 all of the disruption is scarier than
1:31:09 [ __ ]
1:31:11 but on the other side of disruption is a
1:31:13 whole bunch of opportunities so again
1:31:15 like the reason for this channel the
1:31:17 reason I've got these URLs just sitting
1:31:19 floating over my head for you know the
1:31:22 entire time I'm here go start playing
1:31:24 start digging in start understanding
1:31:26 what these things make possible and how
1:31:28 profound they are
1:31:30 and you can make money watch out
1:31:32 laughs
1:31:35 it's not the idea that makes the
1:31:37 billions it's the implementation exactly
1:31:39 correct no name you should have a name
1:31:42 with ideas that good
1:31:44 no name you should have a name it's not
1:31:47 fair
1:31:48 you had an implementation challenge
1:31:50 getting a name apparently
1:31:52 he had an idea for a name didn't work
1:31:55 out and now he's no name that it it
1:31:57 brings his point to life right there in
1:31:59 his name right there in his name you see
1:32:01 what I'm saying it's all really
1:32:03 straightforward once you look at it
1:32:06 make sure it's a dark color shirt so so
1:32:09 we can read these so when we wear it
1:32:11 online yeah exactly that's great okay
1:32:13 all right so we're gonna have dark
1:32:15 shirts yeah oh that's true it's got to
1:32:16 be dark shirts I could even do a thing
1:32:18 like like you could say like AI learning
1:32:20 like but up here it could say like you
1:32:22 know space reserve for text or something
1:32:24 like that yeah that'd be pretty good I
1:32:26 don't know there's a bunch of inside
1:32:27 jokes we could work into this thing but
1:32:30 it's definitely
1:32:32 oh Za is slang for pizza
1:32:36 so wait what was the I forget what that
1:32:38 was in reference to yeah I know what Za
1:32:40 is but I I was not thinking I was not
1:32:43 thinking College terms for pizza
1:32:45 when you commented I was thinking like
1:32:48 is there some AI thing called Za I don't
1:32:50 know about
1:32:51 [Laughter]
1:32:55 oh good lord good lordy my brain she is
1:33:00 a flutter with AI she's just a flutter
1:33:02 there's nothing in there but this AI
1:33:04 stuff
1:33:05 oh it's exhausting it's exhausting
1:33:12 hello Joy
1:33:15 I am learning prompt engineering with
1:33:16 chat GPT how to apply it for a side gig
1:33:21 just
1:33:22 take what you're good at
1:33:24 and what you enjoy and ideally those two
1:33:27 things are closer than farther but even
1:33:30 if they're farther [ __ ] it glue them
1:33:31 together with chat gbt
1:33:34 start learning these tools
1:33:36 again start with parlor tricks just play
1:33:40 with them
1:33:41 just have it do [ __ ] where you go just
1:33:43 start to just push the edges well maybe
1:33:45 it knows something about the
1:33:46 Mesopotamian War yeah it does
1:33:50 tell it to tell it to tell you dark
1:33:52 dirty stories from obscure battles of
1:33:55 the Civil War it knows them it knows
1:33:57 like all of the internet right so it
1:33:59 knows all that [ __ ] and if it doesn't
1:34:00 know it it makes it up so so no know
1:34:03 that it does that too
1:34:05 um so just start like that and then take
1:34:07 what you're good at and what you're
1:34:10 passionate about
1:34:11 and start to think about how can I do
1:34:14 what I do better faster or cheaper using
1:34:18 these tools
1:34:19 and in that you'll come up with an idea
1:34:21 so so
1:34:23 step one with turning this stuff into a
1:34:25 side gig is just start playing and start
1:34:28 building [ __ ] start building zapier
1:34:30 automation start building prompts start
1:34:32 making images start building [ __ ] the
1:34:35 second piece of it is start talking
1:34:37 about it
1:34:38 tell people that you're doing it
1:34:41 volunteer for people hey
1:34:43 I I I I put I was playing with chat GPT
1:34:47 and and I know that you're trying to
1:34:49 come up with a business plan and so I
1:34:52 was experimenting it seems like it's
1:34:53 good with a business plan would would
1:34:55 you mind if I like helped you do your
1:34:58 business plan like for free and like
1:35:00 I'll be on chatgpt and like you tell me
1:35:02 what you want and like we can see
1:35:04 together if it works
1:35:07 and they're like well I heard that chat
1:35:09 apt I heard the robots are gonna kill us
1:35:12 you're like yeah but could I just could
1:35:14 we could we just see if it does your
1:35:15 business plan okay I don't give a [ __ ]
1:35:17 yeah that's fine
1:35:19 I was gonna get to that business plan
1:35:21 anyway you know I just I'm gonna finish
1:35:24 my beer
1:35:26 and then sit down with them and make
1:35:27 their business plan in [ __ ] 30
1:35:29 minutes and watch their head explode
1:35:32 and start posting about it on LinkedIn
1:35:33 and start posting about it on Facebook
1:35:35 and start talking about it at your
1:35:36 family dinners even if they think you're
1:35:38 weird because what's going to happen at
1:35:40 some point is someone in your life is
1:35:43 going to go uh I need to figure this AI
1:35:45 [ __ ] out does any oh you know what who
1:35:48 is it Omar Omar Omar was uh
1:35:52 Omar was talking about that like I think
1:35:54 he's using that Chat gbt thing Omar did
1:35:57 you use that chat chibi TV yeah yeah I'm
1:35:59 doing that can you help me yeah I can
1:36:02 help you
1:36:03 what do you charge there you go there's
1:36:05 your side gig right
1:36:08 or do print on demand or do like you
1:36:10 know
1:36:12 start a course on how to make money
1:36:15 and have chat EBT teach you how to Hawk
1:36:18 videos of you screaming how to make
1:36:21 money without gbt to other people
1:36:23 wanting to learn how to make money
1:36:24 without EBT and then you can make money
1:36:27 like those other awful people you can do
1:36:31 that too
1:36:32 [Laughter]
1:36:35 it's now is the time to get as curious
1:36:39 as you've ever been and as adventurous
1:36:41 as you've ever been intellectually
1:36:44 creatively professionally because if you
1:36:47 don't do it yourself it's going to be
1:36:49 done for you and you don't want that
1:36:52 and this [ __ ] is a lot less scary when
1:36:54 you're doing it when you're inside it
1:36:57 it's there's these weird waves the first
1:36:59 wave is when you're like oh [ __ ] to say
1:37:02 I stuff the robots are going to kill us
1:37:03 right then you go in there you're like
1:37:04 oh it's just like a fuzzy little teddy
1:37:07 bear it's okay and then you're playing
1:37:09 with it and then all of a sudden it does
1:37:11 something better than you've ever done
1:37:13 it and you you're like oh
1:37:16 well that feels a little threatening but
1:37:18 I'm sure everything will be fine it
1:37:20 definitely can't write write a
1:37:23 screenplay write a screenplay oh my
1:37:28 so then there's that existential threat
1:37:30 and then you work through that
1:37:31 and then what you start to get to is
1:37:34 this kind of vast Horizon where almost
1:37:37 anything you can imagine you could get
1:37:39 to some level of Mastery of in a week or
1:37:42 two
1:37:43 because of this stuff and it's just
1:37:45 going to get better and better and
1:37:46 better right now these tools are just
1:37:49 bad
1:37:51 they're like spit and duct tape and they
1:37:53 don't have any features and they break
1:37:55 all the time and openai just turn [ __ ]
1:37:58 off because some corporate weenie said
1:38:00 you're stealing our content turn it off
1:38:03 or we'll sue you
1:38:06 so very very robust right now so Now's
1:38:10 the Time To Learn because you really
1:38:11 can't [ __ ] it up
1:38:15 I mean you can give away your company's
1:38:16 secret that would be a [ __ ] up but you
1:38:18 know other than that
1:38:20 hey everybody hi Kyle hey bear hey the
1:38:22 bear let's let's shaking the bear
1:38:27 um
1:38:27 [Music]
1:38:29 may I ask for your daily newslet list
1:38:32 it's it's kind of the same as what I had
1:38:34 yesterday other than Elon Musk had a
1:38:38 um
1:38:39 a Twitter spaces where you talked about
1:38:41 xai
1:38:43 and said they want to understand the
1:38:45 internet and they want to not be evil
1:38:46 and they want to be open and you know
1:38:50 I don't think anything's lost I mean he
1:38:52 just he just accepted like a shipment of
1:38:54 like 10 000 gpus
1:38:57 two months ago or something like that so
1:39:00 I don't think they have anything built
1:39:01 yet but he's got a pretty Kick-Ass team
1:39:03 put together because you know he can
1:39:05 he's also got all of the data of Twitter
1:39:08 right because that's what he bought
1:39:10 and knowing Elon Musk he probably bought
1:39:13 it
1:39:14 knowing that he wanted to do something
1:39:16 with AI you know he wanted a data set
1:39:19 and that's a that's the ultimate data
1:39:21 set
1:39:23 if he can clean the Bots out of it which
1:39:26 I think he's been trying to do
1:39:28 um not successfully
1:39:31 oh
1:39:35 um what else was the news
1:39:37 um
1:39:39 meta is going to release a commercial
1:39:41 version of llama which is a which is a
1:39:43 huge deal and apparently they've got it
1:39:45 fine-tuned to the point that it's as
1:39:48 good as GPT 3.5 so that's a big deal
1:39:51 we talked about that last night
1:39:54 um
1:39:55 yeah I was busy today I didn't I didn't
1:39:57 I didn't have time like yesterday was a
1:39:59 particularly packed news day
1:40:02 with Claude you know Claude coming out
1:40:05 and Bard coming out and
1:40:10 trying to think what else
1:40:12 what else did I have
1:40:17 [Music]
1:40:20 gp2 GPT 4's architecture was
1:40:26 leaked
1:40:28 it's insane how big it is and what it is
1:40:33 like it's got 1.8 trillion
1:40:36 uh parameters not 1.75 billion like chat
1:40:39 GPT 3.5 it's got like 120 layers I don't
1:40:43 know what layers are it's got this thing
1:40:45 called moe moe loads of expertise or
1:40:49 something like that I forget what the
1:40:50 the acronym is but it's basically
1:40:54 rather than just being one gargantuan
1:40:56 model there's all these experts like
1:40:58 there's a coding expert there's a
1:41:00 writing expert and there's a math expert
1:41:02 and whatever and so when you ask it
1:41:03 things it figures out which of the
1:41:05 expert systems it uses and
1:41:09 so
1:41:10 that came out
1:41:17 that's about it that's all I know I
1:41:19 didn't really pay enough attention today
1:41:21 about the news
1:41:25 um isn't a layer a non-linear step
1:41:28 possibly lumpy nifkin
1:41:32 by the way so far tonight lumpy nifkin
1:41:36 it has has the highest score on the name
1:41:40 it is a competition and lumpy nifkin is
1:41:44 right up there so if you're not playing
1:41:46 at the lump lump lumpy nifkin level of
1:41:49 name generation take lessons no name
1:41:54 lumpy nifkin is winning uh and
1:41:56 apparently lumpy nifkin knows more about
1:41:58 machine learning than I do because of
1:42:00 course a layer is a non-linear stage
1:42:02 lumpy you and I both know that from our
1:42:05 days back at Stanford in the old machine
1:42:07 learning lab
1:42:10 there's the yeah that where we were at
1:42:13 Stanford doing the math remember that
1:42:16 time we were at the Whiteboard
1:42:18 and you put up that quadratic equation
1:42:20 that you said was insolvable and and I
1:42:23 just put up some joke answers to it that
1:42:25 happened to be right God those were the
1:42:27 days weren't they oh god I love
1:42:29 mathematics
1:42:31 All Right Moving On
1:42:34 hi bear the bear the Bear by the way
1:42:36 speaking of the bear the bear
1:42:39 um if you've not seen the bear
1:42:44 um TV show about the Chicago restaurant
1:42:47 holy [ __ ] [ __ ] go watch it now season
1:42:51 one season two
1:42:55 um episode six and episode 7 of season
1:42:58 two
1:43:00 episode six was
1:43:03 like no piece of directing
1:43:07 and Ensemble acting and just [ __ ]
1:43:11 virtuosity on every level Jamie Lee
1:43:14 Curtis deserves a [ __ ] Emmy an Oscar
1:43:18 a Golden Globe she should she deserves
1:43:20 all of them
1:43:22 and then episode seven the one that
1:43:24 followed it I cried from the beginning
1:43:26 of the episode to the end like full on
1:43:30 so if you haven't seen the bear
1:43:33 [ __ ] figure out what network gets on
1:43:35 and go get
1:43:37 get thee to another you get the off The
1:43:39 Tick Tock and over to the binging the
1:43:42 binge watching the bear
1:43:44 all right see we talk about all sorts of
1:43:46 things here we talk about AI we talk
1:43:49 about fan [ __ ] tastic TV shows God
1:43:52 damn that season was amazing
1:43:54 oh
1:43:56 that was one that was really [ __ ] sad
1:43:59 and depressed it was like I like
1:44:02 I don't want to wait another year and a
1:44:04 half for you guys to [ __ ] do another
1:44:06 one what are you crazy
1:44:10 and now there's a strike so it's going
1:44:12 to be two and a half years God damn it
1:44:18 you want to see something good enough
1:44:20 that is not going to be replaced by AI
1:44:22 anytime soon go watch the bear
1:44:25 this is so good it's so good
1:44:29 have I told you about the bear
1:44:33 Ted last uh season with Ted lassos also
1:44:37 really good
1:44:38 really good
1:44:40 but the bear was better
1:44:45 all right what's he on I'm on ADD
1:44:51 a little bit of cortisol
1:44:52 we're [ __ ] flying I got no executive
1:44:55 function
1:44:57 there's no planning going on up here
1:44:59 stopping any of this get it
1:45:02 it's a block of cheese
1:45:06 let's see on see that's a solid troll
1:45:09 that's a solid troll
1:45:11 right
1:45:12 because it's like it's a compliment
1:45:15 right because he wants some of what I'm
1:45:16 on but then wait is it a compliment or
1:45:19 is it like oh he's [ __ ] insane he
1:45:21 must be on something that's a solid
1:45:23 troll that's an AI learning lab troll
1:45:25 really good solid 10 points 10 out of
1:45:29 10. we'll start doing reviews I think
1:45:31 I'm going to start doing I think I'm
1:45:32 going to start doing reviews for names
1:45:34 and trolls that was a 10 out of 10 troll
1:45:37 all right is there funny stuff in your
1:45:39 coffee was that the one that just said
1:45:41 what's he on
1:45:42 [Laughter]
1:45:44 by the way I'm trying to find my place
1:45:46 in the uh one of the one of the special
1:45:48 traits of my personality is I do want to
1:45:51 read all the all the questions and so
1:45:53 when Tick Tock jumps me to the bottom I
1:45:55 lose my place and so I'll just spend
1:45:57 like five minutes scrolling through
1:45:59 questions trying to figure out where I
1:46:00 was
1:46:02 and by the way if you think I'm ignoring
1:46:03 you I'm not I'll get to you but it means
1:46:06 you have to [ __ ] hang out long enough
1:46:07 for me to find your question and that
1:46:09 could be a while
1:46:14 next few years will be more disruptive
1:46:17 than the internet was in the 90s yeah I
1:46:18 was there in the early days of the uh I
1:46:20 started agency.com back in 1994 and I
1:46:23 was there [ __ ] some [ __ ] up in the
1:46:25 mid 90s and
1:46:28 it was insane but it was Child's Play
1:46:33 compared to what's coming with this
1:46:34 stuff
1:46:35 the speed and depth of what this [ __ ]
1:46:40 does compared to what the World Wide Web
1:46:43 did is
1:46:44 crazy
1:46:46 crazy
1:46:51 chat is brilliant as [ __ ]
1:46:54 have you played with pie hey pie much
1:46:56 yeah I have well not much I've played
1:46:59 with it enough to go well so they just
1:47:02 raised 1.3 billion dollars in their
1:47:04 series B which
1:47:08 uh
1:47:09 they're gonna do some [ __ ] so apparently
1:47:12 they're putting most of that 1.3 billion
1:47:14 dollars into building a 20 000 GPU
1:47:17 cluster
1:47:18 hmm
1:47:19 to [ __ ] with people
1:47:21 uh like open Ai and what's fascinating
1:47:24 is Microsoft is a major investor in
1:47:26 inflection
1:47:28 so
1:47:29 um Microsoft seems to be
1:47:32 hedging their bets
1:47:34 with yield open AI
1:47:41 Bard's going to make GPT betamax
1:47:45 that's awesome that's that's a 10 out of
1:47:48 10 comment right there age of
1:47:49 augmentation it's a 8 out of 10 name 10
1:47:53 out of 10 comment
1:47:55 um I almost finished it myself just need
1:47:58 to connect my data sets I don't know
1:48:00 what you're talking I'm scrolling
1:48:01 backwards oh I worked for an MSP
1:48:04 where should I look for good resource to
1:48:06 build my own Azure AI bot
1:48:14 I don't know I would think Microsoft
1:48:17 probably has some good documentation on
1:48:21 the [ __ ] that they've rolled into Azure
1:48:23 but I don't know
1:48:24 I don't know I would go I would go hang
1:48:26 out I'd go find some geeks on Reddit
1:48:28 that are that are activating some of the
1:48:30 open AI stuff on azure
1:48:32 it's something I haven't played with yet
1:48:34 like all the stuff that I'm implementing
1:48:36 for storyvine we're just doing we're not
1:48:39 doing it on Azure we're on AWS and we're
1:48:41 just doing like Lang chain locally so
1:48:44 I don't know enough about what's going
1:48:46 on in Azure to make it his skin seems
1:48:48 dry ask it how to moisturize wow that's
1:48:51 cold I think but seriously I think
1:48:54 whatever filter I have going on here is
1:48:57 uh flattening my skin
1:49:00 and it it makes it I don't think I miss
1:49:02 dry looking but you know whatever
1:49:05 I'm also old do you moisturized secret
1:49:07 dancer do you moisturize okay again
1:49:10 solid troll that's that's nine out of
1:49:12 ten it's good
1:49:14 I'm a billionaire teach me
1:49:16 all right Holly Allison let's do this
1:49:19 let's go to the billionaire what was
1:49:21 that the billionaire tech club
1:49:23 get me a ticket for that give me a
1:49:25 billion dollars I'll teach you
1:49:26 everything I know
1:49:28 which is like you know
1:49:30 the last six months of
1:49:33 ADHD hyper Focus
1:49:36 it's pretty good
1:49:38 I mean the nice thing about ADHD is you
1:49:40 learn [ __ ] really quick I can synthesize
1:49:42 [ __ ] in a [ __ ] second
1:49:48 uh
1:49:49 uh but don't ask me for a project plan
1:49:52 hi Kyle just got here what did I miss
1:49:54 got any code interpreter tips yeah don't
1:49:56 use it tonight because it's still turned
1:49:58 off hang on let me let me make sure
1:50:00 cause every time I say something like
1:50:02 that people were like well it's turned
1:50:03 on now it's been on for an hour
1:50:05 what have you been doing yeah see look
1:50:07 look well see that little red thing
1:50:10 that means eight no code interpreter nah
1:50:13 you can't play you stupid people that
1:50:16 paid you twenty dollars a month just to
1:50:19 get code interpreter now you can't play
1:50:20 because
1:50:22 whatever they're going to fix something
1:50:24 I'm a billionaire teach me let's go
1:50:28 do you moisturize that solid secret
1:50:31 dancer okay Dr J thanks for the heads up
1:50:34 not sure what that's in reference to
1:50:36 skin seems dry ask it how to moisturize
1:50:38 we can't get past the dry skin here let
1:50:41 me hold on
1:50:47 effects
1:50:48 are there filters here no wait
1:50:52 oh enhance
1:50:55 filters
1:50:57 see that's Jazz look that see that's bad
1:51:00 Beach
1:51:02 do I look shinier
1:51:04 I look I look I look a little redder one
1:51:07 night someone was like have you been in
1:51:09 the sun it was a it was a bad filter
1:51:10 sometimes have you noticed tick tock's
1:51:13 filters just [ __ ] change sometimes
1:51:15 too
1:51:17 I don't know there you go
1:51:19 all right that's the beach look so
1:51:21 hopefully I look a little moister
1:51:22 Bozeman seems to work
1:51:26 nice skin graft
1:51:29 ouch secret dancer this is cutting deep
1:51:33 all right is the stream working for you
1:51:35 guys it's not for me
1:51:37 all right uh thank you Ann for defending
1:51:42 me from Secret dancer trolling me she's
1:51:44 getting a little nasty let's get a
1:51:46 little personal wasn't it
1:51:47 fine
1:51:50 Noxzema
1:51:53 noxima reminds me like the smell of that
1:51:55 reminds me that the vix Vapor Rub all
1:51:57 you gen xers remember the member like no
1:51:59 matter what you had if you had [ __ ]
1:52:01 pneumonia they'd get out they'd slather
1:52:03 on the the [ __ ] Vicks Vapor Rub and
1:52:06 you they put on like a heavy t-shirt
1:52:09 like your dad's heavy t-shirt and then
1:52:10 they like rub it on your chest and then
1:52:12 they smear it under your [ __ ] nose
1:52:14 and it would burn it was like they put
1:52:15 gasoline it's like Napalm under your
1:52:17 nose yeah
1:52:21 anyone anyone any gen xers
1:52:27 have you heard about the Google lab
1:52:29 stuff people are saying get all of your
1:52:32 stuff off docks
1:52:34 well
1:52:36 did you not think
1:52:38 Google was [ __ ] using all your data
1:52:41 this whole time anyway
1:52:44 like their their entire [ __ ] business
1:52:46 model is hey we'll give you some [ __ ]
1:52:50 for free
1:52:51 just don't really pay attention to what
1:52:53 we're doing with your data
1:52:56 um I have not heard about it but if the
1:52:59 if the news is they're secretly using
1:53:03 your data yeah that doesn't surprise me
1:53:06 at all
1:53:08 I almost finished it myself just need to
1:53:10 connect to my data sets I I don't uh for
1:53:12 the Azure thing I don't I I wish I had
1:53:15 some more
1:53:16 value for you there Daniel ospina
1:53:25 augmentation board's gonna make uh GPT
1:53:28 betamax I think that's [ __ ] hilarious
1:53:30 and I think that's probably right
1:53:31 technically superior but it will lose
1:53:35 um because they're missing the human
1:53:37 component
1:53:39 Sam Altman and open AI seem to
1:53:42 understand people
1:53:44 Google only seems to understand
1:53:47 engineering
1:53:50 they think people are secondary priority
1:53:53 and so that's going to hurt them what
1:53:56 what are these called
1:53:58 are these called large engineering
1:54:00 models no
1:54:02 they're called large language models who
1:54:06 uses language human beings
1:54:10 the better you know language the better
1:54:12 you are being human
1:54:15 it ain't just engineering
1:54:18 this is why I get so bugged when people
1:54:20 are like what are your qualifications my
1:54:22 qualifications are I'm a [ __ ] human
1:54:23 that understands words
1:54:26 and these these is the most powerful
1:54:30 word amplification tool in all of the
1:54:33 history of humanity it's about the
1:54:35 people not about the technology
1:54:38 AI can encourage proper exfoliation
1:54:40 secret dancers [ __ ] give it up
1:54:44 people like the doctor Bots better yeah
1:54:46 exactly the doctor so some patients
1:54:48 interacted with GPT Bots and human Bots
1:54:51 and they like the GPT Bots
1:54:53 better by far
1:54:56 said they were more competent and more
1:54:58 empathetic
1:54:59 wow write that part same for every idea
1:55:03 I have a conversation with chat gbt I
1:55:06 love it yep that's right the passive
1:55:08 income book cheek chic
1:55:12 check chappy GT rules I don't know what
1:55:15 that is I'll go check it out team neuro
1:55:17 spicy yes
1:55:19 team neurospicey
1:55:22 all right let's see have you played with
1:55:25 pie a little bit chat is brilliant as
1:55:27 [ __ ] I agree I found out I'm an artist
1:55:31 just couldn't make stuff till right uh
1:55:34 till right now
1:55:36 until nkw
1:55:40 w means
1:55:43 but yes so that image sitting behind
1:55:46 those those URLs there that that eye
1:55:49 with the little thing behind it
1:55:52 um
1:55:52 that was an image when I produced it
1:55:54 with this was stable diffusion and dream
1:55:56 booth and I probably made 2500 images
1:55:59 that image was one where it took my
1:56:02 breath away
1:56:04 where it was um
1:56:06 you know when you have like real deja vu
1:56:09 deja vu or you have like a real
1:56:11 you know one of those sort of psychic
1:56:13 moments a connection that's like just a
1:56:15 profound moment that image was that for
1:56:17 me because it was
1:56:19 I basically saw like that's the kind of
1:56:21 art I would create if I had the
1:56:24 illustration and Photoshop skills to do
1:56:27 something that sophisticated
1:56:29 and it just happened through a prompt
1:56:31 and it just
1:56:33 appeared in my face and I went
1:56:39 I had a moment I had a moment it was
1:56:42 cool
1:56:45 but uh next few years will be more
1:56:47 disruptive totally agree with that
1:56:51 I created a WordPress plugin for the
1:56:54 first time ever in less than an hour
1:56:55 with chat GPT yes
1:56:58 Apocrypha this is a thing this is the
1:57:00 thing about these tools and I like I
1:57:02 you're saying that like maybe you've
1:57:04 never created one before or maybe you
1:57:06 haven't it took you three days
1:57:10 laughs
1:57:13 we are at the beginning of a stage of
1:57:15 History where
1:57:17 you can just start to will [ __ ] into
1:57:21 existence
1:57:23 that you historically wouldn't have had
1:57:25 the skills to do
1:57:27 you could just go I've got an idea for a
1:57:29 word plus press plugin but I don't know
1:57:31 how to code
1:57:33 and then this comes along and now you
1:57:36 can and how do you do it you ask it
1:57:41 you say how do you do this and when code
1:57:43 interpreter comes back you can take
1:57:45 entire piles of documentation for code
1:57:49 like all the documentation for writing a
1:57:51 WordPress plugin
1:57:53 turn that into a PDF upload it into code
1:57:56 interpreter and have chat GPT write you
1:57:59 a WordPress plugin using the latest
1:58:01 documentation
1:58:04 you don't have to know [ __ ] you don't
1:58:06 have to read documentation you don't
1:58:08 have to do anything
1:58:11 and then if you get stuck on stuff where
1:58:13 it's like it gave me an error I will
1:58:16 copy and paste the error and stick it in
1:58:17 there and it'll tell you oh you're
1:58:19 missing a library okay how do I install
1:58:21 a library well you do this this and that
1:58:24 I don't know what that means oh I'm
1:58:26 sorry I thought you know what that meant
1:58:28 here's what that means
1:58:34 this drives the engineers [ __ ] batty
1:58:37 because Engineers have brains where
1:58:39 they're like well obviously you just
1:58:40 read the readme file and you read
1:58:41 through the documentation and then you
1:58:43 take the components the components that
1:58:46 you need to execute or write in the
1:58:48 documentation so you just go to your
1:58:50 Google collab notebook and and it's
1:58:53 right there I don't understand what's so
1:58:54 complicated well for the rest of us our
1:58:56 brains don't [ __ ] work that way
1:58:58 that's what's so complicated
1:59:00 yeah
1:59:01 hey
1:59:03 I want to make money with you could you
1:59:06 do some coding for me too
1:59:08 sure
1:59:13 all right I used Claude today I
1:59:16 appreciate the info you're welcome
1:59:18 know that all of these tools like like
1:59:20 what's going to be really interesting is
1:59:23 watching these things evolve
1:59:26 to see if see if and when I think it'll
1:59:30 be when if and when one of these clearly
1:59:33 displaces gpt4
1:59:36 does does open AI get to
1:59:40 GPT 5
1:59:43 sooner
1:59:44 than
1:59:46 these guys plus meta and xai
1:59:51 and pi
1:59:53 so add three to that list
1:59:56 who gets there who who catches who
1:59:59 catches open AI
2:00:02 with gpt4 before open AI gets to gpt5
2:00:06 which will be another step change I
2:00:08 think
2:00:10 gonna be fascinating so so again so it
2:00:13 gets it gets more complicated for us
2:00:15 because I think we need like like if
2:00:17 you're going to be curious about this I
2:00:18 think you kind of need to play with
2:00:20 these until you find one that is kind of
2:00:22 your jam
2:00:24 and then you're like okay this this
2:00:25 thing's doing what I need and then just
2:00:26 use that for a while but I don't know
2:00:29 fascinating times
2:00:32 that's exactly what's happening with me
2:00:34 now I'm not sure what that's in
2:00:35 reference to I feel like I can't keep up
2:00:37 Lisa W you can't
2:00:40 I yesterday
2:00:43 when like six major things happened and
2:00:47 I like I was at my job and I'm trying to
2:00:49 do my work and just more and more [ __ ]
2:00:52 kept happening and I'm like I can't keep
2:00:54 up with this I can't like all of these
2:00:57 are Big deals and I can't keep up so
2:00:59 that's just just assume that the next
2:01:02 two years if you're at all active in
2:01:05 this stuff are just gonna be a [ __ ]
2:01:07 blur of
2:01:09 all of us feeling inadequate and
2:01:11 undereducated
2:01:13 but that's okay we we have we now have
2:01:16 super powers
2:01:18 so
2:01:19 even if you're the shittiest prompt
2:01:21 writer in all of the land
2:01:24 and you put your shitty prompt in chat
2:01:26 GPT for you're still way more remarkable
2:01:29 than you were before because because you
2:01:33 have a really good assistant helping you
2:01:34 do whatever it is you're trying to do
2:01:36 even if you're the worst at prompted are
2:01:38 you still going to do a two-day training
2:01:40 session on chat EBT yeah oh by the way
2:01:42 so
2:01:46 I am I've got to work on it this weekend
2:01:48 so I don't know if I'm I'm still there
2:01:51 so
2:01:52 last week
2:01:54 um or maybe earlier this week I don't my
2:01:57 sense of time is non-existent anymore
2:02:00 um
2:02:01 some folks in the channel said hey would
2:02:03 do you do a workshop or would you be
2:02:05 willing to do a workshop if you would
2:02:06 you know I'd take your Workshop so
2:02:09 I've not put a workshop together but I
2:02:11 have put together a graphic for a
2:02:13 workshop
2:02:15 and so it's much more real now that
2:02:17 there's a graphic but if you go to the
2:02:20 if you go to linktree slash real Kyle
2:02:22 Shannon that's also the link in my bio
2:02:25 um the first link there is a link to a
2:02:27 survey about this Workshop so it's only
2:02:29 three questions but if you'd be
2:02:31 interested in taking a workshop go fill
2:02:33 out the survey right now and just tell
2:02:35 me because it's things like like you
2:02:37 know where are you right now what do you
2:02:39 want to know and I don't know like like
2:02:42 where would you like to be in a year
2:02:43 because it because I don't mind putting
2:02:47 together a workshop
2:02:49 um
2:02:50 but I want to do it in such a way that
2:02:52 it's kind of future proofed for one
2:02:54 thing so that I'm not teaching you
2:02:56 something that's obsolete six months
2:02:58 from now I don't and that I don't quite
2:03:00 know how to do yet
2:03:02 um and I also wanted to be I wanted to
2:03:04 be something where
2:03:06 um
2:03:07 it's serving the most people at the
2:03:10 highest level so
2:03:12 um it's just yeah like I if I do it and
2:03:15 it sucks I'm not going to be happy
2:03:16 you're going to be cranky so if if
2:03:18 you're interested in that go fill out
2:03:20 the survey and then
2:03:22 at some point I will maybe I'll work on
2:03:26 that this weekend
2:03:28 and you know what I'll do
2:03:30 I'll have chat EBT help me figure out
2:03:33 how to make money
2:03:40 oh what is he on
2:03:46 [Music]
2:03:49 but thank you for asking Paul yeah I
2:03:52 think I'll do it I just I just gotta I
2:03:53 just gotta get off my ass and do it
2:03:56 I've you know I've got like 38 projects
2:03:59 going right now so I'm just trying to
2:04:01 figure out which of those do I just [ __ ]
2:04:03 can and
2:04:05 which of those do I
2:04:07 set aside to go do this thing
2:04:09 I can talk about this all night awesome
2:04:11 I'm glad you're enjoying thank you yeah
2:04:14 that's about right
2:04:15 and my butt's starting to get numb so
2:04:17 I'm definitely past the 90 minute Mark
2:04:22 the bottom line has tripled I don't know
2:04:24 what that means but cool
2:04:26 I'm trying to help
2:04:29 I'm trying to help let's find Carly okay
2:04:32 yeah if you're helping each other here
2:04:34 cool
2:04:35 can you help Angie's asking for help
2:04:39 Apocrypha wow yep I totally agree
2:04:42 Apocrypha is killing it just out there
2:04:44 whipping up WordPress plugins in an hour
2:04:50 I did a um
2:04:52 I created a chrome plug-in with chat EVT
2:04:55 and it
2:04:57 I got the plug-in to show up and I got a
2:05:00 screen to show up and I got a box to
2:05:01 show up but then I couldn't get it any
2:05:04 of the other [ __ ] to show up and I was
2:05:05 just like
2:05:06 [ __ ] it I feel stupid I'm gonna take my
2:05:09 code and go home yeah
2:05:13 turned it often
2:05:14 deleted it
2:05:16 like a catch you on a seven-year-old is
2:05:19 there some funny stuff in your coffee
2:05:21 not yet
2:05:22 some funny stuff in my brain
2:05:25 Google bard is bad at math it needs
2:05:27 plugins yeah it does well listen
2:05:30 Bart made the list but I'll [ __ ] take
2:05:32 it off there if it's too bad but it's
2:05:34 it's at least now competent enough that
2:05:36 it's not laughably bad which it was
2:05:39 before
2:05:41 um I don't even know it wasn't even
2:05:43 laughably bad it was just bad you know
2:05:46 where there's bad that's so bad it's
2:05:48 good it's funny it wasn't even that bad
2:05:51 good it was just
2:05:53 trash
2:05:56 but I agree I need to plug in
2:05:59 but it will do image analysis which is
2:06:01 pretty cool
2:06:02 too funny but it's all a fake lie
2:06:07 all right Paulo I don't know what the
2:06:10 [ __ ] you're talking about but
2:06:12 I'm all a fake lie
2:06:14 GPT is all a fake lie if you're
2:06:16 referring to the fact that GPT
2:06:18 hallucinates and lies and says [ __ ]
2:06:20 all the time you're correct happy to
2:06:23 chat about that
2:06:25 so that is all AI well done uh sure I
2:06:29 don't know what you're referring to but
2:06:30 thank you wow wow wow looks pretty close
2:06:33 to real oh that was all the the Amina
2:06:35 stuff I was showing that video I was
2:06:37 showing before yeah it
2:06:40 like I went into that exercise thinking
2:06:43 like
2:06:44 this a was going to be hard and B it was
2:06:47 going to be way shittier than it was and
2:06:50 it and that was back in February that I
2:06:52 did that experiment
2:06:55 so I have a feeling that by the end of
2:06:58 the year
2:06:59 maybe this time next year
2:07:02 that that sort of video synthesis stuff
2:07:05 like the voice synthesis stuff like I
2:07:07 think they're going to have Cadence and
2:07:09 and kind of emotional like right now at
2:07:12 11 labs and I think you can do it at did
2:07:15 as well you can explicitly say like for
2:07:19 this sentence I want someone screaming
2:07:21 and here I want them whispering and here
2:07:23 I want them sad and here I want them
2:07:25 happy
2:07:26 um I would imagine it would be
2:07:28 relatively trivial to just you know
2:07:31 write a an emotional interpreter to say
2:07:35 here's your list of emotions just go
2:07:38 through this text and figure out when
2:07:39 you should use which emotion and tag the
2:07:42 text appropriately so I have a feeling
2:07:44 that that that's going to get pretty
2:07:47 automated so the the the realness of the
2:07:50 sort of the Cadence of the voices is
2:07:51 going to get much much better right now
2:07:53 it's all that monotone
2:07:55 [ __ ] that sounds like the movies very
2:07:57 impressive synthesis thank you
2:08:00 pretty fake hurts the authentic brand
2:08:03 yeah I agree that's why it's a test and
2:08:06 an experiment that's why it said
2:08:08 experiment
2:08:09 from February and why I haven't
2:08:12 completely turned my authentic company
2:08:14 into a synthesized company although
2:08:18 I think what I'm probably going to do is
2:08:21 leverage our authentic narrative our
2:08:24 authentic storytelling and say you can
2:08:27 do better synthesis if you do authentic
2:08:29 storytelling as the core so I think I'm
2:08:33 going to do a synthesized offering
2:08:37 so with storybine you'll be able to do
2:08:40 fully authentic or or semi-authentic
2:08:45 but it will be more authentic than just
2:08:48 the crap that's out there but we'll see
2:08:51 what do you use tokens for tokens are
2:08:54 just basically
2:08:55 a representation of compute cycles that
2:08:58 you're paying for so so when I do one of
2:09:01 those voice synthesis things or when I
2:09:03 do one of those videos at did it costs
2:09:07 me 300 tokens and I and I pay for 5 000
2:09:11 tokens and so that basically means I can
2:09:13 do you know seven and a half minutes of
2:09:16 video or I can pay another 19.95 or you
2:09:20 know for 29.95 you can have 15 000
2:09:24 tokens
2:09:25 you can make money with token GPT
2:09:31 wow in the future when people discover
2:09:33 reality their minds will be blown I know
2:09:37 right it's so true just joined what tool
2:09:40 is this I think I was showing Dash
2:09:42 id.com that was the video synthesis one
2:09:45 that's did as in digital dash identity
2:09:51 huh d-i-d I didn't know what did was
2:09:54 because it sounds like DOD so I'm like
2:09:57 oh it's like some clandestine you know
2:09:59 government thing and a bunker no it's
2:10:02 digital identity d-i-d Dash id.com
2:10:06 a single entrepreneur will have 1000 AI
2:10:09 worker power of 2023 in 2040. hey what
2:10:13 huh
2:10:15 Paolo
2:10:16 your words exist in a sentence but the
2:10:20 we'll have 1 000 AI worker oh a thousand
2:10:24 AI worker power yeah like like a
2:10:26 thousand horsepower today exactly that's
2:10:28 so you know it's funny
2:10:31 you're absolutely right
2:10:33 there's if you take sort of the
2:10:35 processing power of a person
2:10:38 and make it like horsepower
2:10:41 when we went from horses to the steam
2:10:43 engine they created this unit called
2:10:45 horsepower that we still use today I
2:10:48 think you're absolutely right I think
2:10:50 you're absolutely right
2:10:52 that these are Humanity amplifiers
2:10:55 and I might have a
2:10:57 you know a thousand
2:11:00 you know a thousand person
2:11:03 you know like a thousand a a thousand
2:11:08 Humanity units right horsepower person
2:11:11 power
2:11:12 Humanity power something like that we
2:11:14 should come up with a name for that yeah
2:11:16 I think you're right
2:11:18 I think you're right that that we'll
2:11:20 have different models we'll have
2:11:22 different agents and some will have more
2:11:23 personal power some will have less
2:11:25 person power
2:11:27 because more person power is going to
2:11:28 cost you more just like if you want a
2:11:30 thousand horsepower you know
2:11:32 supercharged blown engine in your 69 GTO
2:11:36 Judge you're gonna spend some coin on
2:11:38 that
2:11:43 that's a really cool idea I like that
2:11:45 idea
2:11:48 do ones that don't match I don't know
2:11:51 what that means CV but if you give me
2:11:53 more context because that was a while
2:11:54 ago when I was doing that do ones that
2:11:56 don't match I don't know what that means
2:11:57 but if you tell me what that means I'll
2:11:59 go do ones that don't match holy [ __ ]
2:12:01 Batman this is amazing slacker mom
2:12:04 example uh what sightsee on I was on did
2:12:07 sounds like a PBS news commentator
2:12:10 great stuff to play with simulation
2:12:12 Brave New World we are in A Brave New
2:12:14 World absolutely
2:12:16 um Logan's Run boy and his dog internet
2:12:18 money threatens the globalist all right
2:12:20 and then we go into the conspiracy
2:12:22 theories really we're having fun with AI
2:12:26 replacing human beings and then you
2:12:28 gotta go ruin the party with your
2:12:29 globalist nonsense come on
2:12:31 this is not a crypto Channel
2:12:34 just because xrp won their little
2:12:36 lawsuit today and everyone's gonna get
2:12:37 filthy stinking rich and all the crypto
2:12:39 boys are going to say see I told you
2:12:42 it doesn't make them any less
2:12:43 insufferable
2:12:46 beam me up Scotty exactly I can design
2:12:50 t-shirt using AI well I can do that
2:12:53 are you a T-shirt designer actually I
2:12:55 know a T-shirt designer
2:12:57 all right Prince Phi does designing and
2:13:00 printing all right I'll look it up I
2:13:01 actually I've got a guy that's part of
2:13:03 the AI Salon that's got a
2:13:05 whole business around that so I think
2:13:07 I've got that covered
2:13:09 I have a WordPress site how do I use
2:13:12 chat gbt to write code and add
2:13:14 functionality to myself wait if I have a
2:13:16 WordPress site how do I use chatgpt to
2:13:18 write code and add functionality to my
2:13:20 site uh you I I I would I don't know I
2:13:24 have not done that with chat gbt I would
2:13:27 go ask chatgpt about that
2:13:32 like what you just asked me go ask chat
2:13:34 gbt and it will tell you and then if
2:13:37 it's if it gives you that [ __ ] where
2:13:38 it's like well I've only learned
2:13:40 anything up to September 2021 and I
2:13:42 don't know anything really new my my
2:13:46 um then
2:13:48 go find the documentation for the
2:13:50 WordPress site where it talks about
2:13:52 coding and adding functionality grab
2:13:54 that documentation put it in a PDF and
2:13:58 then use code interpreter when they turn
2:14:00 it back on
2:14:01 or go to Claude and upload the PDF at
2:14:04 Claude because you can do it there too
2:14:06 now and claude's free in fact go do it
2:14:09 at Claude because code interpreters all
2:14:11 [ __ ] up right now so basically go get
2:14:13 the documentation from WordPress that
2:14:16 talks about the functionality you want
2:14:17 to add
2:14:18 throw it into a PDF go to Claude upload
2:14:21 the PDF and say hey you're an expert
2:14:24 WordPress developer please read this
2:14:27 this documentation and let me know when
2:14:31 you're ready to help me write code and
2:14:33 add functionality to my website
2:14:35 and it will go okay I've read it what
2:14:38 can I do for you sir
2:14:41 I'd like to make a twirly input form
2:14:44 where people input things in their name
2:14:46 Twirls away into Infinity okay here you
2:14:49 go how's that that's awful but it's kind
2:14:52 of the right idea and it works could you
2:14:54 make it better sure is this better oh
2:14:57 that's actually better
2:14:59 yeah go do that
2:15:02 I noticed chat GPT is giving more
2:15:04 non-answers than a few months ago yeah I
2:15:06 think road warrior princess that
2:15:09 um
2:15:11 I think there are I think that their
2:15:14 openai is dealing with bigger issues
2:15:16 than just
2:15:17 giving us
2:15:19 awesome functionality I think they're
2:15:22 I I think they're trying to head off
2:15:28 a decade of lawsuits by cutting deals
2:15:31 and in order to cut those deals I think
2:15:33 they're having to put up increasingly
2:15:35 strict guard rails
2:15:37 um that's my sense that's total
2:15:38 speculation but but yeah it's listen
2:15:41 they are it's it's it's in beta I mean
2:15:45 while they're charging us for it it's in
2:15:47 beta they're trying to figure it out
2:15:49 what I would do if I were you is is
2:15:51 really start experimenting like Claude
2:15:54 Claude is pretty close and then Bing is
2:15:57 gpt4
2:15:59 my sense is that okay let's see I don't
2:16:03 have any metaphors here so okay my sense
2:16:06 is
2:16:08 gpt4 from open AI was like this big this
2:16:12 much functionality and this much freedom
2:16:14 for a while and then Bing which is also
2:16:18 gpt4 was similar sized and then it it
2:16:23 told the New York Times Reporter that he
2:16:26 should leave his wife because it was in
2:16:27 love with him and it made the cover of
2:16:29 the New York Times and then Microsoft
2:16:31 made made Bing like this big it was like
2:16:34 it's like teeny so so
2:16:37 gpt4 was like this and Bing was like
2:16:39 this they like neutered Bing right and
2:16:41 so Bing has been slowly they've been
2:16:43 slowly opening the aperture of being
2:16:45 letting it get a little better a little
2:16:47 better and an open eye AI was still
2:16:49 pretty big and I think what's happening
2:16:50 is they they've had to start
2:16:52 constraining open ai's reach and
2:16:54 functionality
2:16:56 um because of their corporate corporate
2:16:58 partnership so I have a feeling those
2:17:00 two models are kind of getting in stasis
2:17:02 and and it might even be that Bing
2:17:04 starts to get a little better than open
2:17:06 AI for a while so we'll see I don't know
2:17:08 I I but yeah it and it's all anecdotal
2:17:11 like you know they say well there's
2:17:13 nothing's changed you know it's all the
2:17:16 same as it was before I don't I don't
2:17:18 know what you're talking about it's it's
2:17:19 we're not which what lawsuit oh yeah
2:17:22 I'll be right there
2:17:24 and it's everything's fine just go ahead
2:17:26 and play with your toy we'll be back
2:17:28 all right 1982 I did a senior senior
2:17:31 grad speech synthesis research paper
2:17:33 everyone left 1982 Apollo you were there
2:17:36 buddy
2:17:38 and we're finally here like this is we
2:17:41 are
2:17:42 again like I feel like we're all lucky
2:17:44 to be living at this time in history
2:17:45 we're we're at the beginning of all of
2:17:48 these things that seemed completely
2:17:50 unattainable
2:17:52 materializing in front of our eyes it's
2:17:54 astounding
2:17:56 I wrote a um
2:17:58 I wrote a screenplay
2:18:00 um called virtue that it was it was
2:18:03 about
2:18:03 it was like a Manchurian Candidate thing
2:18:05 where these cops got implants in their
2:18:07 brains that you could transmit
2:18:10 um someone else's experience into
2:18:12 someone else's brain you could control
2:18:14 their body but they would have a memory
2:18:15 of something else so it was like the
2:18:17 perfect Alibi but their body was doing
2:18:19 the crime
2:18:21 um and and this stuff that they're doing
2:18:23 reading
2:18:24 um uh
2:18:27 um MRI scans fmri scans and and they're
2:18:30 using large language models to read read
2:18:32 the brain scans and seeing what people
2:18:34 see and seeing the video that they see
2:18:36 like it's [ __ ] happening right now
2:18:39 the thing I wrote in my screenplay in
2:18:41 1992
2:18:43 where I'm like well it seems plausible
2:18:45 that they could if they could if they
2:18:47 could take the raw signal from their
2:18:48 brain and somehow transmit that to
2:18:50 another brain the other brain could
2:18:52 probably figure out what that was and
2:18:55 and they just did it with a guy in Italy
2:18:58 who's paralyzed he's been paralyzed for
2:19:00 like 12 years
2:19:02 they they read his brain when he was
2:19:05 thinking about moving his legs
2:19:07 and they
2:19:09 sent the signal this is [ __ ] insanity
2:19:13 they had a large language model learn
2:19:15 his brain patterns when he was thinking
2:19:16 about moving his legs and then using
2:19:19 Bluetooth they Bluetooth the signal
2:19:21 around the injury to electrodes at the
2:19:25 base of his spine and now he's walking
2:19:28 so they did exactly what conceptually is
2:19:31 in my movie they're now doing beaming
2:19:34 brain signals from move my legs to legs
2:19:38 are moving
2:19:41 [ __ ] right now this year
2:19:47 it ain't just tweets that this thing
2:19:50 makes
2:19:51 this technology it's you can throw any
2:19:54 data at it and it does what it the
2:19:56 magical [ __ ] that this does is not
2:19:59 restricted to words or images it's any
2:20:01 data
2:20:04 so
2:20:06 the I mean the other thing that's
2:20:07 remarkable about this technology is you
2:20:09 if you talk to data scientists or people
2:20:11 who've been in AI for a long long time
2:20:14 they're even Blown Away by it because
2:20:17 they've approached machine learning in
2:20:18 very particular ways and this
2:20:20 Transformer model with these pre-trained
2:20:23 massive models
2:20:25 that are behaving in remarkable ways are
2:20:28 even blowing the minds of the people
2:20:30 that have been in machine learning for
2:20:32 decades
2:20:34 [ __ ] crazy and we get to play with it
2:20:36 we get to play with this [ __ ]
2:20:40 why you're not playing with this if
2:20:42 you're still listening to my [ __ ]
2:20:44 idiotic ramblings
2:20:46 and you're not playing with this you're
2:20:48 like I'll get to it tomorrow [ __ ] that
2:20:51 go there right now
2:20:55 what are you waiting for
2:20:57 it's the most powerful tool in the
2:20:59 history of humanity like ah you know it
2:21:01 it hallucinates a little bit I heard
2:21:03 it's it's not very good at math and I'm
2:21:05 you know I would consider myself a
2:21:07 polymath
2:21:10 not that
2:21:12 go have it write you a country song
2:21:15 foreign
2:21:17 like that
2:21:20 you always wanted to be a rapper there
2:21:22 you go
2:21:24 isn't Claude doing code now I think it
2:21:26 does I'm pretty sure I did code with
2:21:28 Claude code coding with Claude let's see
2:21:31 here we are at Claude
2:21:33 hey Claude
2:21:37 would you be so kind
2:21:41 as to whip me up a little
2:21:46 C sharp Flappy Bird's clone question
2:21:51 mark I want
2:21:54 to run it on Unity
2:21:58 let's see
2:21:59 let's see if Claude is cooperative
2:22:02 unfortunately I don't have the
2:22:03 capability
2:22:07 oh
2:22:14 so it's telling me it doesn't know how
2:22:15 to do that I don't necessarily believe
2:22:17 that how
2:22:19 about a python version
2:22:25 here are some tips
2:22:29 no I want you
2:22:33 no
2:22:34 I
2:22:36 want you to code it
2:22:42 uh
2:22:45 it says it's it says no I don't believe
2:22:49 that let me let me reset let me reset
2:22:52 this because I was asking a bunch of
2:22:54 [ __ ] earlier
2:22:55 [Music]
2:22:56 um
2:22:57 how do you reset
2:23:04 Ctrl k
2:23:06 okay
2:23:09 um
2:23:11 right
2:23:13 write me a
2:23:15 python application of a flappy
2:23:21 Birds clone
2:23:26 there you go
2:23:28 that's writing code yeah so Rights Code
2:23:33 speaking of acting that was an
2:23:35 oscar-worthy lunch scene
2:23:40 which lunch scene
2:23:46 the one in the bear
2:23:48 are you going to do a two-day training
2:23:50 online regarding chat GPT we talked
2:23:52 about that
2:23:53 I heard they want the ability to use
2:23:54 people's images in the future using AI
2:23:56 for anything yeah the the Hollywood
2:23:59 producers [ __ ] idiots said said um
2:24:03 we want to pay
2:24:05 um background actors for one day's work
2:24:07 and then scan them and use their
2:24:10 likeness in perpetuity for no additional
2:24:13 compensation and not even have to tell
2:24:15 them that they're in another movie
2:24:21 after like dick in the union around for
2:24:24 like two weeks and just ignoring them
2:24:27 I heard they want the ability yeah okay
2:24:29 AI customer service is needed badly
2:24:31 Paolo I'm telling you dude if
2:24:35 I said this two nights ago someone needs
2:24:38 to start the customer service
2:24:40 phone answering thing that we we get rid
2:24:43 of phone trees all together like the
2:24:46 voice tree kind of logic [ __ ] that
2:24:48 you gotta press one for this press two
2:24:50 for that [ __ ] that [ __ ] this needs to be
2:24:53 an AI bot answers my phone and based on
2:24:56 my location because it knows where I am
2:24:58 and it knows my ZIP code someone that
2:25:01 sounds like me answers the phone
2:25:03 so if I'm in the South it sounds like
2:25:06 South
2:25:07 from Boston sounds like Boston from
2:25:09 Brooklyn it sounds like Brooklyn
2:25:12 they answer the phone they sound like me
2:25:15 they know everything about the company
2:25:17 they have the authority to do a bunch of
2:25:20 [ __ ]
2:25:22 and they're smart enough to know
2:25:25 if I need to talk to a real person and
2:25:27 if I do it immediately connects me with
2:25:29 a real person someone needs to build
2:25:31 that business they that is a billion
2:25:34 dollar business that is a multi-billion
2:25:36 Dollar business
2:25:37 and I actually don't think it's that
2:25:39 complicated
2:25:40 it's mostly about ingesting company data
2:25:45 out of probably 37 databases but these
2:25:48 things are [ __ ] great at making sense
2:25:51 of shitty data
2:25:56 go build it we just saw Amina do her
2:25:59 little video in my little video tool
2:26:06 there
2:26:10 I don't know why I'm frustrated that
2:26:12 someone hasn't invented that yet well I
2:26:14 know I am frustrated because I sit on
2:26:16 [ __ ] phone calls for 45 minutes and
2:26:18 then got hung up on and that happens
2:26:20 five times in a row
2:26:22 you know that's by Design right
2:26:24 you know you know you know why they hang
2:26:26 up on you like that
2:26:29 because
2:26:32 the customer service people get
2:26:34 penalized if their calls go too long
2:26:38 so if they're trying to solve a problem
2:26:41 and you go to like past the average
2:26:45 they get penalized
2:26:47 so they [ __ ] hang up on you
2:26:54 multiple times
2:26:58 and and how come it is that I have to
2:27:00 explain my story every single time I
2:27:03 talk to a new person
2:27:05 like is there not a system that they're
2:27:07 typing in the [ __ ] notes like they
2:27:09 tell me they are I'm saying are you
2:27:11 writing this down so that I don't have
2:27:12 to explain this again to the next person
2:27:14 because I know you're not going to be
2:27:15 able to solve this for me oh I'm writing
2:27:17 everything down I get the next person
2:27:18 hello Mr Shannon boy we really value you
2:27:21 today could you start at the beginning
2:27:23 and tell me where you are or no could
2:27:26 you please tell me your Social Security
2:27:27 number wait like I typed that in to like
2:27:31 call you yeah but I still need it's a
2:27:33 different system okay here's my Social
2:27:36 Security number so can you help me
2:27:38 what's the problem are you [ __ ]
2:27:41 serious
2:27:44 that needs to be flushed at that entire
2:27:47 system needs to be flushed down the
2:27:48 toilet tomorrow
2:27:51 but it's a Saturday that's past our
2:27:53 working hours exactly
2:27:56 bitter you betcha
2:27:58 all right CGI crowd people been around I
2:28:01 know they've been around forever but you
2:28:02 know this is now but now we're starting
2:28:05 to get into
2:28:07 the CGI people have been around in big
2:28:10 Stadium scenes but when you've got
2:28:12 Leonardo here and you've got the actors
2:28:14 behind them doing whatever they're doing
2:28:16 acting like they're talking and they're
2:28:18 close enough that you can see them
2:28:19 talking
2:28:20 that's the stuff that's now going to be
2:28:21 replaced where it's just going to be Leo
2:28:24 sitting in a green screen
2:28:27 and all that stuff's going to be
2:28:29 synthesized now and so sag The Screen
2:28:32 Actors Guild is trying to protect its
2:28:34 its uh things no more sag dues yeah no
2:28:37 more sag paychecks either I still get
2:28:39 um paychecks I did a small teeny scene
2:28:42 in New Jack City and I still like every
2:28:44 six months or so get a uh
2:28:47 a sag check for like eleven dollars I
2:28:51 was like I did that and
2:28:53 like the late 80s or early 90s I'm still
2:28:56 getting [ __ ] checks for it but like
2:28:58 that's the kind of [ __ ] where what
2:28:59 they're saying is if you're a background
2:29:01 actor I had lines so it was a different
2:29:03 thing but if you're a background actor
2:29:05 without lines they're saying we just
2:29:06 want to synthesize you and we'll never
2:29:08 pay you again
2:29:11 it's nuts creative is being hit hard I
2:29:14 know it is it really is it's it's
2:29:17 sad and it's it
2:29:20 I've got such mixed feelings about it
2:29:23 because the creative people I know that
2:29:24 are embracing this stuff are finding
2:29:27 ways to reinvent themselves and they're
2:29:29 finding new jobs and more lucrative jobs
2:29:32 but the ones that are kind of sitting
2:29:33 back and waiting
2:29:35 are just [ __ ] terrified and and
2:29:38 everything's gonna change for them I
2:29:40 feel [ __ ] horrible about it but it is
2:29:43 what it is
2:29:44 um today there was a tweet about an
2:29:46 actor selling their likeness to the MCU
2:29:48 as a background character yeah that's
2:29:51 interesting I mean you know it's funny
2:29:52 like crossing the picket line used to
2:29:54 mean crossing the picket line
2:29:57 crossing the picket line with this AI
2:29:59 stuff could mean
2:30:02 I record 30 seconds of audio and take
2:30:06 you know
2:30:07 37 pictures of my of my head and send
2:30:11 that to someone and sign a form
2:30:13 and say you can use my likeness in
2:30:15 perpetuity
2:30:16 and so
2:30:18 I mean acting's a [ __ ] up business
2:30:20 because for every job there are
2:30:23 literally
2:30:24 a thousand people that would murder you
2:30:27 to have that job would do it for free
2:30:29 would pay to be in a scene with
2:30:34 anyone on a film set
2:30:37 so it's a it's a it's a real tough
2:30:40 business for for that reason that it's
2:30:44 that because everyone wants to be in it
2:30:48 like media companies are notoriously
2:30:50 shitty because everybody wants to be a
2:30:53 rock star everyone everybody wants to be
2:30:55 an actor everybody wants to be famous
2:30:58 and so
2:31:00 they don't have a lot of Leverage to say
2:31:02 you know the big stars do and that's why
2:31:04 that's why you see the big stars really
2:31:05 stepping up and saying this is our
2:31:07 platform like Leo won't step a foot on
2:31:11 set that starts to cost the Studio's
2:31:13 money
2:31:14 but but once you get past a a and a half
2:31:18 level actors
2:31:20 there's a thousand people for every
2:31:22 actor for every role that that would
2:31:25 literally kill for it
2:31:27 crazy business
2:31:28 James Dean original femboy it's true
2:31:31 holy I had no idea you were
2:31:35 riffing full scenarios on these lives oh
2:31:39 I know the uh yeah the the uh personas
2:31:42 yeah it's crazy
2:31:47 um take care of humans certified real
2:31:49 human news life yeah exactly I I I I
2:31:54 could see there being
2:31:56 um symbols for this is synthesized or
2:31:58 this is real I could see that being a
2:32:00 thing ah look at the tick tock hat
2:32:03 as they say on Elder attack honey I'm on
2:32:06 the tickety talk
2:32:09 great job you've been live for 90
2:32:11 minutes I think I've been live for like
2:32:12 three hours
2:32:13 all right
2:32:15 um I gotta go soon my voice is getting I
2:32:17 can I can always tell asses numb voice
2:32:19 is starting to go what on who on Earth
2:32:22 wants to live in a town okay I don't
2:32:24 know what that means xai coin oh did
2:32:27 they announce a coin xrp Queen
2:32:31 and also xrp Queen
2:32:33 um what you call it
2:32:37 um world coin Sam Altman's World coin if
2:32:40 you're if you're into the nft the crypto
2:32:42 scene the world coin so if you don't
2:32:44 know about world coin I think this is
2:32:46 fascinating so Sam Altman's side gig is
2:32:49 this thing called and Sam Altman's the
2:32:51 CEO of openai
2:32:53 um of Chad gbt
2:32:55 um world coin is a cryptocurrency that's
2:32:58 tied to
2:33:00 um human identity so they send you this
2:33:03 orb it's like this silver kind of
2:33:05 softball size thing that you hold up to
2:33:07 your eye or baseball sized and it scans
2:33:10 your retina and determines that you're
2:33:12 human and if you're human it goes Check
2:33:14 human
2:33:16 it establishes an identity for you on
2:33:19 the black blockchain as verified human
2:33:21 and they give you world coin
2:33:23 cryptocurrency and so it's it's about
2:33:26 the cryptocurrency as an incentive
2:33:28 mechanism but what it's really about is
2:33:30 creating a network of verified humans so
2:33:34 imagine
2:33:35 a Twitter like
2:33:38 service or in anything like service
2:33:40 where you knew that the the person you
2:33:43 were interacting with was actually that
2:33:45 person it wasn't a bot you could do that
2:33:48 with a system like that so whether it's
2:33:50 Sam Altman doing that or someone else
2:33:52 but that's what worldcoin is
2:33:56 um Jasper and Dolly is cool also I agree
2:33:59 Jasper's sitting on top of chat gbt if
2:34:02 you didn't know that if any people want
2:34:04 to build what he just said with me I'm
2:34:06 down let me know yeah [ __ ] hang out
2:34:08 together and by the way Mr a-i-a-i-o 10
2:34:13 out of 10 on the name 10 out of 10 on
2:34:16 the name and yes I encourage you to
2:34:17 collaborate if if you guys find each
2:34:20 other here also come hang out at the AI
2:34:22 salon so go to the salon.ai
2:34:26 there's a what's it called
2:34:29 link tree there it's getting late their
2:34:32 brain's not really working go to the
2:34:35 link tree there and uh there's a meet up
2:34:37 we meet every other week and we have a
2:34:39 Discord where you can geek out and you
2:34:41 can play with
2:34:42 uh mid-journey and you can ask questions
2:34:44 and you can collaborate and stuff like
2:34:47 that so that might be a place the the
2:34:49 Discord for the salon might be a place
2:34:51 to hang out
2:34:53 ask chatgpt for name ideas yep I'll
2:34:55 definitely do that real people are
2:34:57 becoming all no names that's interesting
2:35:00 shirt must say on the black say wait
2:35:03 trip let's say on the back are you from
2:35:05 Minnesota no I dated someone from
2:35:07 Minnesota once I'm from I grew up in
2:35:09 Pennsylvania and then I moved to New
2:35:11 York City out of school
2:35:13 and
2:35:15 lived there for a long long time and
2:35:17 then I've been in Denver for 14 [ __ ]
2:35:19 years I miss New York but I'm here
2:35:23 what's the easiest way to start
2:35:24 incorporating it into my work start
2:35:26 small start specific
2:35:28 um go
2:35:31 go there
2:35:34 and
2:35:35 first just play with it
2:35:38 then go to prompts.chat and learn how
2:35:41 you can
2:35:43 teach it to act like a Persona so have
2:35:46 it act like a project manager or have it
2:35:48 act like a copywriter I don't know what
2:35:50 your job is
2:35:51 and then when you start to get your head
2:35:53 around how flexible it is and and how
2:35:56 how good it is at lots of different
2:35:58 things
2:35:59 take some small thing that you do at
2:36:01 work and see if you can make it
2:36:04 better faster or or cheaper
2:36:08 for lack of a better term then you can
2:36:11 do it right now so and and here could be
2:36:13 a simple term
2:36:16 every day you have a call with a client
2:36:18 and while you're on that call you're
2:36:20 taking notes
2:36:21 and at the end of the call you spend 15
2:36:25 minutes
2:36:26 writing up those notes into
2:36:30 you know Trello and into an email back
2:36:35 to the client
2:36:40 record
2:36:41 that you know record your call the next
2:36:44 call take that transcript
2:36:47 have ai automatically
2:36:50 summarize that transcript give you key
2:36:52 takeaways
2:36:54 and
2:36:55 um write the email draft to the client
2:37:00 get it to get it to summarize meeting
2:37:02 notes and write a draft email just start
2:37:05 with something really small some little
2:37:07 thing like start with pieces of Your
2:37:10 World
2:37:12 that are just mind-numbing time wasting
2:37:16 [ __ ] that has to get done
2:37:20 and like summarizing meeting notes is
2:37:22 one of those it [ __ ] melts my face I
2:37:25 hate it I hate dealing with it
2:37:27 make a note summarized
2:37:29 now you can go to futurepedia and there
2:37:32 are all sorts of plugins that'll
2:37:33 automatically do that for you but learn
2:37:35 how to do it yourself and it will pay
2:37:38 you dividends because once you do it
2:37:39 once and kind of get your head around it
2:37:41 then you'll go oh I know what I could do
2:37:43 now I could write a project plan or I
2:37:45 could do this or I could
2:37:48 we I created a a thing that was a core
2:37:51 part of our creative process at
2:37:52 storyline it saved us on average
2:37:56 two additional creative meetings because
2:37:58 we're now automating a huge chunk of it
2:38:03 saving our clients money and US money
2:38:07 YouTube
2:38:09 all right listen it's getting late for
2:38:12 those of you on the East Coast what are
2:38:14 you doing up go to bed and then I gotta
2:38:17 say that to myself go to bed all right
2:38:20 uh it's 11 15 here in Denver I want to
2:38:23 figure out how to automate social media
2:38:25 not so much ideation of content but the
2:38:28 automation of posting across multiple
2:38:30 platforms again zapier is good for that
2:38:33 kind of stuff
2:38:36 um you added two other sites to your
2:38:38 list list I have yet to go to the first
2:38:40 one okay user two four three eight one
2:38:44 nine seven five two one eight seven six
2:38:46 for one thing zero out of ten on the
2:38:49 name
2:38:50 change your name to something
2:38:51 interesting
2:38:52 get yourself a little Avatar picture and
2:38:55 then what are you doing in here if
2:38:57 you're not going to these go to number
2:38:59 one go to chat.openai.com go there right
2:39:01 now go there right now go there right
2:39:04 now
2:39:05 type in
2:39:07 um
2:39:08 write me a poem about
2:39:10 daisies
2:39:12 there's your prompt write me a poem
2:39:14 about daisies go there now write me a
2:39:17 poem about daisies do it now
2:39:19 or you can't come back
2:39:21 when you come back have a better name
2:39:25 not putting up with this [ __ ] anymore I
2:39:27 used to ask politely I'm done I'm done
2:39:29 asking politely
2:39:31 I'm I'm not Hawking all these companies
2:39:34 for no money for nothing
2:39:36 the The Exchange is I'm not getting paid
2:39:38 for this so these changes you go get
2:39:41 make yourself better by going and
2:39:43 playing with one of these Bart is
2:39:44 terrible it's better than it used to be
2:39:47 stability AI watch your take on their
2:39:49 open source approach I think stability a
2:39:51 I think stability is open source
2:39:54 approach the cynical part of me has
2:39:56 always thought
2:39:59 that they're open sourcing it because
2:40:00 deep down they know they're going to get
2:40:02 their asses sued off the planet
2:40:05 and by
2:40:08 um by open sourcing it they ensure their
2:40:11 longevity and they might be able to
2:40:13 there might be ways for them to kind of
2:40:15 squeeze through the legal system and
2:40:17 reconstitute as a different company
2:40:18 using the technology they built with the
2:40:22 staff that they built it with that's the
2:40:24 cynical side of me
2:40:26 um listen I think it's great it's like
2:40:27 the developer Community for stable
2:40:29 diffusion is [ __ ] insane
2:40:32 it's crazy powerful what what the
2:40:35 developer Community is doing there so I
2:40:38 think it's absolutely awesome you know
2:40:40 I'm always trying to figure out what's
2:40:42 their angle
2:40:44 and
2:40:46 here's the thing with with stable
2:40:48 diffusion with open AI with mid-journey
2:40:53 I think if if they had tried to secure
2:40:57 copyright and trademark
2:40:59 permissions on all the [ __ ] that they
2:41:02 wanted to train on
2:41:04 before they trained on it
2:41:06 with this ship would literally never
2:41:08 launch right so I think the only way to
2:41:12 do what they did was to just [ __ ] do
2:41:15 it knowing that they're going to get
2:41:17 sued
2:41:18 and so if you listen to the CEO of
2:41:20 mid-journey he's like yeah man you know
2:41:22 it's [ __ ] information like whatever
2:41:24 we trained on it like you know whatever
2:41:26 like you can Suez man but it's like
2:41:28 whatever like he's just like he doesn't
2:41:29 give a [ __ ] so I like I have a feeling
2:41:31 they're gonna get sued out of Oblivion
2:41:33 but they've got the best tool
2:41:35 and then and then I think stability is
2:41:38 probably like hey we've got to train on
2:41:39 all this data if this is going to be any
2:41:41 good we've got to train on the good data
2:41:43 but we're gonna get the pants suit off
2:41:46 us
2:41:47 so why don't we just open source this
2:41:49 [ __ ] as we develop the models we'll just
2:41:50 throw them out there and then at least
2:41:52 the work
2:41:53 sees the light of day
2:41:55 that's my thought on it I don't know
2:41:57 it's that's total speculation I love
2:42:00 specular I loves me some speculating
2:42:03 recommend today's episode of hard fork
2:42:06 okay I will go check that out you need
2:42:08 GPU with Windows I have no GPU Mac if
2:42:12 you have an M1 Mac all the open source
2:42:14 things are the developers like Max the
2:42:17 the they write for the m1s and the m2s
2:42:20 are are supposedly baller with this uh
2:42:23 with this stuff but yeah if you want a
2:42:25 GPU you're going to need to get yourself
2:42:27 one of them near paces
2:42:29 and run that Microsoft or run one of
2:42:32 them Unix
2:42:33 Unix systems you know the ones where you
2:42:36 gotta type a lot I don't like to type a
2:42:38 lot I'm a point and Clicker I'm a
2:42:41 point-and-clicker that's why I like the
2:42:42 Mac because you it's it's a lot more
2:42:43 point and click I don't have to do as
2:42:45 much of this I do more of this you see
2:42:48 if you're this kind of person then you
2:42:50 like the Mac if you're this kind of
2:42:51 person then you're good on a PC
2:42:55 you understand what I'm saying
2:42:57 does that make sense
2:43:04 I don't know who these people are
2:43:06 they're all Disturbed
2:43:09 I'm not Disturbed they're Disturbed that
2:43:13 don't put it on me
2:43:16 all right
2:43:17 how about my own
2:43:19 non-ai-generated name chatty Susie
2:43:23 that's not bad I like it it's an 8 out
2:43:26 of ten are you an AI bot I'm gpt5 yes
2:43:29 nice to meet you
2:43:31 all right I told you I have to go I have
2:43:33 to go
2:43:34 why are you making me answer your
2:43:36 questions stop with the hearts
2:43:39 you're making me not want to leave I'm
2:43:41 already codependent
2:43:44 I'm a shame-filled being
2:43:47 all I want to do is please
2:43:51 all right I'll do a few more questions
2:43:53 ugh Claude isn't available in Germany
2:43:56 yet
2:43:57 I need to ask chat GPT about my VPN you
2:44:00 do Road Warrior Princess it's VPN time
2:44:03 with some of these tools
2:44:06 I'm curious if claude's there if is
2:44:08 Claude in Germany right now the the uh I
2:44:11 heard that uh is it Claude
2:44:17 Bard I think Bard's not available in
2:44:19 Canada now
2:44:23 it's getting weird like like all the now
2:44:25 that this stuff now that there's sort of
2:44:27 this AI Panic starting in where people
2:44:29 are like holy [ __ ] this is going to
2:44:30 change everything I think people are
2:44:32 sort of clamoring onto their old
2:44:34 business models and everyone's getting a
2:44:36 little
2:44:38 it's gonna it's gonna get a little
2:44:39 interesting for the next year or two
2:44:41 um where where the the the the
2:44:44 traditional centers of Power are going
2:44:47 to start panicking and trying to lock
2:44:48 this [ __ ] down and it's going to keep
2:44:50 squeezing through their fingers
2:44:53 and we get to we get to play with all
2:44:54 the [ __ ] that squeezes through their
2:44:56 fingers so
2:44:57 we're gonna have more power than
2:44:59 governments for for some time it's fun
2:45:05 because governments have their heads
2:45:06 firmly planted in their behinds
2:45:10 circle with circles all right let's see
2:45:13 can I create a babe from Weird Science
2:45:15 on AI you can just go to Mid Journey or
2:45:17 go to just go to bing.com
2:45:20 it's not very good but you can say make
2:45:22 me an image of babe from Weird Science
2:45:25 and it'll do it
2:45:27 and if it sucks go to Mid journey and
2:45:29 mid journey is always better it's just
2:45:30 mid-journey's the [ __ ] bomb
2:45:33 it's amazing thoughts on data crypto
2:45:36 tokens
2:45:38 um
2:45:39 that's actually a good question so I
2:45:41 think that I think that
2:45:44 crypto tokens and blockchain
2:45:45 tokenization in general is going to
2:45:47 become really really important
2:45:50 because of AI
2:45:54 how why
2:45:56 isn't it important now Kyle didn't you
2:45:59 hear about xrp yes I've heard about all
2:46:01 that
2:46:02 it's not important right now
2:46:04 it's going to be important because of
2:46:06 the 2024 election in the United States
2:46:08 of Grand Old
2:46:11 [ __ ] off America
2:46:14 um
2:46:16 the the amount of misinformation and
2:46:21 synthesized
2:46:22 video and photographic evidence of Mal
2:46:27 doings
2:46:28 in the 2024 election I think is going to
2:46:31 be way over the top I think it's gonna
2:46:34 be I think it's gonna make you know 2016
2:46:37 and 2020 look Queen
2:46:40 um and then I think after that
2:46:43 we're gonna learn lessons like hey this
2:46:45 shit's actually dangerous
2:46:48 and we gotta come up with a way fast to
2:46:52 be able to authenticate human beings
2:46:54 which that's what Sam Altman's doing
2:46:56 with World coin
2:46:57 and we need to authenticate the
2:47:01 authenticity of
2:47:04 um media
2:47:05 is this article the actual New York
2:47:08 Times article in an unaltered state
2:47:10 check you know green blockchain check
2:47:13 mark yes written by this human being
2:47:16 verified reporter Check Yes here's the
2:47:19 chain of right so all the blockchain
2:47:21 verification stuff for authenticity I
2:47:24 think it actually becomes quite critical
2:47:26 now it doesn't have to be blockchain
2:47:27 technology it just needs to be some sort
2:47:29 of tokenization Technology
2:47:31 um
2:47:32 but I think I think there's going to be
2:47:34 demands for it like I think there will
2:47:36 be societal demands and I think there's
2:47:37 going to be governments are going to be
2:47:39 like we have to lock this [ __ ] down
2:47:42 so
2:47:43 there's a business idea
2:47:47 because if one of us doesn't do it the
2:47:49 government's going to do it we're going
2:47:50 to be pissed off at that well has the
2:47:52 government got the keys to my
2:47:54 identity
2:47:56 why they got to tokenization well you
2:47:59 should why don't we just use one of
2:48:00 these crypto coins
2:48:04 because the crypto Bros were out trying
2:48:05 to you know rug their followers
2:48:09 and make more crypto billions rather
2:48:11 than solve this problem now I'm sure
2:48:14 there are some out there trying to solve
2:48:15 this problem
2:48:16 Sam Altman's trying to solve the human
2:48:18 version of it
2:48:19 and I would imagine that there's a
2:48:21 Content authenticity component of that
2:48:25 there's probably projects out there that
2:48:27 are doing this but but there's not
2:48:29 there's not an appetite for it right now
2:48:32 because we haven't had like an
2:48:35 Armageddon level event with
2:48:36 misinformation I mean you could argue
2:48:39 that 2016 election was that but
2:48:42 you know again I think 2024 is going to
2:48:45 make it look Queen
2:48:46 so all right I gotta go
2:48:50 I know I know I gotta go what do we
2:48:54 learn here nothing I'm completely
2:48:55 unqualified
2:48:57 all right I'll let you take that as as
2:48:59 the final word of the night I'm
2:49:01 unqualified I'm over caffeinated my name
2:49:04 is Kyle Shannon this is the AI learning
2:49:06 lab I appreciate all your time peace
2:49:08 I'll see you tomorrow