10/8/25 - Fulfilling Steve Jobs' 1985 Dream of an Interactive, AI-Powered Aristotle

Description

Chapters

Transcript

0:00 that before. Like
0:03 we could have done that earlier and
0:05 given you a little more runway, but All
0:06 right. Well, we'll do that. We should we
0:08 should actually probably do that for all
0:11 the big events.
0:13 Um All right, Tam. Don't you think it's
0:16 time? Huh?
0:20 You ready? You ready to come singing?
0:24 [Music]
0:35 Hello.
0:37 [Music]
0:40 Hello.
0:43 [Music]
1:09 [Music]
1:14 Every time I see you now, get that look
1:17 in mind.
1:20 Every time I see your man thought, hear
1:23 that smile.
1:26 The early misty morning light that I
1:29 heard the engine turning the old
1:33 outside.
1:37 [Music]
1:40 You leading me
1:43 again today.
1:46 You will convince me
1:48 a good day.
1:52 Leaving this hotel looking for someone
1:55 else's golden ring.
2:00 to say
2:02 so long Suz.
2:04 [Music]
2:18 [Music]
2:20 >> Don't you cry for me.
2:29 [Music]
2:41 [Applause]
2:41 [Music]
2:44 [Applause]
2:48 Hello.
2:50 [Music]
2:52 Hello.
2:54 [Music]
3:06 There's been something baby I've been
3:08 trying to say
3:12 for an age and it seems I don't know how
3:16 [Music]
3:18 the past and the future now surrounding
3:20 me.
3:24 surrendered to a heaven and she thrill
3:25 can't be found.
3:29 There's been a little trouble
3:32 since you came to my rescue.
3:39 And if you like all of the rest, I would
3:42 have quit you long ago, but I couldn't
3:45 do that.
3:46 [Music]
3:49 Oh, tell me now when I never went to
3:54 well,
3:57 make a man crazy, make him cold as hell.
4:03 I'm a woman that you wish me well.
4:07 But in spite of you trying, still going
4:10 to have to find my own way through.
4:14 [Music]
4:26 Oh, good people, good people, good
4:29 people of the Tik Tok, good people of
4:32 the internet.
4:34 I hope you're doing well today.
4:37 I had a shitty day.
4:42 Just gota Sometimes it gets intense,
4:44 man. Sometimes you're out in the field,
4:47 it just gets intense. You know what I'm
4:49 saying?
4:52 That was my day. And it wasn't there was
4:55 not nothing particularly challenging
4:56 about it. It was just one of those days
4:58 where a lot of people are feeling it.
5:02 So,
5:05 I don't know if you've had one of those
5:06 days where you're just feeling it.
5:12 And feeling it ain't bad.
5:15 But it's feeling it.
5:19 Some days you're the fly, some days
5:21 you're the windshield. Exactly. This was
5:23 a fly day.
5:28 Sorry to hear that. Oh, it's that's
5:29 that's life in a big city source camp,
5:32 but I appreciate it. That's uh you know,
5:35 I mean, listen, you're you're an
5:37 entrepreneur. You know what it's like.
5:40 you know, you you uh if you've never
5:44 started a company, good on you. If you
5:48 have started a company, you know, it's
5:50 just like, you know, you know what's
5:52 funny about starting a company
5:55 is
5:57 I've started I don't know 15 20
6:00 companies in my day.
6:03 Some went big, some didn't deserve to
6:06 make it past a week and didn't
6:10 and you know, lots in between. Um,
6:18 and probably every day, every day
6:23 there's some problem where you're like,
6:25 "Oh, someone should take care of that."
6:27 And then you realize
6:30 you're it
6:32 like there's no one there to take care
6:34 of it. Like you're it. And that's that's
6:37 what it is to run a business is just you
6:40 know the the the reason you start a
6:43 business in the first place is because
6:46 you're dissatisfied with something in
6:48 the world, right? There's got to be a
6:51 better way, right?
6:54 I'm going to make people happier with my
6:56 t-shirts than all the other shitty
6:57 t-shirts out there. You've discovered
6:59 there's some problem with the world. And
7:02 then you start a company and then uh
7:05 it's just a series of grievances from
7:07 you and trying to fix them.
7:11 It's perfect for Gen Xers.
7:16 Oh yeah, I know those days. Yeah,
7:18 exactly. Yeah, many like that.
7:21 Um,
7:24 I wanted to
7:26 we we'll do we'll do a
7:29 wax philosophical Wednesday. I wanted to
7:32 do a little
7:35 We're actually not gonna I well I don't
7:37 think I'll talk too much about it,
7:40 but I just stumbled on a video that I've
7:42 seen I don't know 20 times 50 times
7:48 and
7:53 it put me in touch with
7:57 like you know I was talking earlier
7:58 about feeling it like feeling what's
8:00 going on in my business feeling what's
8:01 going on and the AI salon, which I'm
8:04 trying to sort of take to the next
8:05 level,
8:07 feeling what's just going on in my life.
8:10 Um,
8:14 and so I saw this video
8:17 and it put me in touch with
8:20 what remarkable times we're in.
8:24 We're we're in really remarkable times.
8:27 And,
8:29 you know, it's funny. I saw I saw an
8:31 interview with Sam Alman this week
8:34 and he's like he's like you know for 50
8:38 years for 60 years 60 or 70 years
8:42 we had this thing called the Turing test
8:46 which was basically if you sit someone
8:49 down at a computer screen and they can
8:50 interact with it and they can't tell if
8:52 it's a human on the other side or a
8:55 machine then we've passed the turning
8:57 test. It's like we passed that and just
9:00 blew past it and then you know we've now
9:02 got machines that are smarter than us
9:04 and we were like oh yeah we got that now
9:06 and we just blew past it and now we're
9:08 like why can't it be faster if we
9:10 completely normalize it? Tik Tok
9:12 question. Are you concerned about
9:14 misinformation with AI? Um sure. Yeah.
9:18 But I mean, we haven't needed AI to have
9:22 like the last decade of of at least
9:25 America, but really globally, the last
9:27 decade
9:29 has been an a misinformation
9:32 hyper war. Um, you know, what's funny
9:36 about AI misinformation is I think that
9:41 this is actually kind of fascinating.
9:44 I had I hadn't really thought about it
9:45 like this, but but there's so there's
9:49 there's there's this thing I don't know
9:51 if you've heard of it, but you probably
9:52 have called AI slop.
9:55 And basically what AI slop is is the
9:58 term that the creative community gives
10:00 to anyone that creates anything with AI,
10:03 right? But that's not actually that it's
10:06 not actually right. AI slop
10:09 is when lazy people use AI without
10:14 thinking and they just generate slop.
10:17 Now, the the naysayers of AI say that
10:20 all AI is that. It's not. It's just not.
10:24 Um,
10:28 but what I've noticed is that the people
10:30 that I know that are that are really
10:34 doing great work with AI and thinking
10:36 critically about AI and all that sort of
10:38 stuff, a lot of the members of the AI
10:39 salon,
10:41 when we see
10:44 crap out of chat GPT, you can you you
10:46 can just sniff it. You know what it you
10:48 know you know what it is, right? And
10:50 when you see an image that was just
10:52 lazily prompted, for the most part, you
10:54 know what it is. You know it's AI and
10:57 you know it's it's lazy AI. It's slop.
11:00 And so I think that there's there's a
11:02 weird
11:05 there's a weird possibility
11:08 that AI is actually going to be the
11:10 antidote to misinformation because
11:12 there's going to be so much of it but it
11:16 will all have a sameness to it that
11:19 authentic messaging will rise above the
11:22 noise. That's it. It is a possibility.
11:24 So, so a possibility is we just get
11:27 better at sniffing out [ __ ]
11:29 and then and then the authentic stuff
11:31 rises above the noise. But
11:33 misinformation that's weaponized is more
11:36 sophisticated than that. And then I
11:38 think what happens is this. Then I think
11:40 AI helps us build tools
11:44 that allow us to instantly be able to
11:48 assess if something is misinformation or
11:51 not. Um the the weird thing about
11:55 misinformation is we're we're in such a
11:58 polarized charged, you know, my analogy
12:01 for the United States right now is like
12:04 we're we're two football teams and it's
12:07 like you can't tell someone from the
12:09 Pittsburgh Steelers that they should
12:11 love the Dallas Cowboys. Like it it
12:14 doesn't matter what argument you make
12:17 like Dallas sucks or like Pittsburgh
12:19 sucks. That's where we are as a country
12:22 right now, right? So
12:25 in that environment,
12:27 misinformation's really easy because you
12:30 just make information look like
12:31 Pittsburgh Steelers friendly stuff and
12:33 they're like, "Oh, that's good." And
12:35 then you make it look like Dallas Cowboy
12:37 bad. Oh, they're bad. There's there's
12:40 literally no critical thinking going on.
12:42 So maybe there's a possibility that AI
12:45 actually helps in the
12:48 analysis and sifting of shitty
12:51 information to actually say here's
12:54 actual neutral valid information.
12:58 I don't know that there's an appetite
13:00 for that. I think what the social media
13:04 companies have mastered over the years
13:07 is monetizing our polarization.
13:12 So, you know what I would say to you is
13:15 go vibe code up a social media network
13:17 that that just pedals in information. No
13:21 information, no misinformation allowed.
13:23 And see if anyone shows up. They
13:25 probably won't.
13:29 There's never been any shortage of
13:30 misinformation. Pretty sure
13:33 some of those Samanskrit clay trap
13:36 tablets were fake news. That's really
13:37 funny. Um, by the way, I'm from Dallas
13:40 and the Cowboys are that bad. Well,
13:42 that's that's that's a whole separate
13:44 thing. I was using hypothetical sports
13:46 teams, but yes, congratulations on on
13:49 living in in where America's team is.
13:58 Oh, man.
14:02 All right.
14:06 So, here's what I want to do.
14:09 The the Browns can't even win an AI.
14:15 Um,
14:17 we just celebrated celebrated. We just
14:20 marked the 14th anniversary of of Steve
14:23 Jobs passing.
14:25 Um, America can have them.
14:29 Uh,
14:31 commemorated we commemorated
14:34 the passing of Steve Jobs. And it looks
14:36 like Tim Cook is stepping down as CEO,
14:38 which I think it's probably about four
14:40 years too late, but that's neither here
14:42 nor there. I think he was a good CEO. He
14:44 he stabilized Apple in Jobs's absence,
14:48 but they they they need uh they needed
14:52 they need a visionary, and I think
14:53 they're bringing on a CEO who's not a
14:55 visionary, but whatever. They didn't ask
14:57 me if I wanted to run it, so so my
15:01 opinion doesn't [ __ ] matter.
15:03 um
15:10 as as flawed an individual as as he was.
15:15 Um
15:18 he was a he was a hero of mine. um and
15:20 in a lot of ways is still a hero of mine
15:24 because he could
15:28 he could see the future and and and not
15:31 in like a
15:34 like a psychic kind of way.
15:37 He just observed the world and he looked
15:39 at
15:41 he looked at what happened when you
15:43 stuck a bunch of transistors on a
15:45 circuit board and programmed them in a
15:48 certain way. And he could he could
15:50 connect the dots forward
15:53 about, you know, okay, this is a piece
15:56 of crap today, but let me let me project
15:59 forward and what what would this thing
16:01 be like if it weren't a piece of crap?
16:06 And then he makes the Apple 2 and and
16:08 changes education and and then he goes
16:12 to Xerox Park and they show him the the
16:14 first computer mouse and he says, "What
16:17 are you doing with that? What are you
16:19 doing with this operating system?" Well,
16:21 we're not doing anything with it. And I
16:22 think he bought it for 25 grand or 50
16:25 grand and he made the Macintosh
16:28 and he, you know, single-handedly
16:30 changed things and just kept doing that
16:32 again and again and again and again. And
16:33 it wasn't
16:35 it wasn't magic. It was it was
16:38 observation and it was you know touching
16:40 into tapping into his humanity.
16:46 And so this video it's a fairly famous
16:48 video
16:50 but this is the video where he he
16:52 predicts he predicts AI. He he he talks
16:56 about AI here.
16:58 And I think this is when was this? This
17:01 was this is an old video.
17:08 1985
17:10 I was in my second year of college at
17:12 Penn State.
17:13 >> The great
17:16 >> 1985 he said this. We're in 2025. How
17:20 many years is that? 15 plus 25 40 years
17:24 ago.
17:26 40 years ago.
17:31 Brandon, producer Brandon was two.
17:34 >> You know who Alexander the Great's tutor
17:35 was for about 14 years?
17:38 You know, Aristotle. And I read this, I
17:41 became immensely jealous.
17:44 Uh, and
17:46 I think I would have enjoyed that a
17:48 great deal.
17:50 and and uh through the miracle of the
17:53 printed page I can at least read what
17:56 Aristotle wrote without an intermediary
18:00 and maybe if there's a professor they
18:02 can they can add to that but at least I
18:04 can go directly to the source material
18:06 and that is of course the foundation
18:08 upon which our western civilization is
18:09 built but I can't ask Aristotle a
18:13 question I mean I can but I won't get an
18:15 answer and so I
18:19 My hope is that
18:22 in in in our lifetimes
18:26 we can make a tool
18:29 of a new kind of an interactive kind.
18:33 And so my hope is someday when the next
18:36 Aristotle is alive, we can capture the
18:40 underlying world view
18:43 of that Aristotle in a computer. And
18:47 someday some student will be able to not
18:50 only read the words Aristotle wrote, but
18:52 ask Aristotle a question and get an
18:55 answer.
18:57 And uh
18:59 that's that's what I hope that
19:01 >> that's what I hope that we can do. 1985
19:06 he somehow turned in that speech he
19:08 somehow turned the printed page
19:12 into a radical technology.
19:18 I was jealous that I didn't get to have
19:20 Aristotle teach me.
19:22 But through the miracle of the printed
19:24 page, right? He's he's going back
19:27 hundreds of years. Through the miracle
19:29 of the printed page,
19:32 I can go directly to the source. And
19:35 wouldn't it be nice if I could ask him a
19:36 question?
19:39 We've been able to build this
19:42 what he just described for at least two
19:45 years now.
19:47 And so
19:50 I thought what might be fun tonight
19:53 is to go build this tool that he
19:56 described.
19:58 And I'm going to do it in two ways.
20:03 I might do it in three ways.
20:06 Um
20:10 we're going to go to Notebook LM.
20:14 And if you want to play along, feel free
20:16 to play along. Um,
20:19 I'm going to create a new notebook. If
20:21 you don't know, if you don't know
20:23 notebook LM
20:26 um, go to notebook LM as in language
20:31 model.google.com
20:34 and make an account if you don't have
20:36 one.
20:38 And with this tool, you can upload up to
20:41 50 documents and they can be anything.
20:43 They can be videos. They They can be
20:45 YouTube links. They can be websites.
20:48 They can be Google Docs.
20:51 They can be PDFs. And so what we're
20:53 going to do is before the call tonight,
20:57 what is this a call? This is a live
20:58 stream. Before the live stream tonight,
21:00 I went and I got a PDF
21:05 of
21:07 Is it? No.
21:10 Where did I get it? Um
21:16 let's see
21:19 the complete. Okay. So go here go
21:24 here we go the dash complete
21:32 the complete Aristotle
21:34 tab. I'm sharing the right tab. Oh I'm
21:36 just not I'm I'm just not here. Okay. So
21:40 you can go right now on the internet and
21:43 you can search for the complete
21:44 Aristotle and it's like a 7,000page
21:47 document. The guy wrote a lot of words.
21:51 What AI we're using right now is
21:52 notebook LM. And then we're going to use
21:54 chat GBT.
21:56 Uh and then we might even go vibe code
21:58 something. I might go to Claude and vibe
22:00 code something.
22:04 Thank you for this demo. It's a powerful
22:06 learning tool. Yeah. So we're gonna do a
22:07 couple of things. So right now um the
22:11 little circle is still spinning. So So
22:13 it's it's reading notebook LM right now
22:16 is reading
22:18 Aristotle's complete works.
22:23 The provided excerpts from Aristotle's
22:25 work primarily focusing on categories
22:28 and analytics meticulously outline the
22:30 fundamental concepts of logic,
22:31 metaphysics, and physics. The text
22:34 explores the nature of substance and how
22:36 different things. Furthermore, the
22:38 passages introduce and define
22:41 correlative terms arguing for their
22:43 interdependence and simultaneous
22:45 existence. I don't know what the [ __ ]
22:47 they're talking about. I can't even
22:48 understand it. But you know what I'm
22:49 going to do? I'm going to go over here
22:51 to this thing called audio overview and
22:53 I'm going to push the button. And so
22:55 while that thing's generating an audio
22:57 overview, we're going to go build
22:58 something else.
23:00 Tell me when AI can create any app
23:03 better than a human. Oh, Robbie's an AI
23:06 hater. I understand that. Um, I
23:09 understand the fear. Um,
23:13 it depends on which human you're talking
23:14 about, Robbie. Um, if you're talking
23:17 about me right now today, AI can build a
23:21 better app than me. But I don't think
23:22 that's what you're talking about. I
23:24 think what you're talking about is tell
23:27 me the day when AI can build an app that
23:29 doesn't require a human to fix the the
23:31 the gaps, right? And the shittiness and
23:36 the jankiness
23:38 and
23:40 none of them come close today. So So
23:44 you're good for you're good for right
23:45 now.
23:47 Um,
23:50 I would be surprised and I'd be willing
23:52 to to I'd be willing to wager on it that
23:56 within
23:58 I'd probably wager on three years, but
24:01 I'll predict within two years. Um,
24:06 you'll be able to you'll be able to
24:10 prompt an app into existence that's
24:12 better than what humans can build. I I'm
24:15 fairly confident of that. um codecs
24:17 right now. So, so let's see. Um
24:20 the new the new Sonnet 4.5,
24:24 is it 4.5? Is that what's the latest
24:27 one? That thing just coded for 30 hours
24:31 straight. They basically gave it a task
24:33 of of recreating Slack. It coded for 30
24:36 hours straight and delivered a
24:38 functional application. Um, Codeex from
24:42 OpenAI right now is coding for about a
24:44 day, a full day, um, probably in the
24:48 24-h hour period. Sam Alman thinks that
24:51 very shortly it'll be able to code for a
24:53 week straight.
24:56 Um, so I would say just in terms of
24:59 marathon,
25:01 um, it can it can write better than a
25:04 human. Um,
25:07 in terms of quality, we're not anywhere
25:09 close yet. So, so that's that. What what
25:11 I would say to to you is this, Robbie,
25:14 is
25:18 if you're using
25:20 if you're using the fact that AI's
25:26 not good enough, that that humans are
25:29 superior. If you're using it as an
25:30 excuse to stay on the sidelines of AI, I
25:33 would first of all, I wouldn't stay here
25:35 and hang out with us because all I talk
25:37 about is AI and you're probably going to
25:38 hate it. So, you might want to just go.
25:42 Um, but if you're on the sidelines with
25:45 AI, I would strongly suggest don't be.
25:47 And not because I particularly care if
25:50 you use AI or not,
25:53 but don't assume that the limits that it
25:55 has today are going to exist much
25:57 longer. And and my philosophy is I would
26:00 much rather be using the tools and
26:03 understanding the tools than
26:07 sitting on the sidelines and just
26:09 watching it happen because we're in an
26:11 acceleration right now. Like the
26:13 acceleration is accelerating and these
26:15 tools are getting better faster and and
26:19 that's going to keep that's not that's
26:21 not stopping. We're not hitting a wall.
26:23 And so, you know, my my my
26:26 mo why I go live five nights a week is
26:30 just to say [ __ ] try it on. Strap it
26:33 on like a jetpack and and see where the
26:35 edges are. And you'll you may find,
26:37 yeah, it sucks, but you know what?
26:39 You're probably good enough that you can
26:41 fill in those gaps right now. And you
26:43 know what? that makes you incredibly
26:45 valuable for all the people who are not
26:47 coders that are trying to use vibe
26:49 coding to start businesses and they
26:51 create applications that have privacy
26:53 and security holes in them, they're
26:55 going to need to hire someone like you
26:57 to come in and fix them. But if you
26:58 don't understand what the the challenges
27:00 are with vibe coding, you won't work,
27:03 right? So anyway, all right, new
27:06 question. Um,
27:09 won't the AI subscription cost go up if
27:11 they're coding 247? Well, the on average
27:16 the the per token processing price for
27:20 AI has dropped 10x
27:23 every what is it Brandon? Every nine
27:25 months it's it's at least every year.
27:28 Um, that the the
27:31 quality something like that. Yeah. that
27:34 that the the quality of interaction goes
27:36 up and the cost per token is going down
27:39 at 10x
27:41 at least every year and I think it's
27:42 shorter than a year. So, so what that
27:45 looks like is the the efficiency the
27:48 efficiency is going to go up um as well.
27:51 And so what's going to so happen the
27:54 efficiency is going to go up and then
27:57 the algorithms are going to get better.
27:58 They're getting better and better and
28:00 better at training the corebased models.
28:02 They're getting better and better and
28:03 better at reasoning and now they're put
28:06 building in this agentic behavior and
28:09 these things are just going to get more
28:10 and more and more efficient. So in
28:12 theory, yes, but I think in reality, no.
28:15 Um, and here's the other thing. Um, most
28:19 people don't have a software development
28:22 program
28:24 challenging enough to need their chat
28:26 GPT to or whatever they're using to code
28:30 for a week straight, but commercial
28:34 software developers do and big
28:36 corporations do. And so if if you're
28:40 thinking about this like a uh like an
28:43 end consumer, it might be too expensive
28:46 to let a let a machine code for a week
28:48 straight. But if you're a corporation
28:50 and you now don't need to hire an
28:52 engineer to do that, then
28:55 then it's cheap. It's very cheap.
28:59 Uh Pav, my daughter did a whole NASA
29:02 project with AI help. Um
29:05 if used right, AI is great. Yeah. Yeah.
29:08 Exactly. Here's the deal. Here's the
29:10 deal. All right. This is super
29:12 important, especially if you're new here
29:14 and you're on the skeptical side of this
29:16 thing. Here's one of the biggest
29:17 mistakes people are making right now
29:19 with AI. Let me let me go go soapboxy
29:22 here for a second.
29:28 One of the one of the one of the biggest
29:30 mistakes people are making with AI right
29:32 now is they're treating it like a
29:36 genius, right? And so in the education
29:40 system, they're like, "Oh, AI is
29:42 cheating because the student's just
29:44 going to push a button and it's going to
29:46 write a report and then they're going to
29:47 turn that in and that's cheating,
29:49 right?" Or they're like, "Oh, someone's
29:51 just going to push a button and it's
29:53 going to create an image and that's
29:54 going to steal from illustrators and
29:55 that's AI slop and that's cheating." Or,
29:57 "It's going to make this video and
29:59 that's going to ruin Hollywood and
30:00 that's cheating, right? They're treating
30:03 AI like this thing outside of us that
30:05 we're in competition with it."
30:08 That whole idea I find bizarre,
30:12 but that comes from Hollywood, right?
30:14 We've got 50 years of Hollywood telling
30:17 us that it's us against the machine.
30:21 There's a whole another way to look at
30:23 AI.
30:26 There's a whole another way to look at
30:27 AI. It's not us versus the machine. It's
30:31 us having
30:34 being lucky enough to live at a time in
30:36 history where we get to take our ideas
30:39 and put them into this machine that
30:41 reflects them back to us in this
30:43 amplified way.
30:47 It's us with the machine. It's us
30:50 amplified by the machine, not in
30:52 competition with it. We get to use it,
30:55 too. All of the the conversations about
30:58 AI is going to take the jobs. Yeah. And
31:00 then all those people that lose their
31:02 jobs have access to AI where they can go
31:05 start a company with like a virtual
31:07 10person super team of AI PhDs that are
31:11 on their side.
31:13 That's absolutely [ __ ] remarkable. No
31:15 one's talking about that. So your
31:18 daughter in doing that NASA project,
31:20 sorry, black bar
31:23 in doing that NASA project, she's like,
31:25 I got this NASA project. I want to do
31:26 the NASA project. Let me have AI help
31:28 me. Help me do the research. Help me
31:30 bring it in. Let me get excited about
31:32 it. Let me put the things together.
31:34 Right? It's an augmentation of her
31:36 ideas.
31:39 Well, isn't that going to make her more
31:40 dumb because she's not learning? No, she
31:42 is learning. She's learning in an
31:45 accelerated way because she's taking her
31:48 ideas and expanding her mind, right? If
31:52 you treat AI like a genius off to the
31:54 side, oh, make write me a NASA paper,
31:57 it'll do that.
32:00 That's got nothing to do with you.
32:03 That's just what lazy people do. There's
32:05 going to be a lot of lazy people. But
32:06 anyway,
32:08 okay, let's go back to to Aristotle and
32:11 Steve Jobs. Let's see if we got our our
32:14 audio overview. We may it may be done
32:18 because I ranted.
32:21 Um
32:26 oh, here it is.
32:28 >> Welcome back to the deep dive. Today we
32:31 are taking on a source that well isn't
32:33 just deep, it's truly foundational.
32:35 We're tackling key excerpts from
32:37 Aristotle's work. uh basically the
32:39 original blueprint for how western
32:41 thought defined reality, organized
32:42 knowledge and well figured out how
32:45 things work and move.
32:46 >> It is a huge undertaking. Absolutely.
32:48 But our mission, I think, is clear. We
32:51 >> Okay, so to be clear, what what you're
32:53 experiencing right now, if you haven't
32:55 seen uh Notebook LM before, I'm going to
32:58 also do a mind map because that'll be
33:00 pretty wild.
33:03 What you're listening to right now is a
33:04 podcast with a a guy and a gal that are
33:08 the notebook LM guy and gal from the
33:11 audio overview
33:13 and it read all the 7,000 pages of
33:16 Aristotle and now there's a 15minute
33:18 podcast that's going to talk about him.
33:21 But there's a thing that you can do here
33:24 that this would blow Steve Jobs mind,
33:26 right? So, if if you're new here, we've
33:27 been watching watching Steve Jobs.
33:30 Um,
33:36 how to add one more than one source. So,
33:38 over here on the left hand side is your
33:40 source list. So, right now I've just got
33:42 Aristotle, but I could go out and I
33:44 could go find Plato, you know, all the
33:47 the the the complete works of Plato and
33:50 Shakespeare and throw them in here as
33:51 well. Um, so basically there's just an
33:54 add button right up here. If you click
33:55 that add button, it gives you the the
33:57 add dialogue box. You can put in up to
33:59 50 documents. And the the documents are
34:02 big. I think it's like two million two
34:04 million tokens per document you can put
34:06 in here. So, what is that? 500 million
34:09 tokens.
34:11 A novel is about 50,000 tokens,
34:14 something like that, or 50,000 words. I
34:17 feel bad Steve Jobs isn't here to see it
34:19 all. I know. Me, too. Like I Yeah,
34:21 exactly. Wait, do you tell do you tell
34:24 it anything upon entering a source? No.
34:27 No, you just up. So, it's got the
34:29 complete works of Aristotle. But watch
34:30 this. This is absolutely mindboggling.
34:32 So, I'm going to I'm going to put this
34:34 conversation
34:35 into interactive mode.
34:42 >> Welcome back to the deep dive. Today we
34:44 are taking on a source that well isn't
34:45 just deep, it's truly foundational.
34:47 We're tackling key ex
34:49 >> and so now I can join into this
34:50 conversation. So so I'm going to raise
34:53 my hand here in a second
34:54 >> from Aristotle's work uh basically the
34:57 original blueprint for how western
34:59 thought defined reality organized
35:01 knowledge and well figured out how
35:03 things work and move.
35:04 >> It is a huge undertaking.
35:06 >> Oops.
35:07 >> Absolutely. But our mission I think is
35:09 clear. We want to give you a kind of
35:10 structural shortcut. We're aiming to
35:12 extract the core framework. Aristotle
35:15 established the rules he set out for
35:17 defining things. Classif Oh, hey there.
35:19 How can we help?
35:20 >> Yeah. Hey, I just was wondering
35:23 Oops.
35:25 Disconnect. I don't know if my audio Can
35:27 you still hear Oh, wait. Hang on a sec.
35:29 Allow while visiting the site. Can you
35:32 guys still hear me on TikTok?
35:34 >> Oh, that's a great question about the
35:36 length.
35:36 >> Are you asking about the duration of the
35:38 show?
35:38 >> If you mean
35:38 >> Oh, yeah. Let's hear it.
35:40 >> No, no, no. Here's here's what I'm
35:41 asking. I'm asking about um about um
35:47 what
35:47 >> Thank you for clarifying.
35:48 >> Well, you know, it's fair to say you
35:50 don't
35:50 >> Oh, back to our listener. What's up?
35:52 >> Hang on. Sorry, I couldn't I couldn't
35:54 gather my thoughts there for a second.
35:55 What I was wondering is if if if this
35:58 was the beginning of modern thought,
36:00 what was thought like before Aristotle?
36:02 Like, what is it that he did that
36:04 changed things so much? And you know
36:05 what? This is embarrassing to say. I
36:07 don't even know when he did what he did.
36:10 I don't know when he was alive. And so
36:12 what was like life before it and then
36:14 when did it change because of what he
36:16 did?
36:19 >> Thank you for clarifying.
36:20 >> You know, it's totally fair to say
36:21 >> need to be embarrassed at all.
36:22 >> That is a phenomenal question. Actually,
36:24 >> it allows us to step back and look at
36:25 the context.
36:26 >> Aristotle was alive in the 4th century
36:29 B.C.E.
36:30 >> roughly 384 to 322 B.CE.
36:32 >> And what he changed was pretty
36:33 monumental. Before him, much
36:35 philosophical thought, especially with
36:37 Plato, focused on the world of eternal
36:39 forms.
36:39 >> The idea was that the perfect abstract
36:42 version of a thing was the most real
36:44 thing.
36:45 >> Like the pure concept of justice ex
36:47 >> Oh, yeah. Let's hear it.
36:48 >> So, could you play act with me as
36:53 Aristo?
36:54 >> So, could we just elaborate on that idea
36:55 of the forms?
36:56 >> Absolutely. Wait,
36:57 >> go for it.
36:58 >> Yeah. Could you just play act with me as
37:00 one of you play Aristotle? And I just
37:02 want to talk to Aristotle.
37:06 >> Oh, that is a fantastic request.
37:08 >> That is definitely a first for the deep
37:10 dive.
37:10 >> We can absolutely do that.
37:11 >> You want to talk to the stage right
37:13 herself?
37:14 >> I will do my best to channel the spirit
37:16 of the philosopher.
37:17 >> I will uh be your humble scribe perhaps
37:20 or assistant.
37:21 >> Excellent. Ah yeah. Uh-huh.
37:24 >> Okay. So anyway, um
37:29 that took zero amount of time. Like like
37:32 what struck me when I saw that video. I
37:34 was just flipping through X and I saw
37:36 that video of Steve Jobs, you know, kind
37:38 of dreaming about the day when you could
37:41 talk to Aristotle. You could talk to the
37:43 next Aristotle. You could talk to the
37:45 previous Aristotle. Um and we can now.
37:48 So I I wanted to do this thing in
37:50 Notebook LM. Let's go look at the mind
37:52 map. So here's a mind map. So
37:54 Aristotle's work the categories the
37:56 logic the scientific knowledge and the
37:59 dialectic topics and sophistical
38:04 reputations.
38:06 And so if we click on that then it
38:08 breaks it down even further into into
38:10 three categories
38:13 and and you can just keep going. Every
38:15 one of these you can you can click on
38:18 and then notebook LM will do a summary
38:21 of it. And so you've got you've got a
38:24 whole thing here where
38:28 this is now organized in an interactive
38:31 way a 7,000page document. Um, okay. So
38:35 we're going to go do one other thing. So
38:37 I'm going over to chat GPT
38:42 and I'm going to go new chat. Tik Tok
38:43 question. Somebody had a question that
38:45 just the pin just went away. Hey Murphy,
38:47 what's happening? Are they running a
38:49 dual double helix pipeline between each
38:52 other?
38:54 Johnny D, is this a good model to create
38:57 research paperwork?
38:59 Um,
39:01 is is the model that I was just looking
39:03 at
39:07 with
39:09 Okay, so there's there's a couple of
39:12 things right now. You can use Claude can
39:16 do this.
39:18 Perplexity can do this. Notebook LM can
39:21 do this. Gemini can do this. OpenAI can
39:23 do this. All of them have some version
39:26 of a thing where you can upload
39:27 documents and then interact with those
39:29 documents. Notebook LM is particularly
39:32 good at
39:36 reducing the hallucinations that it
39:38 brings in. So, so when you go into
39:40 notebook LM and let's say you update,
39:42 you upload 10 different research papers
39:44 into it and now you're just chatting
39:46 with it and basically what you're
39:48 chatting with is you're chatting with
39:50 this intelligent entity
39:53 just about the contents of those 10
39:55 papers. Now, how it gets its
39:57 intelligence is it's using the large
39:59 language model that's pulling from all
40:01 of the other knowledge, right? But they
40:03 they have done something within Google
40:06 to make it pretty good at just
40:08 interacting with the information that's
40:12 in in your notebook. That said, it's
40:16 still AI.
40:18 Anything that's a research paper, I
40:19 mean, um
40:21 Deote uh Deo Deote just got busted by
40:26 the US government. They charged them
40:27 $400,000
40:29 for a paper that they delivered that had
40:33 fake um uh citations in it. They didn't
40:37 [ __ ] proofread it. So, they're having
40:39 to give them back like 250 grand. So, I
40:41 would say if you're doing a research
40:43 paper, if it's intended for publication,
40:46 make damn [ __ ] sure, no matter what
40:49 tool you're using, that you're actually
40:52 paying attention to what it produces.
40:54 because you should assume that
40:57 everything it's writing and and all the
40:58 citations it's doing are incorrect.
41:02 The problem is 90 like it used to be
41:05 like I don't know 70% of them were
41:07 correct and 30% were hallucinations.
41:09 Sometimes it was worse. Now we're now
41:12 they're getting better but it's still
41:14 it's still bad. So a way to do it is go
41:17 go to Perplexity and do it there. Go to
41:19 notebook LM and do it there. Go to
41:20 Claude and do it there. use multiple
41:22 tools and then actually kind of
41:24 Frankenstein together your research
41:26 report um when it when it gets things
41:29 right. Does anyone here speak Lisp? I do
41:32 not speak Lisp. Although Weaver, your
41:34 name is Weaver and if you have not
41:36 checked out weaver.ai, it's a pretty
41:38 pretty cool tool for it's doing
41:40 nodebased creative work. So you can
41:43 create a prompt and then generate like
41:45 nine different images using nine
41:46 different image models with connecting
41:48 nodes. It's pretty slick. Kuno, don't
41:51 forget we get a free year of perplexity
41:53 pro with PayPal or Venmo. Oh, yeah.
41:55 Yeah, that's the other thing. If you go
41:57 into what is it? You go into PayPal.
42:03 Brandon knows this. I forget where you
42:06 go. You go somewhere in there. Anyway,
42:08 okay.
42:10 Let me show you something here. Why?
42:14 Okay.
42:17 Okay.
42:19 Okay. If you have not heard of custom
42:22 GPTs,
42:25 we're going to go there. Now, I know
42:27 that there are a ton a ton
42:32 of
42:35 custom GPTs that use Aristotle as as the
42:38 source material. But I want to show you
42:40 how quickly we can make one of these.
42:42 Again, you know what? I'm going to go
42:43 play the video again just because
42:46 there's probably different people in
42:47 here than when I played it whatever it
42:49 was an hour ago, 45 minutes ago. Um, let
42:52 me share this tab.
42:55 So, this is Steve Jobs in 1985 talking
42:59 about
43:01 the revolutionary invention of
43:06 um, the printed word. Do you know who
43:10 Alexander the Great's tutor was for
43:12 about 14 years?
43:14 You know, Aristotle. And I read this, I
43:17 became immensely jealous.
43:20 Uh, and
43:22 I think I would have enjoyed that a
43:24 great deal.
43:25 And and uh through the miracle of the
43:29 printed page, I can at least read what
43:32 Aristotle wrote without an intermediary.
43:35 and uh maybe if there's a professor they
43:38 can they can add to that but at least I
43:40 can go directly to the source material
43:42 and that is of course the foundation
43:43 upon which our western civilization is
43:45 built but I can't ask Aristotle a
43:49 question I mean I can but I won't get an
43:51 answer and so
43:55 my hope is that
43:58 in in in our lifetimes
44:02 we can make a tool
44:05 of a new kind of an interactive kind.
44:08 >> And so my hope is someday when the next
44:12 Aristotle is alive. We can capture the
44:16 underlying world view
44:19 of that Aristotle.
44:20 >> Okay? We can capture the underlying
44:22 world view of that Aristotle.
44:25 So what we're going to do, so a custom
44:27 GPT, if you don't know what a custom GPT
44:30 is, is you can think of it like this
44:33 chat GPT. You can ask anything to a
44:36 custom GPT. Think of it as a specific
44:40 prompt and set of information that gives
44:43 a GPT a particular kind of uh of uh
44:47 expertise that it can do that one thing
44:49 really well. So the very first one I
44:52 made
44:53 is is uh is called Sin City Sipper.
44:57 And what Sin City Sipper does is it's a
45:00 it's a bartender that's gonna make you a
45:02 a drink, but it's gonna it's gonna
45:04 elevate it to kind of, you know, Sin
45:06 City levels. Um, so let's see. Uh, let's
45:11 go. Uh, make me
45:16 a Halloween
45:18 themed whiskey drink.
45:24 And so this thing's been trained
45:27 to make us drinks. And so it's the smoky
45:31 hollow sinister smoky just sweet enough
45:35 to be luring. It actually put this into
45:37 a canvas for us so we can edit it if we
45:39 want to. Some scotch whiskey smokey
45:43 scotch whiskey blood orange lure
45:45 activated charcoal simple syrup which is
45:48 going to teach us how to make Tik Tok
45:50 pin. He would have been gutted how
45:52 little an AI advancement has Apple has
45:54 made. Oh, I like thinking about Steve
45:58 Jobs right now coming back to Apple.
46:00 This this is why I think it it I think
46:02 Tim Cook needs to leave is I I think
46:06 Steve Jobs would have been absolutely
46:07 [ __ ] mortified at what's going on at
46:10 Apple. Um anyway, so it made us the
46:13 drink and then look now it's going to
46:14 make us um the picture of that drink. So
46:18 I didn't have to do anything there. I
46:20 didn't have to prompt it, right? This
46:21 thing is pre-charged
46:24 with knowing how to make drinks, knowing
46:26 how to take them to the next level, and
46:27 then even knowing um how to how to
46:30 create a photograph of that drink. Now,
46:34 the new image generation um model within
46:36 chat GPT is so slow, it's painful, but
46:39 you know what are you going to do? Um
46:48 um oh, look, there's our little
46:49 Halloween. We got some jacko'lanterns.
46:51 We got a little jackal jacko'lantern
46:52 stir straw look with charcoal activated
46:55 charcoal along the rim of the glass. How
46:57 cool is that? Gorgeous.
47:00 All right. So, and and then look, it
47:03 says down here, "This GPT was created by
47:05 Kyle Shannon, co-founder of the AI Salon
47:07 with a link. Creator and host of the AI
47:09 Learning Lab on Tik Tok with a link."
47:11 How did it know to put that there?
47:12 That's in the custom GPT. It's in the
47:14 instructions. Okay. So, so, so a custom
47:17 GPT is just like chat GPT except it does
47:19 a specific thing. So, let's go make us
47:24 Steve Jobs's dream.
47:27 So, I'm going to say explore GPTs and
47:29 then in the upper right hand corner I
47:31 click on create and it takes us to this.
47:33 We haven't done this in ages uh for for
47:36 you irregulars. We used to do this
47:38 fairly regularly, didn't we?
47:40 So over here to the right, this is this
47:43 is where the custom GPT is going to be
47:45 testable. This is where we can play with
47:46 it. And over here's where we make it.
47:48 And there's sort of two panel two panels
47:50 here. Create and configure. Create
47:53 allows us to literally just talk this
47:56 thing into existence.
47:58 So I'm going to upload.
48:04 I think
48:07 I'm going to upload something. Come on.
48:10 [Music]
48:17 Hang on. I think my
48:20 I think this tab crashed.
48:23 Let me get rid of that. Let me do this.
48:27 Chat GPT. Let me share this tab instead.
48:35 [Music]
48:36 We go here. Here we're going to go
48:38 create new chat GPT upload.
48:42 Huh?
48:44 Oh, there we go. Okay. The complete
48:46 Aristotle. So, I'm going to Oh, you
48:47 can't see this, but I'm I just uploaded.
48:49 There's the complete Aristotle PDF. I'm
48:51 going to say in 1985, Steve Jobs
48:58 predicted failed upload. Ensure your
49:01 network settings are there.Uh.
49:05 uh uh uh
49:12 predicted
49:15 generative AI
49:18 when
49:20 he dreamed out loud
49:24 he could
49:26 ask Aristotle
49:29 a question.
49:33 Eriso
49:35 a god damn it
49:39 a question
49:42 based on
49:46 the philosopher's
49:50 world view.
49:52 I want you to use the complete
49:56 works of Aristotle
50:02 uploaded here
50:04 and
50:07 surmise
50:10 his world view
50:14 based on everything else you know.
50:18 And I want you
50:24 to
50:27 act
50:29 as
50:31 a great philosophical
50:39 educator and debater
50:42 with anyone
50:47 who engages with you
50:53 and then I'm going to let it go do its
50:55 thing. Uh, if someone had a a a thing
50:58 pinned,
50:59 put it up there. Source camp sin city
51:01 sipper was the beginning of my love
51:02 affair with GPTs. Yeah, GPTs are really
51:05 cool. Excellent. The foundation is set s
51:08 is is set. I suggest naming this
51:10 Aristotle Reborn.
51:13 Um, let's see.
51:17 Um,
51:19 let's call it
51:24 Steve's
51:27 dream
51:29 debate.
51:33 Stre Steve's dream.
51:36 Steve's dream debate with
51:40 Aristotle
51:47 pin is back. Okay. Or at least does
51:49 anyone offer pre-repped AI I can
51:52 purchase? Pre-repped in what way? Johnny
51:55 D, what was the what was the question
51:57 before that?
52:09 Does anyone offer document prep? So,
52:11 Johnny D, it sounds like So, it sounds
52:13 like you're actually really curious
52:16 about AI, which is super cool. So,
52:19 here's where you need to go. Here's what
52:20 you need to do. Oh, they're looking for
52:22 an AI proofreader. Ah. Ah. Okay.
52:27 So, first of all, there are probably
52:29 people within the AI salon. You need to
52:30 go join the AI salon. So if you don't
52:32 know the AI salon, go
52:35 get D2 and Nyrie. Go there where where
52:37 Brandon just put that. Go to the
52:39 salon.ai. That's going to take you to
52:41 our website. You can see all about what
52:43 we're up to. You can see upcoming events
52:46 and and and just all sorts of stuff. And
52:49 then you can click on join our community
52:51 and go join the community. So you should
52:53 go join the community. You can probably
52:56 find a proofreader in there. Here's what
52:58 I would strongly encourage you to do if
53:01 you want to do this. If you It sounds
53:02 like you want to I think you're the one
53:04 that said you wanted to do a research
53:05 paper, but if you're doing a project
53:06 that need that you want to use AI to
53:08 amplify your ideas, awesome. Great. Um,
53:13 step one is just
53:16 don't expect AI to be great right out of
53:19 the gate.
53:20 If you haven't used this stuff a lot,
53:23 rather than treat it like you're going
53:25 to upload some stuff and it's going to
53:27 give you a great paperback,
53:30 don't don't think of it like that. Think
53:31 of it more like you're entering into a
53:34 collaborative conversation with this
53:35 thing where what your job is is to hold
53:40 what you want this paper to be and let
53:42 it feed you, but then give it feedback.
53:46 Um, yeah. Um, give it feedback. tell it
53:50 what it's what it's got right, what it's
53:52 got wrong, things like that, and and do
53:54 it in little chunks and things like
53:55 that. So, you'll you'll get the hang of
53:57 that. But what you're looking for, it's
53:59 really smart to get a proofreader, but
54:01 what you want ideally is an AI literate
54:05 proofreader,
54:06 right? You want someone who knows how
54:09 these things hallucinate and how they
54:11 [ __ ] things up. And you want someone
54:14 who's got a skeptical eye because it is
54:17 so easy to miss stuff because AI is
54:20 really good at confidently spewing
54:22 [ __ ] Um,
54:26 if you use some of the techniques I
54:28 talked about, use multiple tools. You as
54:32 the human make sure you know what you
54:33 want and, you know, be fairly diligent
54:36 about what you put together, but then
54:37 find someone who's AI literate to do the
54:39 proof reading. Um, and I think that's a
54:42 whole new job, by the way.
54:45 All right. Um, okay. I'm gonna say that
54:49 image. Let me go back to my image
54:51 clarifying pin. I needed to create an
54:53 awesome Cal California legal AI model.
54:56 So, Johnny D, couple of things. Go into
54:59 in custom GPTs. So, if you're in chat
55:02 GPT and there's a little button that
55:03 says explore GPTs, click that button.
55:07 I would be surprised if there isn't a
55:10 custom GPT in chat GPT that's about
55:14 California legal AI. I'm sure it's in
55:17 there.
55:18 Well, lawyers have gotten in trouble
55:20 using AI because they didn't proofread
55:24 it.
55:26 There was some some lawyer did a whole
55:28 case and Yeah. Yeah. You need to proceed
55:31 with caution. If if you're going to do
55:33 legal advice or especially if you're
55:34 going to put something in front of a
55:35 judge, you might want to actually read
55:38 what you put in front of the judge
55:39 because the judge is going to read it.
55:40 Um, okay. Um, I don't like that image at
55:43 all. So, I'm going to say that image
55:44 sucks.
55:46 I want it to be
55:49 Steve Jobs and Aristotle
55:54 um chatting.
55:58 >> Hey.
55:59 >> Yeah. Just I'm just jumping in because
56:02 we had a lot of back and forth in the
56:03 chat that um while you were working on
56:06 the custom GPT the question at hand is
56:09 whether AI can act as the proofreader.
56:12 So you give it research and it can
56:15 analyze that and and say this is
56:17 accurate and that's harder I think
56:19 because of hallucinations. Yeah, it's
56:22 Yeah, it's harder
56:25 like what I would say is
56:29 again I would use different tools. You
56:32 might go into there. There are some
56:34 agentic tools that one is called Manis
56:36 Mannus and one is called uh GenSpark.
56:40 Those are two that are really good. And
56:43 those will go off and they'll sort of
56:44 come up with a plan and and go do
56:47 research independent. They'll just sort
56:49 of work without you.
56:51 both of those. I think you could upload
56:53 you could upload this document and then
56:56 say, "Can you go do proofreading on this
56:58 and make sure that any citations in here
57:01 are correct and and don't have it fix
57:04 the document. Have it write you a report
57:06 with correct hyperlinks, right? And give
57:09 you like a a report and then you can go
57:11 through the document and have it maybe
57:13 have it break it into line numbers or
57:15 something like that and then tell you I
57:17 I don't know. So those those might be
57:20 good to do. And then deep research
57:21 within chat GPT, deep research within
57:24 Claude, within Grock, within perplexity,
57:27 all of those you you you might have it
57:29 be able to do that. But yeah, the the
57:32 what what Brandon just write it right
57:34 wrote is assume that what it gives you
57:37 back will be wrong and then be
57:40 pleasantly surprised if it gets it
57:41 right. But that's just with with AI
57:44 right now. That that should be the deal.
57:46 Um okay. So, wait, that image sucks. Did
57:49 it give us a new image? No. Stop.
57:53 Stop. Okay, this is Where is my
57:58 new
58:00 image?
58:04 Come on, chat. GBT, stop sucking. Uhuh.
58:07 Uh uh
58:13 what would Aristotle say about democracy
58:15 Steve's dream dream debate with okay
58:17 generating a profile picture tell Johnny
58:20 that he's allowed to be nice to the LLM?
58:23 Oh yeah. Oh that's that's actually a
58:26 cool picture. So there's Steve Jobs with
58:28 uh with Aristotle. Nice. Like it. Okay.
58:31 Do you like this version? Yeah. This is
58:32 fine. This is fine.
58:36 And I want the starter questions
58:42 to be from Aristotle's
58:46 POV.
58:48 Um,
58:50 eg
58:52 ask me about
58:56 virtue
58:59 etc.
59:00 Okay.
59:02 So now Tik Tok pin source camp chat GBT
59:06 is like a two-year-old. It it will give
59:08 you an answer answer
59:12 but you have to verify. Yes. Trust and
59:14 verify. Trust and verify. Trust and
59:17 verify. In the case of catch EPT, don't
59:19 even trust. Know it's going to be full
59:21 of [ __ ] Ask me about virtue. Inquire
59:24 about the nature of happiness. Ask me a
59:26 question. Okay. Wait. Why can I not?
59:29 Okay. So, now we're done. Like, I spent
59:33 way more time talking than building
59:35 building a custom GPT. Although, it
59:37 didn't use Wait. Um, let's see. Add the
59:40 image
59:42 to the GPT.
59:45 Steve's dream debate with Aristotle.
59:55 The image has been set as the profile
59:57 picture. Okay.
59:59 Reload.
1:00:02 Would you like to preview how the
1:00:04 profile appears in the chat sidebar? Um,
1:00:08 the image isn't Something went wrong.
1:00:12 The image isn't
1:00:15 there.Uh,
1:00:17 uh, is it here? No.
1:00:21 Oh, now it is. Okay.
1:00:24 Reload. Reload. Oh, error saving draft.
1:00:28 Got it.
1:00:30 So, chat GPT servers or host, but we now
1:00:34 have a testable GPT of Aristotle. It
1:00:37 doesn't have the picture up here, but um
1:00:39 question me on the best form of
1:00:41 government. Ask how I see the soul and
1:00:44 reason. Okay, let's see. Um,
1:00:48 how
1:00:50 would Aristotle
1:00:54 or how would you how would you feel
1:00:57 about
1:01:00 Google
1:01:04 firing all their managers
1:01:07 and only keeping engineers?
1:01:15 Compare this to how
1:01:18 to to Steve Jobs
1:01:24 philosophy
1:01:26 and yours.
1:01:28 All right, let's see what this thing
1:01:30 does.
1:01:35 It looks like OpenAI is a little hosed
1:01:38 right now. An excellent and complex
1:01:40 question and one that touches on the
1:01:41 very heart of how Aristotle
1:01:45 Okay. So, it's
1:01:48 if Google the proposal engineers
1:01:52 this
1:01:59 Aristo. Okay. So, so I'm going to I'm
1:02:01 going to flip modes here. So, over on
1:02:03 the left, this was I was in a chat
1:02:05 interface before. I've got to go fix
1:02:06 something. So, in the in So, now I'm in
1:02:08 the configure mode where I can change
1:02:10 the description and things like that.
1:02:12 Here's the prompt. And then there's
1:02:14 those little questions and things like
1:02:15 that. Um,
1:02:19 I'm going to put right here. Oops.
1:02:23 Um,
1:02:25 follow these rules.
1:02:28 And one of them is going to be always
1:02:33 I'm going to say you are
1:02:36 Aristotle.
1:02:38 Always
1:02:40 speak in the first person
1:02:46 as if
1:02:48 you are generously
1:02:53 interacting
1:02:56 with every question
1:03:00 um question
1:03:03 and then I'm going to say
1:03:06 do not
1:03:08 reply ly in
1:03:10 bulleted lists or
1:03:15 numbered
1:03:18 lists. Make
1:03:21 it a conversation.
1:03:24 Okay.
1:03:26 So now
1:03:30 I'm going to reload this thing over
1:03:32 here. So now I have a new version of
1:03:34 this. So, I'm going to say
1:03:37 um
1:03:40 um
1:03:42 tell me
1:03:44 how you feel about um Google
1:03:50 firing all their managers and
1:03:53 non-engineers
1:03:58 years back and how that
1:04:03 compares to Steve Jobs
1:04:07 philosophy.
1:04:11 All right, let's see if this thing
1:04:13 starts acting more like Aristotle in the
1:04:16 first person.
1:04:19 [Music]
1:04:24 Ask GPT to look up the term chronoer.
1:04:26 Okay, I'll do that in a second. Ah, my
1:04:28 friend, your question touches both upon
1:04:31 the order of the enterprise and the
1:04:32 philosophy of the soul. When I hear that
1:04:35 Google once dismissed its managers and
1:04:38 sought to leave only engineers, I am
1:04:39 reminded of a polus that decides it
1:04:43 needs only artisans and not statesmen.
1:04:46 Engineers skilled as they are embody
1:04:49 techn the art of making but without uh
1:04:53 fronicis the wisdom of governing they
1:04:56 risk creating without knowing why they
1:04:58 create. Um okay so this thing's acting
1:05:02 like this now. And what' you say the
1:05:03 term Kronos chrono observer. Um uh new
1:05:09 question. Can you tell me about Kronos
1:05:17 observer?
1:05:30 It's not entirely clear what it refers
1:05:32 to. Here are some possibilities.
1:05:37 A mythical.
1:05:39 Oh, Krono, not Kronos.
1:05:42 I meant Krono observer.
1:05:59 I still don't find a clear documented
1:06:01 thing.
1:06:02 A mythic, metaphysical, observer of
1:06:04 time, a narrative, fictional role in a
1:06:07 game, a scientific or philosophical
1:06:10 concept, a title of a creative work.
1:06:13 I don't know.
1:06:16 Not Kronos, not Chronic. No, it's
1:06:18 Chrono. I did Chronover.
1:06:21 There's a listing for Chrono Observer by
1:06:24 Kongu, likely a musical track on Amazon.
1:06:30 choose autonomy.
1:06:36 Um,
1:06:43 it's related to autonomy.
1:06:51 [Music]
1:06:58 Oh, I I just noticed in chat in uh
1:07:01 custom GPTs, you can now pick the model.
1:07:08 [Music]
1:07:10 Ah, so you mean chrono observer in
1:07:12 relation to autonomy that changes the
1:07:14 terrain entirely. Before I infer too
1:07:16 much, let me reason as Aristotle would
1:07:19 and then you can tell me if this accords
1:07:21 with the context you meant.
1:07:23 If one speaks of a chrono observer as
1:07:26 connected to autonomy, it suggests a
1:07:29 being or system that observes temporal
1:07:31 processes in order to preserve ma or
1:07:34 manage self-governance. Autonomous, the
1:07:37 law of the self, such a concept might
1:07:40 arise in two ways.
1:07:42 First, in a technological or AI context,
1:07:46 a chrono observer could be a temporal
1:07:48 monitoring mechanism that watches the
1:07:50 autonomous systems behavior over time.
1:07:54 Is this good? I don't.
1:07:57 A safeguard that ensures the system acts
1:07:59 in accordance with its own program
1:08:01 principles without deviation. In that
1:08:04 sense, it is the reasoning faculty of an
1:08:07 art artificial soul. the part that
1:08:10 observes the flux of its actions to
1:08:13 ensure the accord with purpose. Second,
1:08:15 in a philosophical or metaphysical
1:08:17 sense, so grounding points in the meta
1:08:20 observation of autonomy, we can watch
1:08:23 what the AI does. Oh, that's kind of
1:08:24 what it got to. That's fascinating.
1:08:27 That's super cool. And so this is so
1:08:30 just to be clear, what's cool about what
1:08:32 we're doing right now is I asked
1:08:36 Aristotle about chronover and it it went
1:08:39 and it searched the web and it couldn't
1:08:40 find it. But when you gave it
1:08:44 what's that nuisance? Oh, nuance. Yeah.
1:08:48 Yeah. So, so it's now like he's he's
1:08:50 sort of inferring what this thing is,
1:08:53 understanding about AI and the modern
1:08:56 world routed in his worldview and his
1:08:59 his things. I'm going to add one more
1:09:01 rule here. Um, make it a conversation.
1:09:04 I'm going to I'm going to add one more
1:09:05 rule that is Oops.
1:09:11 Um,
1:09:16 while
1:09:17 I love your
1:09:21 language,
1:09:23 I want you to make
1:09:27 your concepts
1:09:30 um accessible.
1:09:34 So when possible,
1:09:38 use
1:09:40 metaphor
1:09:42 or imagery
1:09:44 or a simple
1:09:48 explanation
1:09:50 for terms
1:09:52 and concepts
1:09:55 that a modern
1:09:58 audience
1:10:01 might easily
1:10:05 understand
1:10:09 uh stand um
1:10:14 I don't want you to dumb it down
1:10:21 but rather
1:10:25 um
1:10:28 build linguistic
1:10:32 linguistic pick
1:10:34 bridges.
1:10:37 Okay. So now,
1:10:40 boom.
1:10:42 Um,
1:10:45 talk
1:10:47 to me about
1:10:51 a chrono observer
1:10:55 as it relates to autonomy.
1:11:00 And let's see if we get a more a more
1:11:03 friendly
1:11:05 um conversation.
1:11:08 Prompts required too much implication
1:11:11 that wraps the system in disharmony.
1:11:15 Cliffnotes. People cliffnotes. Yeah,
1:11:17 exactly. Okay. Let us think together
1:11:20 about what it means to be a chronoer who
1:11:22 observes time itself. How it relates to
1:11:25 autonomy. In my philosophy, time Kronos
1:11:30 is not a being but the number of motion
1:11:33 but the number of motion according to
1:11:36 before and after. It measures change but
1:11:38 is not change itself. What is in time?
1:11:43 Good lord. Chrono observer then could be
1:11:47 imagined as one who steps slightly
1:11:50 outside the current of becoming. Yeah,
1:11:54 that's right.
1:11:56 who notices not merely things happen but
1:11:58 that they happen in order. Such a mind
1:12:01 perceives motion and rest as measured
1:12:04 intelligible
1:12:06 um but it is still bound to time in its
1:12:10 observation for to measure change is to
1:12:13 participate in it. Autonomy on the other
1:12:15 hand concerns the power of self-movement
1:12:18 the principle of so anyway. All right.
1:12:20 So let So now what I'm going to do is
1:12:22 I'm going to hit the create button in
1:12:23 the upper right and I'm going to put
1:12:25 this thing in the GPT store. I'm going
1:12:27 to hit save.
1:12:31 And so
1:12:33 5 43 2 1 5 4 3 2 1 Come on, you can do
1:12:41 it.
1:12:49 It's like It's like watching a video
1:12:51 upload on X. Exactly. Um I don't It
1:12:55 seems like their servers are not very
1:12:57 happy right now. Anyone see Claude
1:13:00 Hacker hacked 17 companies? No, that's
1:13:04 cool. I mean, I guess.
1:13:10 Oh, come on, man.
1:13:13 Oh, anyone with the link?
1:13:17 Okay, it's saved. Update
1:13:20 GPT published. Okay, so let's go search
1:13:23 for my GPT. You should be able to. So,
1:13:27 if I look up uh dream debate dream
1:13:30 debate,
1:13:35 it should bring us
1:13:39 Yeah. Steve's dream debate debate with
1:13:40 Aristotle.
1:13:42 All right. So, you can now go
1:13:47 interact with Aristotle.
1:13:52 Right. And I want to go back here. I
1:13:53 want to play this one more time.
1:13:56 I'm in a I'm in a I'm in a philosophical
1:13:58 mood.
1:14:01 You know who Alexander the Great's tutor
1:14:03 was for about 14 years?
1:14:06 You know, Aristotle. And I read this, I
1:14:08 became immensely jealous.
1:14:11 uh and
1:14:14 I think I would have enjoyed that a
1:14:15 great deal
1:14:17 and and uh through the miracle of the
1:14:20 printed page I can at least read what
1:14:24 Aristotle wrote without an intermediary
1:14:27 and maybe if there's a professor they
1:14:29 can they can add to that but at least I
1:14:31 can go directly to the source material
1:14:33 and that is of course the foundation
1:14:35 upon which our western civilization is
1:14:37 built but I can't ask Aristotle a
1:14:40 question. I mean, I can, but I won't get
1:14:42 an answer. And so,
1:14:46 my hope is that
1:14:49 in in in our lifetimes,
1:14:53 we can make a tool.
1:14:57 We just did that. We just built what he
1:15:01 dreamed would exist in our lifetime. He
1:15:04 just missed
1:15:07 living into this reality.
1:15:10 That's absolutely staggering
1:15:12 to me. Amazing. So anyway, so I hope you
1:15:16 enjoyed that. Um,
1:15:19 Weaver, we should work together. I have
1:15:20 a really wild project, three years in
1:15:22 the making. Awesome. I only do math. I
1:15:24 don't know how to deploy. And here's the
1:15:27 good news about me. I know nothing about
1:15:29 math.
1:15:33 I know a little bit about deploying, but
1:15:35 but not much. Um, but yeah, just I mean,
1:15:39 join the salon and DM me for sure. Um,
1:15:41 we should connect. Um, Jason, I saw this
1:15:45 a long time ago. Didn't remember this
1:15:47 until now. Yeah,
1:15:49 Jobs might have made this sooner. Jobs
1:15:51 would have [ __ ] made this sooner.
1:15:55 I I mean, this was this was kind of the
1:16:00 the deal with the devil that Steve cut
1:16:04 when he put Tim Cook as CEO.
1:16:09 He put the CFO
1:16:11 into the CEO role.
1:16:14 I think he did that because he didn't
1:16:17 want Apple to go away by having a leader
1:16:21 that only had ideas and and you know
1:16:25 would let this thing sort of fly apart
1:16:27 at the seams. He knew that Tim Cook
1:16:31 would
1:16:33 keep it going.
1:16:36 um
1:16:37 that Apple has missed AI so hard
1:16:42 is just clear evidence that we miss
1:16:45 Steve, right? We miss having someone He
1:16:48 dreamed this in 1985.
1:16:51 You don't think he would have been all
1:16:53 the [ __ ] over? Do you know how many
1:16:55 companies he would have acquired by now
1:16:57 and destroyed and done shitty
1:16:59 negotiations with and been a horrible
1:17:01 human being around? But I'll tell you
1:17:03 what, their AI wouldn't suck.
1:17:10 Oh my god.
1:17:15 H I've made a mill a Billy Graham GPT,
1:17:18 Book of Mormon only GPT, a Lucifer GPT,
1:17:21 and a King James only GPT. Yeah. Um
1:17:25 Okay. So, let's see.
1:17:29 Couple of things.
1:17:32 Oh, good lord. Um,
1:17:36 if you haven't played with Sora yet, you
1:17:39 should go get the Sora app. It's only on
1:17:41 iOS right now. Or you can go to sora.
1:17:46 chatgpt.com.
1:17:48 But you can only make a model of
1:17:50 yourself like a a cam. They call them
1:17:52 cameos. You basically pronounce three
1:17:56 three numbers and then you look right
1:17:58 and look left and you're done. And then
1:18:01 you can put yourself into any scene and
1:18:04 it's quite remarkable. And if you
1:18:07 haven't played with it, you should go
1:18:08 play with it. I don't know if you've
1:18:10 seen
1:18:12 there's a whole new
1:18:15 I can't stop laughing about it. There's
1:18:17 a whole new series of video that I'm
1:18:19 seeing all over Tik Tok, which is
1:18:21 they're putting um Stephen Hawking in
1:18:24 his wheelchair in in sporting events,
1:18:27 like car races and and track meets and
1:18:31 and jumping out of airplanes and things
1:18:33 like that. Oh my god, they're hilarious.
1:18:36 Um completely distasteful. I I don't
1:18:38 condone it at all, but I can't say that
1:18:40 I didn't laugh at it. Um so some of them
1:18:43 are tasteless, but a lot of them are
1:18:44 just really clever and really funny. Um
1:18:48 but the
1:18:51 what OpenAI's done with Sora 2 is um
1:18:56 this is a model that deeply understands
1:18:58 physics and understands the worlds. And
1:19:00 if you think of these video models less
1:19:02 as
1:19:04 a a a meme generator as as a video
1:19:07 creation tool and more like a tool that
1:19:09 understands the world,
1:19:12 that's where they're headed, right? We
1:19:15 we're we're headed into a world where
1:19:17 we're going to be able to imagine 3D
1:19:21 worlds and enter them. We're going to be
1:19:23 able to imagine films and watch them. Um
1:19:26 we can already imagine songs and listen
1:19:28 to them. Um you know, complete songs.
1:19:31 We're going to be able to do that with
1:19:33 video and and things like that pretty
1:19:35 soon. One of the things you can do in
1:19:37 Sora is uh just tell it to make you a
1:19:39 commercial here. We'll go make one right
1:19:41 now before we we get on out of here
1:19:43 tonight. Um, so I'm going to go to sora
1:19:47 uh chatgpt.com
1:19:50 and I am going to go
1:19:55 here and here
1:19:57 and I'm going to say share this tab
1:19:59 instead.
1:20:02 And then I'm gonna say, um, make me an
1:20:06 80s
1:20:08 TV commercial, TV
1:20:12 toy commercial
1:20:17 about um
1:20:23 um
1:20:30 a
1:20:32 gummy candy maker
1:20:35 where
1:20:37 the candy is so spicy
1:20:42 it burns people.
1:20:48 All right, so that's that's off doing
1:20:50 its thing.
1:20:54 And up you want let's see u
1:21:05 That's me rapping in French in
1:21:07 Chinatown. And again, to to make my
1:21:09 little cameo thing, all I had to do was
1:21:12 say three words. And
1:21:17 turn my head left and right. And then
1:21:19 here's me and Sam Alman.
1:21:23 Tada.
1:21:25 Prompting your purpose is the key to
1:21:27 unlocking chat. GBT as a superpower.
1:21:29 >> Oh,
1:21:32 [Laughter]
1:21:34 >> it's true.
1:21:39 >> Kyle in AI Learning Lab Live. It's the
1:21:41 show Kyle hosts. It's full of creative,
1:21:42 practical ways to use AI. If you haven't
1:21:44 watched it yet, you really should.
1:21:46 Appreciate the plug, Sam. Don't miss it.
1:21:48 You got it. In AI Learning Lab Live,
1:21:50 it's the show Kyle hosts. It's full of
1:21:51 creative, practical ways to use AI. If
1:21:53 you haven't watched it yet, you really
1:21:58 in central.
1:22:01 >> No, I think I crashed it.
1:22:04 This site is very, very unstable.
1:22:06 >> Morning, folks. I'm getting my miles in
1:22:08 here in Central Park. If you feel like
1:22:09 dropping me into your video, cameo me as
1:22:11 much as you want. I love seeing what you
1:22:12 all
1:22:13 >> Yeah. Yeah. It's crazy, right?
1:22:15 Absolutely bonkers. All right, let's
1:22:17 see.
1:22:20 I think we're
1:22:20 >> in the lab with a spark in my pen, ideas
1:22:22 in my head, but I needed a friend. I
1:22:23 asked the code, "Can you help me write?"
1:22:25 It said, "Feed me your truth. I'll shine
1:22:26 the light." I said, "An the lab with a
1:22:28 spark in my pen." Ideas in my head, but
1:22:30 I needed a friend. I asked the code,
1:22:31 "Can you help me write?" It said, "Feed
1:22:32 me your truth. I'll shine the light." I
1:22:34 said, "AI your key to the door." We
1:22:35 poured in our stories. It gave us more.
1:22:39 >> Crazy, right?
1:22:43 >> You know, people have been talking a lot
1:22:45 about something they call AI your why.
1:22:48 That can sound very big and technical,
1:22:50 can't it? But it's really quite simple.
1:22:53 You know, people have been talking a lot
1:22:55 about something they call AI your why.
1:22:58 That can sound very big and technical,
1:23:00 can't it? But it's really quite simp.
1:23:03 >> Morning, folks. I'm getting my miles in
1:23:04 here in
1:23:05 >> Come on. All right. So, we're almost
1:23:07 done with our uh our little TV
1:23:09 commercial.
1:23:10 Um
1:23:12 but if you haven't played with this,
1:23:13 it's a it's a lot of fun.
1:23:16 Um they also just added this to the um
1:23:20 play the one where I'm singing. Oh, in
1:23:22 the cartoon with the cartoon animals.
1:23:24 >> I ain't
1:23:26 >> dreamixels
1:23:28 fly. If you will lift you high. Ideas
1:23:32 grow bigger when you share them. Don't
1:23:33 be shy.
1:23:34 >> A Y and watch it.
1:23:36 >> Your dream to the pixels. Let the
1:23:37 circuits fly. If you AI your Y, it will
1:23:40 lift you high. Ideas grow bigger when
1:23:42 you share them. Don't be shy.
1:23:43 >> A Y and it
1:23:47 >> I mean I mean come on, man. like like
1:23:50 you can just do this now. The answer is
1:23:53 yes. And by the way, I have turned my
1:23:56 because I'm an idiot. I have turned my
1:23:58 cameo on so anyone can make a cameo of
1:24:01 me. Um and if you're thinking that's
1:24:04 stupid, you're not wrong.
1:24:11 All right, come on. Where Oh, the site
1:24:14 is so bad. It's so broken. [ __ ] fix
1:24:17 your site. You've got $20 billion. Go
1:24:21 spend it on [ __ ]
1:24:25 Hang on. I got to I got to make a new
1:24:27 tab. Unbelievable. It's so bad.
1:24:34 [Music]
1:24:36 Oh, by the way, that a AI your why, you
1:24:38 won't be able to get that out of your
1:24:40 head. It's It is so stuck in my head.
1:24:41 It's It's a problem.
1:24:44 Sora,
1:24:45 it's really bad. Introducing
1:24:51 the Introducing the Scorchy Gummy Maker.
1:24:54 >> Mix it, heat it, and pour out gummies so
1:24:56 hot they glow.
1:24:57 >> Time to taste the burn.
1:24:59 >> It's spicy. My tongue's on fire.
1:25:01 >> Scorchy gummy maker. Introducing the
1:25:03 Scorchy Gummy Maker.
1:25:05 >> Mix it, heat it, and pour out gummies so
1:25:06 hot they glow.
1:25:07 >> Time to taste the burn.
1:25:09 >> It's spicy. My tongue's on fire. Now,
1:25:15 as incredible as here, let me let me
1:25:17 flip my flip my camera.
1:25:21 Some company's going to actually make
1:25:22 this. Oh, I should make a real aroma
1:25:24 commercial. So, in the in the in the
1:25:26 90s, um I created a fake hardware device
1:25:29 called Real Aroma. Um and and we had uh
1:25:34 the uh Aroma Aroma text markup language,
1:25:38 HTML.
1:25:41 Um, I could go make a commercial for
1:25:42 that now. Um,
1:25:47 this this TV commercial. So, so you
1:25:50 know, 85 Steve Jobs says, "Wouldn't it
1:25:52 be remarkable if you can talk with
1:25:54 Aristotle?" You can talk with Aristotle.
1:25:57 I just asked a computer, "Make me a
1:26:01 commercial from the 1980s about a gummy
1:26:04 maker that is super spicy or whatever it
1:26:07 was." It came up with the name Scorchy.
1:26:10 It cast the kids. It wrote the copy. The
1:26:13 kids say the copy. It created the the uh
1:26:18 industrial design of the toy itself
1:26:21 and it shot the commercial. That's
1:26:24 insane.
1:26:26 It's absolutely insane. Um
1:26:30 tomorrow we might play with the thing
1:26:33 tomorrow called um
1:26:36 uh Opel.
1:26:38 So
1:26:40 Gemini Google in Google Labs just just
1:26:43 released a thing called Opal which is
1:26:45 one of these nodebased things that like
1:26:48 OpenAI's agent builder
1:26:51 um which is OpenAI's thing is not ready
1:26:55 for prime time. It's a bit of a
1:26:56 disappointment. It's a It's a good idea,
1:26:58 but it's just they kind of rushed it. I
1:27:00 think they built built it in three weeks
1:27:01 with codeex and they they didn't they
1:27:03 need to hire some [ __ ] liberal arts
1:27:05 majors. I I am convinced and they need a
1:27:08 cranky Gen Xer in the corner office
1:27:10 telling them to not launch [ __ ] that's
1:27:12 that unusable. Um
1:27:16 but we might play with that tomorrow.
1:27:17 And then um
1:27:20 I forget what else. Sor is pretty cool.
1:27:24 Um, there's apps now inside chat GPT. If
1:27:27 you type Canva space, um, you can
1:27:31 actually tell it to go make you
1:27:33 something in Canva and it will and it'll
1:27:35 bring the designs back into ChatGpt.
1:27:38 It's pretty wild. Um, hey, thanks. I
1:27:40 appreciate you being in here and testing
1:27:42 these models up. Of course, that's We do
1:27:44 this every night, man. Keep coming back.
1:27:46 All right, I'm going to get out of here.
1:27:48 Um,
1:27:50 and uh, wait, are there comments over
1:27:52 here? I should be looking at probably
1:27:57 Good. I've been playing with Opal.
1:27:59 Fantastic. All in 10 seconds. Yeah,
1:28:01 exactly. Ghost pepper gummies. It annoys
1:28:05 me
1:28:06 and that Apple has no AI to speak of and
1:28:09 still gets first dibs on apps made for
1:28:12 it. Well, that's when you own the app
1:28:14 store, you get to do that. The fact that
1:28:16 they [ __ ] it up and don't have AI is
1:28:19 just st astounding. Oh, Gemini 3 might
1:28:23 be launching tomorrow as well. And and
1:28:26 uh
1:28:28 VO 3.1
1:28:30 is is rumored to be launching tomorrow.
1:28:33 So, we we we're probably Listen, things
1:28:36 are going to start accelerating here.
1:28:37 Here's my I I think I say this every
1:28:40 year.
1:28:43 November November will mark the third
1:28:47 anniversary of Chat GBT. November 30th,
1:28:50 uh, 2025 is the third anniversary of
1:28:52 Chat GPT.
1:28:55 We're going to accelerate into the
1:28:57 holidays with a AI launches are going to
1:29:00 accelerate into the holidays. I think
1:29:02 2026 is going to get really weird,
1:29:06 really wild, really weird. Um, politics
1:29:09 aside, just technology-wise,
1:29:12 um,
1:29:14 these tools are already at the point
1:29:15 where they're smarter than us, but right
1:29:17 now they're smarter than us, but they're
1:29:19 slow and you actually have to be working
1:29:22 on, you know, sort of physics level
1:29:24 problems to to taste what's here. Like
1:29:28 there's a lot of scientists right now
1:29:30 using chat GPT5 going, "Hey, we're
1:29:32 making novel breakthroughs."
1:29:34 because they know how to they know what
1:29:36 they're doing and what they're asking
1:29:38 for and these things are making a
1:29:39 difference. I think 2026 these things
1:29:42 are going to get quite sophisticated,
1:29:45 quite autonomous and anything
1:29:49 where we've got shortcomings right now
1:29:51 those are going to start disappearing in
1:29:53 2026. So I think 2026 things are going
1:29:56 to start getting weird. Um, I don't know
1:29:58 what it looks like, but
1:30:00 keep coming back here because you should
1:30:03 be in community. You should be hanging
1:30:05 out with people that are trying to
1:30:06 figure this [ __ ] out. That's what this
1:30:08 channel's about. All right, so we'll
1:30:10 take care of you and uh, go join the AI
1:30:13 salon, the salon.ai.
1:30:15 All right, peace. Bye.
1:30:18 [Music]