// transcript — 1793 segments
0:00 Introduction to Marc Andreessen
0:02 If we didn't have AI, we'd be in a panic right now about what's going to happen
0:04 to the economy. We've actually been in a regime for 50 years of very slow
0:08 technological change in the face of declining population growth. The timing
0:11 has worked out miraculously well. We're going to have AI and robots precisely
0:13 when we actually need them. The remaining human workers are going to be
0:16 at a premium, not at a discount. >> How big of a deal is the moment in time
0:21 that we are living through right now? >> This is a very, very historic time. AI
0:25 is the philosopher stone. Now we have a technology that transfers the most
0:28 common thing in the world which is sand converted into the most rare thing in
0:30 the world which is thought. >> We spent a lot of time with the most
0:34 cutting edge AI forward founders. The most leading edge founders are thinking
0:37 of can you have entire companies where the founder does everything.
0:40 >> There's all this concern that young people jobs are not going to be there
0:43 for them. AI is replacing them. >> Everybody wants to talk about job loss
0:46 but really what you want to look at is task loss. The job persists longer than
0:49 the individual tasks. >> What's your sense of just the future of
0:52 three very specific roles? Product manager, engineer, designer. There's
0:55 like a Mexican standoff happening between those three roles. Every coder
0:59 now believes they can also be a product manager and a designer because they have
1:02 AI. Every product manager thinks they can be a coder and a designer. And then
1:05 every designer knows they can be a product manager and a coder. They're
1:07 actually all kind of correct. What happens is the additive effect of being
1:11 good at two things is more than double. The additive effect of being good at
1:14 three things is more than triple. You become a super relevant specialist in
1:18 the combination of the domains. >> People aren't fully grasping how much
1:21 this changing. And people who really want to improve themselves and develop
1:23 their careers should be spending every spare hour in my view at this point
1:26 talking to AI being like, "All right, train me up." Today my guest is Mark Andre, one of the
1:34 most seinal figures in tech and in business. He invented the web browser,
1:39 built the world's largest venture firm. He's also a multi-time founder and an
1:42 investor in essentially every generational tech company and is also
1:46 one of the most clear-minded, lateral, and insightful thinkers about both the
1:51 past and the future of technology. In this very special conversation, we chat
1:55 about how unique and significant the moment that we are all living through
1:59 right now is, what skills he's teaching his kids to thrive in the AI future,
2:04 what happens to product managers, designers, and engineers in the coming
2:09 years. where moes exist in AI, what the most AI native founders are doing
2:13 differently, and so much more that is just scratching the surface of this very
2:17 deep and important conversation. You are going to walk away from this chat being
2:21 smarter about what is going on in the world right now and where things are
2:24 heading. A huge thank you to my newsletter community and focus on X for
2:28 suggesting topics and questions for this conversation. If you enjoy this podcast,
2:31 don't forget to subscribe and follow it in your favorite podcasting app or
2:35 YouTube. It helps tremendously. And if you become an insider subscriber of my
2:40 newsletter, you get a year free of over 20 incredible products, including a year
2:45 free of lovable, replet, bold, gamma, nad, linear, superhuman, devon, post
2:49 hog, dscript, whisperflow, perplexity, warp, granola, magic patterns, raycast,
2:53 chappd, mob, and stripe atlas. Head on over to lenny'snewsletter.com and click
2:57 product pass. With that, I bring you Mark Andre after a short word from our
3:02 sponsors. Today's episode is brought to you by DX, the developer intelligence
3:05 platform designed by leading researchers. To thrive in the AI era,
3:10 organizations need to adapt quickly. But many organization leaders struggle to
3:14 answer pressing questions like which tools are working? How are they being
3:19 used? What's actually driving value? DX provides the data and insights that
3:23 leaders need to navigate this shift. With DX, companies like Dropbox,
3:28 Booking.com, Adion, and Intercom get a deep understanding of how AI is
3:32 providing value to their developers and what impact AI is having on engineering
3:36 productivity. To learn more, visit DX's website at getd dx.com/lenny.
3:46 If you're a founder, the hardest part of starting a company isn't having the
3:50 idea. It's scaling the business without getting buried in back office work.
3:54 That's where B comes in. Brex is the intelligent finance platform for
3:58 founders. With Brex, you get high limit corporate cards, easy banking, high
4:03 yield treasury, plus a team of AI agents that handle manual finance tasks for
4:07 you. They'll do all the stuff that you don't want to do, like file your
4:12 expenses, scour transactions for waste, and run reports, all according to your
4:17 rules. With Brexit AI agents, you can move faster while staying in full
4:22 control. One in three startups in the United States already runs on Brex. You
4:27 The historic moment we’re living in
4:34 Mark Andre, thank you so much for being here and welcome to the podcast.
4:38 >> Awesome, Lenny. Thank Thank you. It's great to be here.
4:40 >> I want to start with just a big picture question. I have a billion directions I
4:42 want to go, but I think this is going to give us a little bit of a frame of
4:47 reference. How big of a deal is the moment in time that we are living
4:50 through right now? >> This is a very very historic time. I
4:54 think 2025 was maybe the most interesting year in my entire career and
4:58 and probably life and I think I would expect 2026 to exceed that.
5:01 >> Wow, that says a lot. >> Yeah, I've se I've seen some stuff. So,
5:04 um it feels like two things are happening. one is the the the trust that
5:09 a lot of people have had in kind of what you describe as kind of legacy
5:12 institutions around the world is I I think in kind of full scale collapse
5:15 right now. By the way, there's a lot of data data to support that. And so I
5:18 think there's just there's there's like a lot of structures and orders and uh
5:23 institutions that people have just relied on for a long time that have just
5:26 proven to not be up for the up for the challenge. And then kind of
5:29 corresponding with that is the national and global conversation have become like
5:34 let's say liberated. Um, and so, you know, this sort of incredible revolution
5:37 that we have in in kind of, uh, you know, what I've described as freedom of
5:41 speech, freedom of thought, um, ability for people to openly discuss things that
5:44 maybe they couldn't discuss even a few years ago, you know, is just
5:47 dramatically expanded. And I think that's that's now on on a one-way train
5:50 for just a much broader range of discourse. And then, you know, there's
5:54 also just these like incredibly massive geopolitical shifts that are happening.
5:57 And obviously, the the US is changing a lot, Europe is changing a lot, China is
6:01 changing a lot, Latin America, by the way, is changing a lot. very dramatic,
6:03 you know, events playing out down there right now, you know, kind of all over
6:06 the world. Like I think a lot of assumptions are being pulled out in the
6:10 into the daylight and and re-examined. And and then it's kind of the fact that
6:12 all these things are happening at the same time, right? And so you've got all
6:17 of these countries and industries, you know, where things are kind of
6:20 increasingly upheaval, but you have AI is this kind of new technology that's
6:22 going to really affect things. And then you've got, you know, people, you know,
6:25 citizens being able to fully participate, uh, and being able to argue
6:28 things out. So, it's it's kind of like those three kind of big mega things are
6:31 kind of all colliding um at the same time. And I I think we're probably just
6:34 the very beginning of all three of those. And those all feel like kind of,
6:37 you know, historical, you know, moment shifts, you know, comparable in
6:42 magnitude to maybe the fall of the Berlin Wall in 1989, you know, maybe
6:46 maybe the end of World War II. Um you know, kind of moments like that. It
6:48 certainly feels like that. >> Good God, what a time to be alive.
6:52 The impact of AI on society
6:55 >> Yeah. In terms of the AI piece, which is where a lot of people are trying to
6:57 figure out what to do, what do you think isn't being priced in yet in terms of
7:01 the impact AI is going to have on say the world or just people listening?
7:05 >> The I think at I think at this point I think it's pretty clear with it with you
7:08 know our technology hats on that like this stuff is really working now right
7:11 and so there there was this you know kind of you know when when there was a
7:14 chat GPD moment you know three years ago it was only by the way only three years
7:18 ago right? um was the chat GPD moment and and the big question was all right
7:21 this this is like incredibly fun and creative and like we have machines now
7:24 that can compose Shakespeare and silence and rap lyrics and like you know this is
7:28 amazing but then there was there you know there's this big question like can
7:30 you can you harness this technology for reasoning um and for you know problem
7:34 solving in domains that like really matter you know medicine and science and
7:39 and and law and so forth um and and you know it turns out the answer to that is
7:42 yes right um and you know the the last 12 months and especially the last even
7:45 just the last three months have really proven that like AI can really do like
7:48 you know you're seeing it all now you know you can actually you know AI is now
7:52 developing new math theorems um you know there you know over the holiday break
7:55 you know there's sort of the what it feels like the AI coding thing you know
7:59 really hit critical mass uh and the world's best you the world's best
8:02 programmers right including like Lisbald's you know for for the first
8:05 time over the holiday break basically said yeah AI is now coding better than
8:08 we can and so that you know that's that's incredibly incredibly powerful
8:12 and I think we we all you know kind of I think assume that AI now is going to get
8:15 really good at reasoning um in in any domain do in which there are verifiable
8:19 answers and so that that you know that's going to include like many very
8:23 important domains. So um so like for the technology feels like it's it's it's
8:26 moving fast and and it's going to be working really well. Um I think the
8:30 thing that is not well understood I I think a lot of people have a I think you
8:34 know a lot of people in the industry have kind of what I would describe as
8:36 this one-dimensional thing which is okay as a result of the technology not
8:39 working AI just kind of sweeps sweeps the world and changes everything. And I
8:43 think that's that's kind of the wrong that's kind of the wrong frame. I think
8:45 it's based on an incomplete understanding of of the world that we
8:47 live in or the world that we've been living in for the last you know 80 years
8:52 and I would call out two things in particular. So one is it has I think
8:57 it's felt to us like in the US and the west for the last you know whatever 30
9:00 years or 50 years it's felt like we've been in a time of great technological
9:04 change but actually if you look for actually evidence of that like in stat
9:07 in statistical evidence of that analytical evidence of that like you
9:11 basically can't find it. Um and in particular um economists have a way of
9:14 measuring the rate of technological change in the economy that is
9:17 productivity growth which which we could talk about what that means but basically
9:20 it's it's a it's sort of the mathematical expression of the impact of
9:24 technology uh on the economy and productivity growth for the last 50
9:28 years has actually been very low not very high so we all feel like it's been
9:31 very high there's been lots of technological change what's actually
9:33 happening is it's it's been very low and in fact the pace of productivity growth
9:39 like in the US is is running at like a half of what it in my lifetime, in our
9:43 lifetimes, it's been running at about a half the pace um that it ran in um
9:48 between 1940 and 1970. And it's been running at about a third the pace that
9:53 it ran between about 1870 to about 1940. And so statistically in the US in the
9:58 west technology progress in the economy, technology impact the economy has
10:02 actually slowed way down. And so we, you know, the AI thing is is going to hit,
10:05 but it's hitting an environment in which we, we have actually had almost no
10:08 technological progress in the actual economy for a very long time. So we
10:11 could talk about that. And then there's this other like just incredible thing
10:14 that's happening, which is the the, you know, s the de demographic collapse,
10:18 right? It's sort of a western phenomenon, an increasingly global
10:21 phenomenon, which is, you know, the rate of reproduction of the human species is
10:25 is in rapid decline. And you know there are many countries you know including
10:28 the US where you know the rate of reproduction is you know under two you
10:32 know meaning meaning that you know many many countries around the world by the
10:35 way including China which is a really big deal are actually going to
10:39 depopulate over the next century um and so you have this kind of precondition
10:42 that says there's actually been very little techn technological progress
10:45 happening in the world um and the world is going to depopulate um and so AI is
10:50 going to enter the world a world in which those two things are true and I
10:53 think it's inc this is incredibly important because we actually need AI to
10:57 work in order to get productivity growth up, which is what we need to get
10:59 economic growth up. And we actually need AI to work because we're going to need,
11:02 you know, we're going to need machines to do all the jobs that we're not going
11:05 to have people to do because we're we're literally going to depopulate we're
11:07 going to depopulate the planet over the next hundred years. And so I I think the
11:11 interplay of these factors is is going to be much more interesting and and
11:14 AI’s role in education and parenting
11:14 frankly more more more complex than a lot of people have been thinking.
11:16 >> I'm going to follow this thread about kids. I know you have a kid and one of
11:20 my most my favorite lenses into how people think and what they value is what
11:24 they're teaching their kids, what they're steering their kids towards.
11:28 >> Are there specific skills or I don't even careers that you're steering your
11:31 kid towards? >> The way I think about this and you know,
11:35 yeah, we we have a 10-year-old and so, you know, we and we actually homeschool
11:38 and so we we we think a lot about this. Um so I think the way to think about the
11:43 impact of AI on on people on specifically people as individuals I
11:47 think it's it's it's actually you know a lot of people just focus on kind of this
11:50 you know this kind of very I would say straightforward and or overly simplistic
11:54 view of just literally job gains you know job losses which we could talk
11:57 about but there's two specific things at the level of like an individual person
12:00 and individual kid so I think it's pretty clear that AI is going to take
12:05 people who are good at doing things and it's going to make them very good at
12:08 doing things right and so It's going to be a tool that's going to sort of raise
12:11 the average kind of across the board. And you know, look, you see that playing
12:14 out already. You know, anybody who's in a position where they need to, you know,
12:16 write something or design something or write code or whatever, if they're if
12:19 they're pretty good at it today, they use they use AI and all of a sudden
12:22 they're very good at it. And so there there's sort of that aspect to it. And I
12:25 think the the the way the education system very large is going to teach is
12:28 going to kind of teach AI is is going to be based, you know, hopefully a lot on
12:32 that. But then there's this other thing that's happening which we're also
12:35 starting to see and we're really seeing it particularly in coding right now. Um
12:40 where the really great people are becoming like spectacularly great,
12:46 right? Um and so you just you kind of use it use the term you think about like
12:50 the supermpowered individual, right? So the individual who is like really good
12:54 um at coding or really good at making movies or really good at making songs or
12:59 really good at designing you know making art or whatever whatever those things
13:03 are or or you know or podcasting or you know hopefully venture capital you know
13:06 if if you're very good at it and you can really harness AI you can become
13:11 spectacularly great uh and like super productive right and you know I'm sure
13:14 you have a lot of friends in this in this category as well but like you know
13:18 the the really really good coders are experiencing this right you know, my
13:20 friends who are really good coders are like, "Oh my god, all of a sudden I'm
13:23 not twice as good as I used to be. I'm like 10 times as good as I used to be."
13:28 And so I think at the at the unit of like n equals one of like an individual
13:32 kid, I think the question is kind of how do you get them in a position where
13:34 they're kind of this kind of supermpowered individual such that
13:38 they're going to be really kind of deep in whatever it is they're going to do,
13:40 but they're going to they're going to be deep in a way that's going to let them
13:43 fully use the power of AI to be not just great, but to be like spectacularly
13:47 great. Um, and and I think that that that's that's going to be the real, you
13:50 know, that that that that that's the real opportunity and that, you know, at
13:52 least that's what we're shooting for and that's what I would encourage parents to
13:53 shoot for. >> So, what I heard there is essentially agency, this word that we see on Twitter
13:58 all the time is building uh agency, them not waiting for someone to tell them
14:01 what to do, figuring out what to do. >> Yeah. Yeah. So, this this this thing
14:04 with this this term agency that's become very very um you know, very popular um
14:09 certainly California for the last couple years. It's really interesting because
14:11 it's it's I had a lot of trouble with this early on because I'm like agency.
14:14 What are they talking about? And what what they're kind of talking about is
14:17 like, you know, initiative, you know, um you know, willingness to,
14:22 you know, you could just do things. Um you know, uh what is it? Uh the the demo
14:27 bird has the great term live player. Um you know, you you you can be like a
14:30 primary participant in events. And at first I was like, well, yeah, like
14:34 that's kind of obvious, right? like of course and and then I'm like oh actually
14:39 it's not so obvious anymore because kind of your your point I think so much of
14:42 our society is based on like there are all these rules and everybody gets
14:46 taught kind of by default you're supposed to follow all these rules right
14:50 and then everybody if you like break the rules like everybody gets freaked out
14:52 it's like oh my god he broke the rules and so like we we we have somehow worked
14:56 our our way our way kind of you know I don't know psychologically
14:59 sociologically you know kind of into a state in which I guess the natural
15:02 assumption for a lot of people is you know the thing that you for example you
15:05 want to train kids to do is like follow all the rules. Um, and you know, you
15:08 could argue that kind of you know, for example, the you know, the school
15:10 system, the K through2 school system or whatever has gotten kind of more and
15:13 more focused on that over time. And it's like yeah, it's like no, you you should
15:16 actually and again, especially at unit unit n equals one, like of your kid.
15:20 It's like and look, there's there's something to be had. We I just had this
15:23 conversation my 10-year-old last night actually. I I I rolled out uh uh the
15:28 concept of uh you know, in order to lead, you must first learn to obey,
15:31 right? In order to you know, issue orders, you must learn how to follow
15:35 orders. and you know you kind of try to keep keep him with some level of
15:39 structure in his life and not just and not just pure agency but yeah I mean and
15:42 so look you know some rules are important and so forth but yeah no look
15:45 there there is like a huge b there's just a huge premium in life on being
15:48 somebody who is able to like fully take responsibility for things fully take
15:52 charge run an organization lead a project create something new um and you
15:58 know maybe yeah that that has been maybe a little bit diminished in our culture
16:02 over the last 30 years it you know it's it's healthy you know that that you know
16:05 that that there's now a term for that that that is coming back back into vogue
16:09 and then and then and again that's how I view AI for kids is like okay AI should
16:13 be the ultimate letter on the world for a kid with agency to be able to say okay
16:17 I can actually be a primary contributor right whether that's I can be a primary
16:20 contributor in everything from you know developing new areas of physics to
16:24 writing code to being an artist uh you know to writing you know to writing
16:27 novels like you know whatever that thing is I I can fully participate in the
16:30 world I can really change things and I and I that that feel that The
16:33 combination of that idea combined with this technology feels very healthy to
16:35 me. >> What is that quote about? Give me a lever and I'll move the world.
16:39 >> And I'll move the world. Yeah, that's exactly right. Well, so it's actually
16:42 funny you mentioned that. So the the um the uh the the early kind of scientists
16:46 including like Isaac Newton were super obsessed with with you know this concept
16:50 of alchemy, right? It's like you know they you know they you know they
16:52 developed like you know Newton he's like developed Newtonian physics and he
16:54 developed like calculus and all these things but the thing he was really
16:57 obsessed with was alchemy which was the thing he could never get to work right
17:00 and and and alchemy was the transmutation of lead into gold which
17:04 meant the transmutation of something that was very common which was lead into
17:07 something that was very rare and valuable which was gold. And you know
17:10 they there was this the he spent you know decades trying to figure out this
17:13 thing called the philosopher stone which would be basically the the machine or
17:16 the process that would would be able to transmute the rare you know the common
17:20 thing into the rare thing led into gold and he never figured it out and you know
17:22 it's incredibly frustrating nobody ever figured that out and now we literally
17:27 with AI have a technology that transfers >> just blew my mind
17:34 >> right the the most common thing in the world which is sand converted into the
17:38 most rare thing in the world which is Right. And and so AI is it is it is the
17:43 it is the philosopher stone. Like it it is that it it actually is that and it's
17:47 just this incredibly powerful tool. Um and and that's where I that's where I
17:49 get so excited. I mean and again this is what we're doing with our 10-year-old
17:51 which is like all right a primary thing that we want to make sure to to do is to
17:55 make sure that he knows fully how to leverage and and get and get benefit out
17:59 of the philosopher stone, right? Which is uh you know which is to say AI and
18:02 that that and then you know that's certainly central to everything we're
18:04 teaching him. you know, there's there's this meme going around that um you know,
18:07 Silicon Valley people don't let their kids use computers. And I I just I I
18:10 there may be a handful of people who are like that. I I don't you know, I don't
18:14 know. Um I I think it's more honestly the other way around, which is uh the
18:17 you know, the more you're kind of plugged into stuff in Silicon Valley,
18:19 the more important it is to make sure that your kids actually fully understand
18:22 this and know how to use it. And that's certainly the mode that we're in. And
18:25 that's that's certainly the mode that I would encourage parents to think about.
18:27 >> I did not know your kid was homeschooled. That is super interesting.
18:30 There it's almost a statement on, you know, education in today's day. Maybe is
18:35 there any thoughts there? I'm just for folks that maybe aren't in your tax
18:39 bracket that want to help their kids be successful, maybe homeschooled, maybe
18:42 not. What what advice would you have? >> This is the challenge and again this
18:45 this kind of goes to how you're you know kind of your original question which is
18:50 education there's two completely different ways to talk about think about
18:54 education. The way that's usually thought about and talked about is kind
18:57 of at the level of like a nation, right? So, so you know, it's like a national
19:01 level issue or maybe a state level issue in the US, which is basically like how
19:04 do you educate all the kids? And of course, that's incredibly important. And
19:06 of course, you're going to need like some level of large scale system like
19:09 the, you know, the national K- through2 school system or something like that,
19:13 you know, in order in order to do that. Um, but then there's this other question
19:17 which is like at n equals 1 for an individual kid like what can you do with
19:22 with an individual kid? Um, and so I'll just give you kind of the ultimate, you
19:24 know, kind of the ultimate answer to that question, which is it's been known
19:29 for centuries that the ideal way to teach a kid at the unit of n equals 1,
19:34 by far the ideal way to do it is is with one-on-one tutoring. Like if you just
19:38 have an individual kid and the goal is to maximize an individual kid, by far
19:41 you get the best results with one-on-one tutoring. And and this is something that
19:46 like every royal family knew in history. It's something that every aristocratic
19:49 class knew in history. There's all these amazing examples. Alexander the Great
19:53 was tutored by Aristotle. He took over the world, right? Like, you know, many
19:57 of the great kings and queens and you know, royal families and aristocrats and
20:00 so forth, you know, over the course of centuries. Um, you know, kind of always
20:04 had always had this approach. There's actually also statistical evidence, um,
20:07 analytical evidence that this is correct. Um, there there's this, you
20:11 know, massive question in the field of education, which is how do you improve
20:14 educational outcomes? And basically it turns out it's just it's very hard to
20:16 improve educational outcomes except there's one method that always does it
20:19 which is called the it's called the bloom two sigma effect which is there's
20:22 one method of education that routinely raises student outcomes by two standards
20:26 of deviation and will take a kid from the 50th percentile to the 99th
20:30 percentile and that's oneonone tutoring right so again if you go back to like at
20:34 n equals one you have a kid and a tutor and they're in this like you know very
20:37 tight loop with each other you know where the kid is able to constantly kind
20:40 of be on the leading edge of what they're capable of doing and they can
20:42 they you know they they can move incredibly past and they get kind of
20:45 correction in real time, you get these better outcomes. But, you know, to your
20:47 question, like it's never been economically feasible for anybody other
20:51 than the richest people in society to be able to provide one-on-one tutoring for
20:55 kids. AI provides the very real prospect of being able to do that, right? Because
20:58 obviously now, right, if you have a kid that's like super interested in
21:01 something and they can talk to, you know, an LLM about it and they can ask
21:05 an infinite number of questions and they can get instantaneous feedback. Um, and
21:09 in fact, you can even tell an LLM it's like, you know, teach me how to do the
21:11 following. And you can say, you know, wow, that's like I don't quite
21:13 understand what you're saying. Like, dumb it down for me a little bit. Um,
21:17 okay, now quiz me, you know, do I actually understand this? Like, people
21:21 can just do this today, right? Um, and so I I think there's this like massive
21:25 opportunity for for parents, you know, in in many walks of life to be, you
21:28 know, with with with a little bit of time and focus, uh, to be able to say,
21:31 okay, you know, my my kid's probably still going to go through a traditional
21:33 education system, but I'm going to augment this with AI tutoring. Um, and
21:37 of course there, you know, and of course there's going to be tons of startups,
21:39 right? And there already are that that are going to try to build on all the all
21:42 the products and services for this. Khan Academy, you know, on the nonprofit side
21:46 has a big push to do this. Um, and so, you know, I think the the broad answer
21:49 might be a hybrid approach with schools plus onetoone tutoring through AI. Um,
21:53 there's also this great, you may have heard there's this great school new
21:56 private school system called Alpha, um, in which everything I just described is
21:59 kind of the basis of their philosophy, which is, you know, it's a combination
22:02 of in-person schools and teachers, but it's also, you know, heavily based on AI
22:05 and AI tutoring. And so I I think there's like a there is a magic formula
22:10 in here um that I think is going to apply much more broadly. Um and I and it
22:13 really for parents interested in this I now would be a great time to really
22:15 The future of jobs in an AI-driven world
22:16 start to think hard about that um and and to look at the options.
22:19 >> It's interesting because there's all this concern that young people uh jobs
22:23 are not going to be there for them. AI is replacing them. On the flip side,
22:26 there's what you're describing here. It feels like people coming into learning
22:29 today are going to be move so fast and learn so much more. And where where do you sit on this
22:34 divide of like young people are in big trouble or they're actually going to be
22:37 the ones winning in the end? >> Yeah. So the job the job substitution
22:40 job loss thing is just it's very reductive. It's it's I think it's an
22:43 overly simplistic model. And again it goes back to what I said at the very
22:46 beginning which is we've actually been in a regime for 50 years of very slow
22:50 technological change in the economy. And so you know again like I said it's like
22:53 at a half the rate of of the previous era and then a third the rate of like
22:57 100 years ago. And so we're we're coming out of this kind of phase where we've
23:00 had like almost no technological progress in the economy. We've had
23:03 remarkably little job turnurn as a result of that relative to to any
23:07 historical period. And so even if AI like ticks up, even if AI triples
23:10 productivity growth in the economy, which would like be a massively big
23:13 deal, it would take us back to the same level of job turnurn that was happening
23:18 between 1870 and 1930. And if you go back and you read accounts of 1870 to
23:21 1930, people just thought the world was a wash with opportunity. Right? at that
23:25 rate of technological transformation, kids were able to like develop new
23:29 careers into new areas of of of the economy, building new kinds of products
23:32 and services. I mean, you know, a huge part of our of everything in our modern
23:35 world today was kind of invented and uh and proliferated kind of during that
23:39 period. Um, and so even if AI like triples the pace of economic change in
23:42 the economy, it's going to just translate to like a much higher rate of
23:45 economic growth is going to transfer translate to a much higher rate higher
23:48 rate of job growth. And you know there there will be some level of like task
23:51 level and job level substitution that will take place but that will be swamped
23:55 by the macro effects of economic growth and innovation uh that will happen and
23:58 that then corresponding to that there will be you know there there will be
24:01 hiring blooms you know I quite honestly I think all over the place and then
24:05 again go back to the the other thing which is like this is all happening in
24:08 the face of declining population growth and and and increasingly population
24:13 shrinkage. Um and so human workers in many many many countries over the next
24:16 you know 10 20 30 years are going to be at more and more of a premium u
24:20 literally because you're going to have shrinking population levels. You know we
24:23 don't really want to get into you know politics particularly but it does feel
24:27 like the world broadly is going is is going to reverse course on on on the
24:29 rates of immigration that we've had for the last 50 years. it seems to be kind
24:32 of a broad-based you know kind of thing happening um you know kind of with you
24:36 know rise in nationalism you know concerns about the rate of immigration
24:38 and immigration historically in countries like the US you know it's it's
24:42 kind of eb and flowed over time based on kind of how you know kind of how the the
24:45 national mood shifts and so if you sort of combine in a country like the US or
24:49 any country in Europe if you combine declining population with less
24:53 immigration you the the remaining human workers are going to be at a premium not
24:57 at a discount um and so I think I think that combination of kind of faster
25:01 productivity growth, faster economic growth, and then slower population
25:04 growth and less immigration. Um, actually means there's going to be much
25:07 less of this kind of dystopian, you know, no jobs thing. I I just think it's
25:10 probably totally outpaced. >> That is extremely interesting. So, what
25:14 I'm hearing is you're not super worried about job loss. Is the key here that the
25:18 timing kind of just works out, this population decrease, you know, like all
25:21 these kind of have to line up for there not to be this massive job loss with AI?
25:26 >> Yeah. Well, look, if we didn't have AI, we'd be in a panic right now about
25:29 what's going to happen to the economy. Right? Because what we what we'd be
25:33 staring at is a future of depopulation and like depopulation without new
25:36 technology would just mean that the economy shrinks. Right? So so it would
25:40 mean that the economy kind of itself kind of shrinks over time. You know the
25:43 opportunity diminishes. There are no new there are no new jobs. There are no new
25:47 fields. There's no new there's no new source of consumer demand for spending
25:51 on things. Um and so you you would you would you would be very worried about
25:53 going into period of like severe decline of stagnation. Um, and you know, you
25:57 know, essentially you'd be looking at these like very dystopian scenarios of
26:00 like an economy kind of self- euthanizing itself uh over time. Um, and
26:05 and so you'd be very worried about like the opposite of what everybody, you
26:08 know, thinks that they're worried about. The only reason we're not worried about
26:11 that is because we now know that we have the technology that can substitute for
26:14 the lack of population growth and then, you know, also for the for the lack of
26:17 immigration that's likely. And so, you know, I would say the timing has worked
26:20 out miraculously well in the sense that we're going to have AI and robots
26:24 precisely when we actually need them, uh, to keep the economy from actually
26:27 shrinking. Um, and and I just think like that that's just like a a fundamentally
26:31 a fundamentally good news story. Um, to get to the mass job loss thing that
26:35 people are worried about, um, on the other side of things, you know, you have
26:38 to you'd have to look at like far far far higher rates of productivity growth.
26:42 you'd have to look at rates of productivity growth that are 10 20 30
26:45 50% a year, you know, something like that, which are, you know, orders of
26:48 magnitude higher than we've ever had in any in an economy in the history of the
26:52 planet. Um, you know, it's possible that we get that. I mean, look, I'm, you
26:55 know, I I have my utopian kind of, you know, kind of, you know, temptation
26:59 along with everybody else. If if AI like radically transforms everything
27:02 overnight, then maybe you, you know, let's let's play out the kind of utopian
27:06 scenario. Uh you get to a much higher level of of of productivity growth. You
27:09 get to a much higher level of technological change. corresponding to
27:13 that you'll have a massive economic boom. Uh you'll have a you know massive
27:16 growth in the economy and then corresponding with that you'll have a
27:21 collapse in prices. Um and so the price of goods and services that are that are
27:23 that are sort of you know whatever you want to call it affected by or
27:26 commoditized by AI the prices of those goods and services will collapse right
27:30 there'll be price deflation and then as a consequence of price deflation
27:32 everything that people are buying today gets a lot cheaper and that's the
27:35 equivalent of a gigantic increase in wealth right across the society right
27:40 think it this way this is actually worth talking about because people I think get
27:45 get kind of sideways on on this issue so if AI is going to transform the economy
27:49 as much as the you know whatever or utopians or dystopians or whatever kind
27:53 of think that it will. The necessary economic calculation of what happens is
27:57 massive massive productivity growth. The consequence of massive productivity
28:00 growth, what that literally means mechanically is more output requiring
28:04 less input, right? So you get more economic output for less input, right?
28:08 So you're substituting in AI for human workers or whatever. And as a
28:11 consequence, you get like this massive boom in output which with much lower
28:15 input costs. The result of that is you get lots of goods and services in all
28:19 those affected sectors. The result of those gluts is you get collapsing
28:23 prices, right? The collapsing prices mean that the thing today that cost you
28:27 $100 now cost you $10 and now cost you $1. That's the equivalent of giving
28:31 everybody a giant raise, right? Because now they have all this additional
28:34 spending power. That additional spending power then translates to economic
28:37 growth, right? The development of new fields. Everybody's like materially like
28:41 much better off very quickly. And then by the way, if you to the extent that
28:44 you do have unemployment coming out the other side of that, it's it's now much
28:47 cheaper to provide the kind of social safety net to prevent people from being
28:51 emiserated, right? Because the prices of all the goods and services that like a
28:54 welfare program has to pay from, they're all collapsing, right? And so the price
28:57 of healthcare collapses, the price of housing collapses, the price of
29:00 education collapses, the price of everything else collapses because this
29:04 this this this incredible impact that AI is having. And so in this kind of
29:07 utopian dystopian scenario that people have, it's not there there's no scenario
29:11 in which like everybody's just poor. In fact, it's it's quite the opposite,
29:13 which is everybody gets a lot richer because prices collapse and then it's
29:17 actually much easier to pay for the social safety net for the people who,
29:20 you know, for some reason can't find a job. And so, like like maybe we end up
29:25 in that scenario. I mean, the the kind of optimistic part of me says, yeah,
29:28 maybe AI is that powerful and maybe the rest of the economy can actually change
29:31 to to accommodate that and maybe that'll happen. But the result of that is going
29:33 to be a much better news story than people think it's going to be. Um, and
29:37 again, everything I've just described, by the way, is like just a very
29:39 straightforward extrapolation on very basic economics. I'm not making any like
29:43 bold predictions of what I just said. This is just like a straightforward
29:46 mechanical process that that that plays itself out if you have higher rates of
29:48 productivity growth, which are necessarily the results of higher grade
29:52 rates of technological growth. And so, I think we're I think we're looking at,
29:55 and to be clear, I think we're looking at a world that's not like radically
29:58 transformed the way that maybe the utopians think that it will be or the
30:01 the dystopians think it will be. I think it'll be more incremental for races we
30:04 can discuss. But I think that incremental is overwhelmingly I think
30:08 that process is going to be a good news process. And then even if it's much
30:10 faster, it's also going to be a good news process. It'll just be a good news
30:13 process in the other way that I described. >> I love hearing optimism and good news. I
30:15 Marc's past predictions
30:19 will also add that you've been I was researching you ahead of this chat and
30:23 you've been right so many times about where the world is heading. That's why
30:26 I'm especially excited to talk to you. I'll give you a short list. I imagine
30:29 there are many more things. Uh okay. Okay. So, one, you were right about the
30:34 web and web browsers becoming important. You were right about software eating the
30:38 world. Check. You uh in 2011, you said that in 10 years we're going to have 5 billion
30:45 people using smartphones. And I believe the actual number ended up being six
30:48 billion. You also you had this debate with Peter Teal that I came across where you were
30:54 debating whether technologies stop progressing or if new technology will
30:59 continue to emerge. and you were arguing there is progress. Progress will
31:02 continue. And he he was like, "No, I think we're done with cool technology."
31:06 You were right. Uh imagine there are many more things you were right about.
31:12 So, so again, I'm just I I love hearing your predictions because I feel like
31:14 they're actually going to turn out to be correct. >> So, I should start by saying I've been
31:18 wrong about tons of things, but you know, I buried those out back behind the
31:21 shed. >> Delete them from the internet. No web browser can discover them.
31:25 >> Yes, I have them nuked out of the internet archives so they they're never
31:29 seen again. Um, so, uh, you know, I'm wrong plenty of times also. Um, but
31:33 yeah, I mean, look, I think, yeah, some some of those I got right. By, by by the
31:36 way, I will say on the on the Peter one, I I have come I've come much more around
31:39 to Peter's point of view. >> Um, I would probably argue that one like
31:42 quite a bit differently today than I did, and I would give his view I think I
31:46 think a lot more credit. Um, and and it actually goes to kind of the discussion
31:49 that the kind of conversation we just had, which is the the real form of what
31:53 Peter was arguing was we have lots of process in bit. We have lots of progress
31:56 in bits, right? But we have we have very little progress in atoms, right? Um and
32:00 and that's the real core of what he was arguing. And I think I I I think I I was
32:03 a little bit I don't know missing that or kind of you know kind of glossing
32:06 that over a little bit um because I was so focused on making sure people
32:09 understood no there actually is still progress happening in in bits. But I
32:12 think you know a lot of his critiques around the lack of progress in Adams is
32:15 real and and again this goes back to this thing of like in the and he you
32:18 know he's talked about this for a long time. In the last 50 years there has
32:22 just been very little technological innovation in most of the economy.
32:24 there's been very little technological innovation in particular anything
32:27 involving atoms that you know there's been very little real world
32:30 technological change there just there just hasn't been like the the the built
32:33 world is just not that different today than it was 50 years ago and if you and
32:37 again if you contrast that you know if you if you compare and contrast 1870 to
32:41 1930 it was a dramatically different world if you contrast 1930 to 1970 it
32:45 was a dramatically different world if you contrast 1970 today it's not that
32:49 different right and look you just see that you could just like walk around and
32:52 it's just like oh yeah there's a bunch of buildings that were built built in
32:55 like 1960, right? And there's a bridge that was built in like 1930 and there's
32:59 a dam that was built in like 1910 and there's a city that was founded in, you
33:04 know, 1880 and like what have we done, right? Like where are
33:09 new cities? Where are new dams? Where, you know, where's where's the California
33:12 highspeed rail? Like you know, you know, like what's going on here? And so like I
33:17 think he is I I think he is right about a lot of that. Um, again, this is also
33:21 why I think that AI is not going to have as rapid an imp. It's not going to be
33:25 again this kind of utopian or dystopian view of like everything changes
33:29 overnight. I think it just kind of can't happen because of the reasons that Peter
33:32 articulates which is there's just there's so much about how the world
33:35 works that's basically just like wrapped up in red tape like bureaucratic
33:41 process, rules, restrictions. um you know the the the politics um by the way
33:47 you know unions cartels opolies there there's all these structures in the world that are kind of
33:52 economic or political or regulatory structures that basically prevent things
33:57 from changing and so I mean let's take let's take a great example like a AI's
34:01 impact on the healthare system like by rights AI is going to have a dramatic
34:05 impact on the healthare system and in and in in very positive ways but you
34:09 know large parts of the medical system today are they are cartels, right? And
34:13 so there's like a there's the doctors are a cartel and like nurses are a
34:16 cartel and like hospitals are a cartel and then there's this push to like
34:18 nationalize all the healthare systems and then you've got, you know, then
34:21 you've got a government monopoly, right? And it's like and and and guess what
34:25 cartels of monopolies don't like is they don't like like rapid change, right? Um
34:29 and so, you know, you show up as a kid and you're like, "Wow, I've got like
34:31 this new technology to do like AI medicine." And they're like, "Oh, well,
34:34 does it threaten Dr.'s jobs?" Well, in that case, we're going to we're going to
34:37 block it. So, and I think a lot of consumers, by the way, you know, I I I
34:40 see this in my life and you you'll probably see this in your life also,
34:43 which is, you know, like Chet GPT is like almost certainly a better doctor
34:46 than your doctor today, but like Chad GPT can't get a license to practice
34:50 medicine, right? So, it can't substitute for a doctor. It can't prescribe
34:53 medications, right? It can't, you know, perform procedures, right? And so there
34:57 there there are these any anyway so Peter Peter I think was very articulate
35:01 and has been for a long time on like no there are actually real structural
35:04 impediments in the economy and in the political system that we have that
35:08 actually prevent any the rates of change that are anywhere near the rates of
35:11 change that people had in the past. And and you can maybe say optimistically you
35:15 know maybe the presence of it of the new of the new magic technology of AI maybe
35:19 it causes us to revisit a lot of these assumpt assumptions for the first time
35:21 in decades to really say okay is this really the world we want to live in?
35:24 Don't we actually want to get to the future faster? So maybe that would be
35:26 the optimistic view. >> It's time to build. Somebody famously
35:30 said, I uh in my calendar, I actually have that as my when I start to work.
35:34 It's time to build. That's my block in the morning of the day. Thank you for
35:35 The Mexican standoff of tech roles
35:36 that. >> Okay. I love I love the way you go from just like macro to just like end of one.
35:42 And I want to go to end of one. A lot of the listeners of this podcast are
35:46 product managers. They're engineers. They're designers. They're not a lot of
35:49 There's a lot of founders, but there's also a lot of non-founders. There's a
35:52 lot of people building product that aren't founders and uh obviously a lot
35:56 of people are worried about where their career is going. Is one of these roles
35:59 going to disappear? Is one of these roles going to do really well? How do I
36:02 stay up to date? You're close with a lot of teams, a lot of product teams. What's
36:05 your sense of just the future of these three very specific roles? Product
36:09 manager, engineer, designer. >> This I think is a really funny question.
36:12 So these three roles in particular obviously are kind of the central roles
36:15 for for building you know for tech companies. So, the way I've been
36:17 describing it is, you know, you know the concept of the Mexican standoff, right?
36:21 Which is the the movie scene where the, you know, the two guys have guns
36:23 pointing at each other's heads. >> Um, and then there's, if you watch like
36:26 John Woo movies, he loves to have he does the three-way Mexican standoff
36:30 where you've got like a triangle, you know, people like, you know, and of
36:33 course it's John Woo movie, they've got, you know, guns in both hands.
36:37 >> So, they're all each each is aiming at the other two.
36:38 >> Yeah. >> Um, and you got this kind of standoff
36:41 situation. And so the way I've been describing this is there's like a
36:44 Mexican standoff happening between those three roles between product manager,
36:48 designer and coder. Specifically the following which is every coder now
36:51 believes they can also be a product manager and a designer right because
36:55 they have AI. Every product manager thinks they can be a coder and a
36:57 designer. And then every designer knows they can be a product manager, right?
37:01 And a and a coder, right? And so people in each of those roles now, you know,
37:06 know or believe that with AI they they don't need the other two roles anymore,
37:09 right? they they they can do that because they can have AI do that. And
37:11 then of course and then of course there's the real irony which is you know
37:14 all the the all three of them are going to realize that AI can also be a better
37:18 manager, right? So they're going to they're going to end up a aiming the
37:21 guns up the order chart. But that's probably that's the next phase. And what
37:26 I think is so fascinating about this Mexican staff is they're actually all
37:30 kind of correct I think right which is AI is actually a pretty good you know
37:33 it's now it's actually now a really good coder. it's actually now a really good
37:35 designer and it's also a really good product manager, right? It's actually
37:38 good at doing all three of those things or at least doing a lot of the tasks
37:41 involved in in in those three jobs. And so again, this this goes back to the the
37:45 the superower this kind of idea of the supermpowered individual. Uh where if if
37:49 I'm a coder like you know I mean step one is like I need to make sure that I
37:52 really understand AI coding and like what that means and what how coding is
37:55 going to change in the future. you know that that I need to you know
37:58 specifically how to go from being a coder who writes code entirely by hand
38:02 to being a coder who you know orchestrates you know a dozen instances
38:06 of of of you know coding bots you know you know there's there's a change in the
38:09 actual job of coding itself which is which is happening right now but the
38:12 other part of it is okay how do I become that superpowered individual how how do
38:15 I become a coder that also then harnesses AI so that I can also be a
38:18 great product manager and I I can also be a great designer right and then the
38:22 same thing for the product manager which is how do I make sure that I can now use
38:25 coding tools how do I make sure I can also, you know, do AI AI based design.
38:29 And the same thing for the designer, which is how do I use AI to be be also
38:32 become a coder and also become a product manager. And then what you get is maybe
38:36 the maybe the those individual roles change like maybe those are not anymore
38:40 sort of stovepipe roles the way that you know they have been for the last 30
38:43 years or whatever. Uh but what happens is the the talented people in any of
38:46 those roles become superpowered and they become good at doing all three of those
38:49 things. Um and then and then those people become incredibly valuable
38:52 because then those are people who can actually like you know build and design
38:55 right new products right from scratch which is like the you know which is
38:58 which is the most valuable thing. And so I I think I think that's I think I think
39:01 that's the opportunity. >> So I love this answer. So what I'm
39:05 hearing is essentially uh if you're amazing at any of these three roles you
39:08 will do well. >> Number one if you're amazing at these
39:11 roles that's great but also you part part of being amazing these roles is
39:14 also being being able to fully harness the new technology right. So if you're
39:19 if you're a master coder today and you you don't ever get to the point where
39:21 you you figure out how to use AI to leverage your coding skills, you and and
39:25 do more, right? Like at some point you are going to hit an issue, right? Here's
39:28 Adapting to changing job tasks
39:29 another way economists talk about this, which is there's the concept of the job,
39:34 but the job is not actually the atomic unit of what happens in the workplace.
39:37 The atomic unit of what happens in the workplace is the task. And so and and
39:40 then what what the way the economists think about it is a job is a bundle of
39:44 tasks. And everybody wants to talk about job loss, but really what you want to
39:49 look at is is task task loss, right? Tasks changing. I mean the the the the
39:54 classic the classic example of task changing. Classic example of task
39:58 changing was once upon a time executives never used typewriters or personal
40:02 computers themselves, right? You know, if you were a vice president of a
40:05 company in 1970 or whatever, you did not have like a typewriter or computer on
40:08 your desk typing things. You had a secretary who you dictated memos to,
40:11 right? And then there and then there was this change where like emails started to
40:14 show up. And what would happen was the job of the secretary then went from, you
40:17 know, it went from, you know, the the job of the secretary changed from
40:20 sending out letters with stamps on them to like sending or receiving emails with
40:24 the other admins. And then and then the secretary would print out the email and
40:26 bring it into the executive's office. And the executive office would read the
40:31 email and paper, scroll scroll the reply um and and and give and give that
40:34 message back to the secretary who would go back and type it into the computer on
40:37 on on his or her desk and send it as an email. Fast forward to today, none of
40:42 that happens. Now executives just do all their own email. They still have
40:46 secretaries or admins, but they're now doing different tasks. You know, they're
40:50 travel planning and orchestrating events and like doing all these other things,
40:53 you know, that that you know that the great admins do. And then and then the
40:57 task the task set ironically of the executive has expanded to do actually
41:01 more of the clerical work themselves actually like sit there and like type
41:04 their own memos, which again 50 years ago they never never would have done
41:07 that. And so the executive job still exists. the secretary job still exists u
41:11 but the tasks have changed and and I think that's like a great example of
41:14 what's going to happen in coding the tasks are going to change is what's
41:16 product management the tasks are going to change designer tasks are going to
41:21 change and so the the the job can p the job persists longer than the individual
41:26 tasks and then as the tasks change enough then that's when the jobs change
41:31 and so at the at the level of individual you kind of want to think of like okay I
41:34 have this job the job is a bundle of tasks I need to be really good at making
41:38 sure that I can like swap the tasks out, right? I can I can really adapt, use the
41:41 new technology, you know, get really good at AI coding, for example. I can,
41:44 you know, and then and then you want to kind of add skills. I can also get
41:47 really good at design. I can also get really good at product management
41:50 because I've got this new tool. So, you want to kind of pick up more and more
41:53 scope as you do that. And then, you know, 10 years from now, is your job
41:57 title coder or coder designer, product manager, or is it just I build products
42:02 or is it just I tell the AI how to build products? It's like whatever that
42:05 whatever that job is called, who even knows what it's going to be, but it's
42:07 going to be incredibly important because the people doing that job are going to
42:10 be orchestrating the AI. And so that that that's the track that the best
42:14 people are going to be on. Um and and I think that that's the thing to lean hard
42:15 The shift to scripting languages
42:17 lean hard into. >> I think people aren't fully grasping
42:20 just specifically software engineering and how much that is changing. Like it's
42:25 pretty clear we're going to be in a world soon where engineers are not
42:28 actually writing code, which I think a year ago we would not have thought. And
42:31 now it's just clearly this is where it's heading. It's like there's going to be
42:34 this artisal experience of sitting there writing code which is so crazy how much
42:39 that job is going to change. >> Yeah. So again here I go back and again
42:42 pardon maybe the history lesson but like I go back like coding. So the first you
42:48 may know that do you know the original definition of the of the term
42:50 calculator. Do you know what that referred to? >> No. >> It referred to people.
42:56 Right. So back before there were like electronic calculators or computers or
43:01 any of these things um the way that you would actually do computing the way that
43:04 you would do calculating like the way an insurance company would calculate
43:07 actuarial tables or the military would like calculate you know I don't know
43:10 whatever troop logistics you formulas or whatever it was the way that you would
43:13 do it is you would actually have a room full of people um and by the way these
43:16 like big rooms you could have hundreds or thousands or tens of thousands of
43:20 people doing this and you would actually you would actually figure out you have
43:22 somebody at the head of the room who was like responsible for like whatever the
43:26 mathematical equation was and then they would parcel out the individual
43:28 mathematical calculations to people sitting at desks who were doing them all
43:32 by hand, right? And and those that that job title was those people were
43:37 calculators, right? Um and so we've gone from a world in which you literally have
43:40 people doing mathematical equations by hands by hand. Then we got the first
43:44 computers. The first computers of course didn't have programming languages,
43:48 right? They they only had machine code, right? So the first computers were
43:51 programmed with ones and zeros. And so the task of the programmer became do the
43:55 ones and zeros and then that became punch cards and you can still you know
43:58 there's still people you know kicking you know today who you know whose job as
44:01 a programmer was to like deal with the punch cards and then you got actually
44:04 this big breakthrough which was called assembly language which was basically
44:08 the way to do machine code but like with some level of like English kind of added
44:12 to it and then the best programmers did assembly language and then you know when
44:15 I was coming up it was higher level languages like C that compiled into
44:18 machine code and that's what programmers did and then I still remember when when
44:22 scripting you know when scripting languages you know we developed
44:24 JavaScript at Netscape and then you know Python took off and Pearl and these
44:27 other scripting languages but scripting languages you know took off in the in
44:31 the in the in the in the 2000s there was this big fight in in the technical
44:34 community which is is scripting real programming or not right because it's
44:37 it's like it's kind of cheating right because real programmers write code that
44:41 compiles to machine code and like real programmers like do like memory
44:43 management themselves and they do all you know this this this whole craft of
44:46 writing writing uh you know writing writing C code and you know these these
44:50 The importance of understanding code
44:50 JavaScript or Python programmers are is doing this kind of lightweight thing and
44:52 does it even really count as as coding and of course the answer is yes it very
44:55 much counted and now most coding is done with the scripting languages right um
44:59 which have you see my point the scripting languages have abstracted away
45:03 like five layers of detail underneath that that people used to do by hand and
45:06 they don't anymore and then and then there's and then to your point like AI
45:09 coding is the next layer on that AI coding actually abstracts away the
45:12 process of actually writing the scripting code right and so in one sense
45:16 this is a really big deal for all the obvious reasons but on the other hand
45:19 it's like okay this is the next layer of the task redefinition under the job of
45:25 programmer right now what's the job of the programmer it's to your point it's
45:29 not necessarily to write the code by hand but what it is now is all right now
45:32 you know if you talk to the world's best programmers today what they'll tell you
45:35 is oh my job is I'm sitting there and I'm orchestrating 10 code bots right
45:39 coding bots that are running in parallel right and and literally they sit there
45:41 and they shift from browser you know browser to browser or terminal to
45:44 terminal and they're and they're they're watch their their day their day job now
45:48 is kind of arguing with the AI bots trying to get them to like write the
45:50 right code, right? And then and then debug it and and fix the problems and
45:53 change change the spec and and do all these things. And so now now the job of
45:56 the programmer is to argue with the coding bots, but like if you don't know
46:00 how to write the code yourself, you don't know how to evaluate what the
46:03 coding bots are giving you, right? And so, you know, you asked about the 10,
46:06 you know, our 10-year-old is, you know, super into computers and super into
46:09 programming. And what I'm what I'm tell, you know, he's he's using claude and
46:12 chat GPD and co-pilot and all these things. What I'm telling him is like,
46:15 look, and by the way, he lo coding. He's on Replet all the time doing vibe
46:18 coding, you know, doing g doing games, you know, he's sitting there, you know,
46:20 it's hysterical, right? Because he's sitting there, it's a 10-year-old
46:22 basically who's, you know, spends two hours at dinner arguing with an AI for
46:27 fun, right? Um, right. But but what I'm telling him is, no, look, you need to
46:31 still fully understand and learn how to write and understand code because the
46:35 coding bots are giving you code. If it doesn't work or if it's not doing what
46:37 you expect or it's not fast enough or whatever, like, you need to be able to
46:40 understand the results of what the AI is giving you, right? in in the same way
46:43 that somebody who's writing scripting language code does need to understand
46:46 ultimately how the microprocessor works. Um, and so again, it's it's kind of this
46:51 upleveling of capability where you actually want the depth to be able to go
46:54 down and be able to understand what the thing is actually doing even if you're
46:57 not spending your day actually doing that by hand. And again, I look at that
47:00 and I'm like, okay, now programmers are going to be 10 times or 100 times or a
47:03 thousand times more productive than they used to be, right? And and and that is
47:07 overwhelmingly a good thing. The the the the tasks are definitely changing. The
47:11 nature of the job is changing. Um, but are human beings going to be involved in
47:15 like in the coding process and overseeing the the AI coding and all
47:19 that? And and the answer is of course absolutely 100%. Like no question.
47:23 >> So you're in the camp of still learning to code, still a valuable skill.
47:26 >> Oh yeah, totally. Well, again, if you want to be one of these super Look,
47:29 look, if you just want to put your like self on autopilot and like I can't be
47:33 bothered and I'm just going to have AI write the code and it's going to
47:35 generate whatever it does and that's fine and I'm going to be, you know, I'm
47:38 going to be if if the goal is to be a mediocre coder, then just let the AI do
47:42 it. It's fine. The AI is going to be perfectly good at generating infinite
47:45 amounts of mediocre code. No problem. It's all good. If if if the goal is I
47:48 want to be one of the best software people in the world and I want to build
47:51 new software products and technologies that like really matter then yeah you
47:56 100% want to still be you want to go all the way down you want your skill set to
47:58 go all the way down to the assembly to assembly and machine code you want to
48:01 understand every layer of the stack you want to deeply understand what's
48:04 happening at the level of the chip right and and and the network and so forth by
48:08 the way you also really deeply want to understand how the AI itself works right
48:11 because you want to right because if people understand how the AI works are
48:15 able to they're clearly able to get more value out of it somebody doesn't
48:18 understand how it works, right? I mean, you're always more productive if you
48:20 know how the machine works, right? When you use the machine and so yeah, the the
48:24 supermpowered individual on the other end of this that wants to do great
48:27 things with the new technology, yes, you 100% want to understand this thing all
48:30 the way down the stack because you want to be able to understand what it's
48:33 giving you, right? And and and when something doesn't work or when something
48:35 isn't right, you want to be able to really quickly understand why that is.
48:39 Um, by the way, again, this goes back to education. AI is your best friend at
48:43 helping you learn all that, right? because it's like, oh, I need to
48:45 understand, I don't know, like this isn't fast enough. Um, I need to go I
48:49 need to figure out as a coder, I need to figure out how to do a different
48:51 approach to memory management or something. And you can be like, well,
48:54 you know, like I, you know, I don't quite know how to do that. Okay,
48:58 AI, let's spend 10 minutes. Teach me how to do this, right? Teach me what this
49:02 all means, right? So, all of a sudden, you have this like incredibly
49:04 synergistic relationship with the AI where it's also helping you get better
49:07 at the same time as doing a lot of work for you. >> By the way, I was going to say I was a
49:11 big Pearl uh programmer. I was an engineer for 10 years and that was my my
49:13 language of choice. >> You do you remember I don't know when
49:16 you were doing it but do do you remember at the at least early on do you remember
49:19 did you ever did you ever hit this where like coders were like looking down their
49:22 nose at you being like >> for sure for sure it's like this is so
49:25 slow it's not going to scale what are you what are you spending all your time
49:28 on this thing? Yeah, exactly. And of course, you know, and again, it was sort
49:30 of this thing where, you know, they were they were sort of correct, which is at
49:33 the beginning it wasn't, you know, fast enough or whatever. By the end, they
49:36 were definitely wrong, right? Which is it got much better, much faster, and you
49:40 know, it's it swept the world. U you know, most coding today happens as
49:43 scripting languages. And and then by the way, the people along along the way, the
49:46 people who really understood the scripting languages and the people who
49:49 understood all the lower level systems, they were the ones who were able to
49:51 actually make the scripting languages actually work really well, right? And so
49:54 that that was that was a great example of this kind of adaptation. And then and
49:57 then again the result of that was you know a far higher number of people
50:00 writing code with scripting languages than were ever writing code with lower
50:02 level languages. And I I think this will just kind of be a more dramatic version
50:05 of that. I love that Pearl was designed by a linguist. I don't know if you
50:08 remember that part and that's what made it so nice to to code with.
50:11 >> Well that's funny because of course it was so notorious for being impossible to
50:15 understand. So how ironic. >> Yeah. >> This episode is brought to you by Data
50:22 Dog, now home to EPO, the leading experimentation and feature flagging
50:25 platform. Product managers at the world's best companies use Data Dog, the
50:29 same platform their engineers rely on every day to connect product insights to
50:35 product issues like bugs, UX friction, and business impact. It starts with
50:39 product analytics where PMs can watch replays, review funnels, dive into
50:43 retention, and explore their growth metrics. Where other tools stop, Data
50:47 Dog goes even further. It helps you actually diagnose the impact of funnel
50:52 drop offs and bugs and UX friction. Once you know where to focus, experiments
50:56 proved what works. I saw this firsthand when I was at Airbnb, where our
50:59 experimentation platform was critical for analyzing what worked and where
51:03 things went wrong. And the same team that built experimentation at Airbnb
51:07 built EPO. Beta do then lets you go beyond the numbers with session replay.
51:12 Watch exactly how users interact with heat maps and scroll maps to truly
51:16 understand their behavior. And all of this is powered by feature flags that
51:20 are tied to real-time data so that you can roll out safely, target precisely,
51:26 and learn continuously. Data Dog is more than engineering metrics. It's where
51:30 great product teams learn faster, fix smarter, and ship with confidence.
51:36 Request a demo at dataq.com/lenny. That's data dogq.com/lenny.
51:37 The value of design in the AI era
51:41 Coming back to these this kind of triad, the other element that I hear more and
51:45 more of is just as is the skill of taste and design and user experience. It feels
51:49 like that's a very hard skill to learn and to me tells me design is going to be
51:54 much more valuable in the future. >> Yeah, that's right. And again here this
51:58 this is a great example. So the again the task level the the the the task
52:04 level of like design the perfect icon, right, is going to be like all right,
52:07 the A's going to do that all day long. is g give you a thousand icon designs.
52:09 It's going to be great. Like it's going to be fantastic. Like whatever, you
52:11 know, and there will still, by the way, there will still be some level of human
52:14 icon design or whatever, but like AI is going to get really good at that. But
52:19 like what are we trying to do? Like the, you know, kind of capital D design of
52:23 like, all right, what is this thing for? And how does this how is this going to
52:27 function in a world of human beings? And like, you know, what what's going to is
52:29 this going to make people happy when they use it? Is it's going to make
52:32 people feel good about themselves? Um, is it going to fit into the rest of
52:34 their life? Is it going to you know I don't know challenge them in the right
52:38 way? You know all these kinds of higher level questions that the great designers
52:40 have always thought about like that the the job of designer right will involve
52:44 much more of those higher level more important components and and then again
52:48 with with AI doing a lot more of the underlying tasks. And so, you know, one
52:51 way to think about it is, you know, I don't know, you you think of like, I
52:53 don't know, the world's best designers, you know, Johnny Ibe or whatever, you
52:56 could be like, "Wow, like if I'm a designer today, if I'm a 25-year-old
53:00 designer and I and I aspire to be, you know, Johnny Ibean in a decade, um, it's
53:05 it's all of a sudden I have a new path that I can use to kind of get to to get
53:08 there, which is I, you know, because Johnny I did everything he did without
53:12 AI." Now, you know, a young designer can be like, "Wow, if I really harness AI in
53:15 a decade, I'm going to be like the best designer the world's ever seen because
53:18 it's not just going to be me. it's going to be me plus being so super empowered
53:22 by this technology to be able to do so much more. Um, and then so much more of
53:26 my time and attention is going to be is going to be able to be focused on these
53:28 higher level things that most most designers never get to. And I think that
53:30 The T-shaped skill strategy
53:30 that's going to be another great example of that. >> So maybe what I'm hearing here is kind
53:35 of this T-shaped strategy of be if you want to be successful in any three of
53:39 these roles, be very very very good at that specific role, product management,
53:42 engineering, design, and then get good enough at these other two roles.
53:46 >> Well, so I think that's great. I think that's really really relevant. And then
53:48 you know the Scott you know Scott Adams unfortunately just passed away um you
53:52 know which which is a real tragedy but um I was always I I referred for years
53:57 to actually Scott's Scott Adams he had this famous um he had this famous kind
54:00 of career advice he would give people which I I think makes a lot of sense
54:03 which which doveetails with what you're saying which is he he used to say he
54:07 used to say it's like look he said um you know I I could he he said you know I
54:11 could have been a pretty good cartoonist um or I could have been like pretty good
54:15 at business but the fact that I was a cartoonist who understood business made
54:18 me like spectacularly great at making Dilbert, right? Because even the world's best
54:23 cartoonist who didn't understand business could have never written
54:25 Dilbert. And then the world's best business people who didn't know how to
54:27 do cartoons couldn't have done Dilbert. It took somebody who actually had both
54:30 of those skills to be able to make Dilbert, right? Which is one of the most
54:33 successful cartoons in history, right? And so so the the way Scott always
54:37 described it was that that the from a career development standpoint that the
54:40 additive effect of being good at two things is like more than double, right?
54:45 um the additive effect of being good at three things is more than triple, right?
54:50 Um because you you you be you become a super relevant specialist in the
54:54 combination of the domains. Um and and you you look you see this all I mean you
54:56 see this all over you know you see this all over the economy. Yeah. I mean you
54:59 see this all over the economy but I'll you know give you an example Hollywood
55:02 you know just Hollywood as an example you know there are a lot of writers who
55:05 can't direct a movie and they can be very successful writers. There are a lot
55:08 of directors who can't write a movie. They can be very successful directors.
55:11 But the superstars in the entertainment industry are the people who can write
55:15 and direct, right? And you know they they have a term for those. They call
55:17 those auras, right? And that's you know those are the people who are like the
55:19 real creative forces that move the field. And so and so again and by the
55:23 way Hollywood actually it's really funny spend been spending a lot of time
55:25 talking to Hollywood people about AI. Hollywood has the same Mexican standoff
55:29 going um right now that we that we described in tech except in Hollywood
55:32 for example for filmm it's the director it's the writer and the actor right
55:37 because the director is now thinking wow I don't need the writer anymore because
55:39 the AI can write the script and I don't need the actor anymore because I can
55:43 have AI actors the writer is saying I don't need the director because AI can
55:46 direct the movie and the AI can do the actors and the actor is saying I don't
55:49 need either one of these guys I can have the AI direct the thing I can have the
55:51 AI write the thing and I'm just going to show up and do my performance right and
55:55 so so it's it's it's the same it's the same kind of tri triangular
55:58 configuration. And again, what what's great about it is they're all correct,
56:02 right? Each person in each of those three fields is going to be able to
56:06 expand laterally and pick up those other those additional skills. And then as a
56:08 consequence, you're going to have more people who can write and direct or write
56:12 and act or direct and act or do all three. And and I think, you know, to
56:15 your point like your your your T-shift thing, like I I think that's going to be
56:18 true basically across the entire economy. And and and if you think about
56:21 the T is if you think about the T configuration, it's like yeah, the bre
56:25 the breath the breath the top of the tea is like how many individual domains are
56:29 you familiar enough with to be able to use the AI tools to be able to do really
56:33 good work. And then the the this part of the tea is how deep can you go in at
56:37 least one of those domains so that you really really deeply know what you're
56:40 doing. But like if you're like super deep on coding and you can use AI to do
56:44 design and you can use AI to do product management, right? That that's your T
56:47 right there. and and you're a triple threat at the top of the tea, but with
56:49 this level of technical grounding underneath that. And I mean, at that
56:52 point, you're again, you're the superpowered individual, you're going to
56:55 be able to just perform like feats of magic, uh, for example, in terms of
56:58 designing and building new products, you know, that people in my generation
57:01 couldn't have even dreamed of. And so I I I think I think that this is a
57:04 universal kind of theory that I think could can apply across the entire
57:06 economy. >> I'm going to invent a new framework right now. Okay, forget the T framework.
57:13 Uh, I'm picturing an F sideways or an E where there's three two or three
57:18 I don't know, downward parts. And so what I'm hearing is get good at least
57:20 two. >> I think that's right. I think that's right. Yeah. The combination. Yeah. Um
57:27 my my friend Larry Summers had a had a different version of the Scott Adams
57:29 thing which is he he used to tell people he said the key for career planning is
57:34 he said don't be funible, >> right? And you know that's he's an
57:37 economist and so that was economics speak and and what that means is what
57:40 that means essentially is don't be replaceable. And so don't be a cog.
57:43 Right? So, and what that meant was don't just be one thing, right? So, if you're
57:47 if you're if you're quote unquote, you know, again, just a designer, it's just
57:50 a product manager, just a coder, like then in theory, you can be swapped in or
57:53 out. But if if you have this if you have this Yeah. to if you have this E or F,
57:56 you know, laying on it side kind of thing. And if you have if you have this
57:59 combination of things that's actually quite rare, then all of a sudden you're
58:02 not fungeible. Not not only you're not funible, like you're actually massively
58:04 important because you're one of the only people in the world who can actually do
58:07 that combination of things. Um, and yeah, that that your ability now to
58:11 become one of those people is like just titanically enhanced uh with AI as
58:14 compared to anything we've ever seen before. >> This is so interesting because I've
58:18 worked with people that are good at these two skills and they were always
58:21 called unicorns at the company. She can code and design. Oh my god. And what I'm
58:25 hearing here is this is what you need to become. You need to become really good
58:28 at at least two things there. I think you use the term smoke stack or
58:31 something where it's like PM over here, engineer design. And what I'm hearing
58:34 here is you need to get good at at least two of these skills. the silos of these
58:37 two roles are disappearing. >> That's right. That's right. And again, I
58:40 can't I can't overstress the following for for anybody listening to this. The
58:44 thing about AI that I think people are just like not getting enough benefit out
58:49 of yet is just it will teach you. Like this is amazing. Like there's never
58:54 been a technology before where you can ask it like teach me how to do this
58:59 thing, right? So it's I always feel like it's like it's like people spend too
59:02 much. It's one of these things where it's like so much focus on figuring out
59:04 how to use like a large language model. is like, okay, what am I going to try to
59:07 get it to do for me? Right, which is of course very important, but the other
59:11 side of it is what can I get it to teach me how to do, right? And it's it's just
59:16 as good at that, right? Um, and so again, this is this level this level of
59:19 latent superpower like you know, people who really want to like improve
59:22 themselves and like develop their career should be spending every every spare
59:25 hour in my view at this point talking to an AI being like, "All right, train
59:27 train me up like tell me tell supermpower me, tell me how to, you
59:30 know, train me train me how to be, you know, I'm a coder. Train me how to be a
59:33 product manager." It will happily do that. It's it knows exactly how to do
59:36 that. You know, run me, you know, make me problems, you know, make me
59:40 assignments, then evaluate my results, right? And it will it will do that just
59:43 as happily as it will do work quote unquote for you. >> Two tricks I've heard along those lines.
59:48 One is uh to watch the output. What the agent is doing and thinking as it's
59:51 doing the work. So, if you're not an engineer, is just sit there and watch it
59:55 think and make decisions. And it's almost become this like layer on top of
59:59 learning to code is learning to see what the agent is doing and thinking because
60:02 that teaches you about architecture. And the other is uh a couple podcast guests
60:06 have mentioned this. When you get stuck and then you figure out how to unstuck
60:10 yourself, you ask it, "What could I have done differently? What could I have said
60:13 that would have avoided this error in the first place?" >> Yeah, that's right. That's right. Yeah.
60:17 Look, on that first one, and this again, this is what I'm doing with my
60:21 10-year-old. Yeah. Look, if if if you ask an Yeah, this is this is a really
60:23 good point. So, if you ask an AI, write me this code, and then and then it
60:25 doesn't and it comes back and it doesn't work right. Like if if all you know is
60:30 like single function I asked and it gave me back something that's not good like
60:32 what do you like what do you even do with that right like you don't
60:35 understand why it gave you that result do you really understand even what do
60:38 you even understand what to tell it to try to get it to do something different
60:41 but to your point like if you actually w if you actually watch what it's doing um
60:46 and and and then and then you you have the grounding you know kind of that leg
60:50 of the of your ear or your F um if you have that grounding then you can be like
60:54 oh I see what it's doing I see where it made the mistake I see where it went
60:57 sideways and then you're all of a sudden able to intervene and able to say no no
60:59 that's not what I meant do this other thing right and so and again this is
61:03 this this this is a big part of having having the actual kind of you know
61:06 synergistic relationship um is that you understand and by the way look I mean
61:09 this is like everything I'm saying is you know everything everything that
61:11 we're saying right now also is the same as if you're working with human beings
61:14 right like you know if you and I are colleagues and I you know ask you to do
61:17 something you'd come back with something completely different like I I do need to
61:20 understand what was happening in your head right in order to in order to be
61:24 able to get do need to give you give you feedback right if I just tell you oh
61:27 that's wrong it doesn't like nothing happens. I need to actually understand I
61:30 need to have theory of mind, right? I need to understand what you were
61:32 thinking in order to really give you the right feedback. Um and so and and you
61:36 know and again the great thing with AI is AI will happily sit there and explain
61:39 all day long why it's doing what it's doing. It'll you know it'll happily
61:44 critique itself. You know, you can do this. By the way, this has a very fun thing where you can
61:48 have have one AI critique the other AI, right? Um, which is another thing, which
61:51 is like you have one AI write the code, you have another AI debug the code, and
61:54 so you can actually use you can play the AIs off against each other and get them
61:57 to argue with each other. Um, and yeah, the these are all these are all the
61:59 kinds of skills that are going to become, I think, incredibly valuable.
62:03 >> I think people call those LLM councils. Yes. >> They're talking to each other.
1:02:05 AI’s impact on founders and companies
62:06 >> Yeah, that's right. That's right. >> I do feel like if I were like I'm I have
62:10 no design background. I've always wanted to design. I would I've always wanted to
62:13 be a great designer. Uh, it feels like that's the hardest one to learn of all
62:17 these three by just watching and talking, right? Because there's a lot of
62:20 exposure hours as as folks have used this term just like how do you learn to
62:23 be a great designer? That feels like that's going to be really hard and
62:25 valuable. >> So, my my true confession is I've always kind of wanted to be a cartoonist,
62:32 >> but I have no like art skills, but as we're talking, I'm like, it might
62:35 be time. >> Their time has come, Arc. >> Yes. >> I want to pivot to founders, your maybe
62:41 your bread and butter. You spent a lot of time with the most cutting edge AI
62:46 forward founders. I'm curious to what you see them do, how you see them, some
62:50 way they operate that's maybe blowing your mind about how the future of
62:54 starting a company looks, how the future of AI forward companies look.
62:59 >> Yeah. So, this is a great very, you know, topical topic that's all playing
63:02 out in real time right now um on on the leading edge. So, I I think there's like
63:06 three layers of it and see see if this makes sense. I think there's like three
63:09 layers of it. I think layer one is they're thinking all right how how does
63:13 AI redefine the products themselves right um and and this is kind of the
63:16 this is kind of the timehonored you know kind of thing that happens at technology
63:19 transitions and this is kind of what you know a lot of venture capital is based
63:23 on which is um you know okay there's a new technology that comes out and you
63:26 know maybe it's the personal computer or the iPhone or the internet or now it's
63:31 AI and it's like all right um is this a new capability that gets added to
63:35 existing products right so all of a sudden you've got I don't know an
63:39 existing you software business and now you got your you know PC version of it
63:41 and now you got your iPhone version of it and you just kind of keep on going
63:44 and you know you kind of add the new technology kind of gets kind of added
63:47 into the mix um you know it's kind of another ingredient into an existing
63:51 formula and and of course you know a lot of new technologies are like that right
63:53 um you know I don't know when I don't know when flash when flash storage came
63:57 out or something you know it didn't really you didn't really redefine the
64:00 the software industry because people just went from using you know hard disk
64:06 using flash storage or something um uh but when the internet came out like
64:09 basically old school onrem software for the most part you know not not entirely
64:13 but like a lot of it died and just got replaced by like web software um right
64:17 and so so sometimes you get the kind of it's additive to an existing thing
64:20 sometimes you get the actually it redefineses an entire product category
64:24 redefineses an industry the actual you know in many cases the companies
64:27 themselves turn over and so so so you know so there's sort of this question
64:29 and like you know an example you just mentioned nano banana so like a great
64:33 example is there you know there are these businesses like you know just take
64:36 Adobe like you know Photoshop is built a whatever a 40-year franchise in image
64:41 editing. Um, okay. Is AI a sort of a feature now that gets added to Photoshop
64:46 to be able to do AI based image editing or, you know, do you just like stop
64:50 editing images entirely because you're using Nano Banana and your all images
64:53 are just being generated and it's just easier to just have AI generate a new
64:57 image than it is to try to edit edit an old one. So I think you know there's
65:01 many areas of of tech in which that question is being asked and you know the
65:04 answers I think will vary by domain but u you know obviously as as a venture
65:07 firm we're batting hard on many of these categories being being totally
65:10 reinventsted and a lot of the a lot of the best founders are trying to figure
65:13 out how to do that. So that so that's kind of AI you know changing the
65:17 definition of the product. I think the next layer is actually a lot of what
65:20 we've already talked about which is AI changing the jobs. Um, and so it's, you
65:24 know, a lot of what we've already talked about, but like, okay, if I'm a founder
65:27 of a company and I've got, you know, if I have, you know, room in my budget for
65:31 100 coders, you know, how do I get those coders to be super empowered AI coders,
65:34 not, you know, not the kind of coders I used to have? And if they're super
65:37 empowered AI coders, then does that mean, you know, do I still need the
65:40 hundred? Maybe now I only need 10. Or does that mean I still want 100, but now
65:44 they're doing 10 times more, right? And so that, you know, as you know, like a
65:46 lot of the best founders are are working on that right now. And then I think the
65:51 third shoe to drop hasn't quite dropped yet, but it's it's you know it's kind of
65:54 the big one which is like all right like the the the the basic idea of having a
1:05:58 The concept of one-person billion-dollar companies
65:59 company right you know does that change and and again here you've got this
66:02 concept of the superpowered individual which is like okay um you know can you
66:07 have entire companies where you have basically the founder does everything
66:11 right because what the founder is doing is like overseeing an army of AI bots
66:14 and and there's sort of this you know there's kind of this holy grail in our
66:16 industry that's been running for a long time which is like can have the can you
66:19 have like the one person billion dollar outcome and you know we've had a few of
66:24 those over the years Bitcoin is probably the most spectacular example you know
66:28 with Ethereum right behind it um you know which wasn't quite one person but
66:31 you know a very small team you know you had you know kind of Instagram and
66:34 WhatsApp that had very big outcomes with very small teams you know every once in
66:37 a while you get one of these things where you just you know some something
66:40 hits and you just have a you know very small number of people associated with
66:43 it you know but that said you know most most software companies obviously end up
66:46 with you know huge numbers of employees um and So I I think you know some the
66:50 most leading edge founders are thinking of like okay how how do I reconstitute
66:55 the actual varied definition or idea um of a um of having a company and and you
66:59 know can you have a company that's that's literally basically just all AI
67:03 um and so and and if you're doing so you know if you're doing anything in the
67:04 real world that's hard but if you're doing software like that that that seems
67:08 like it might be feasible in some cases and then you know there's like the
67:10 ultimate example of that which is like you know can you have like AI can you
67:13 have like autonomous like AI economy stuff happening where you have like AI
67:16 bots on the blockchain or something you know that are out basically out there
67:19 like functioning as a as a as a business and like making money and just you know
67:23 literally where the the AI does all the work itself and just get you know issues
67:26 me dividends and so you know maybe that that you know maybe that maybe that's
67:29 the the final outlier result we have we have a few founders who are chasing that
67:32 kind of thing. Um so I would describe that as I would describe that as kind of
67:36 the the latter that the best founders around. >> Super interesting. this whole idea of a
67:40 oneperson billion-dollar company. I think it depends on your definition of
67:43 what this is like an outcome I could see. Uh having run running my newsletter
67:48 uh as one person with some contractors, there's so many little annoying things
67:51 that I have to deal with with just support tickets and issues and bugs and
67:55 like it's hard for me to imagine actually a oneperson billion-dollar
67:59 company even if AI is handling so much of your support because there's just so
68:02 many random edge cases that I'm just const like filling out forms. Uh and so
68:06 I guess depends on do you have contractors? Does that count? You know,
68:09 like what does it count? What does it mean to be a one person? But I'm just
68:12 like I can't see that happening. >> Yeah. I mean, look, Bitcoin's Satoshi
68:16 pulled it off. >> But like, you know, the open source community, you know, like does that
68:20 count? I don't know. I guess I guess guess it counts. Okay. >> Yeah. Exactly. Right. So, yeah, that
68:25 that Yeah. And I would say I don't propose to have answers here, but more
68:27 just like >> the smartest people I know are are many of the many of the smartest people I
1:08:33 Debating AI moats and market dynamics
68:33 know are are thinking hard about this. >> Yeah. What do you think about Moes? a
68:38 big question constantly in AI, you know, the fact that everything's changing.
68:42 Just what's your guys' thesis on Moes in AI? Does is that even a thing? Do you
68:45 care? >> My experience with like really big technological transformations, and of
68:49 course, I I kind of lived this directly with the internet, and I saw this
68:53 happen, is the really big technological transformations, they they take a long
68:56 time to play out, and there's there's all of these structural implications
69:00 that just kind of cascade out over time. And then there's kind of this this
69:04 there's this like rush to judgment up front where people kind of say, "Oh,
69:08 it's therefore obvious that you know XYZ. It's therefore obvious that this
69:11 kind of company is going to be the company of the future, not that kind.
69:14 It's obvious that this incumbent is going to be able to adapt and this other
69:17 one isn't. It's it's obvious that there's economic opportunity and this
69:20 kind of startup and not in these others. Um it's obvious that the moes are going
69:23 to be in this area of the technology but not in this other area. And and there
69:26 and you know what everybody does is they they kind of state those things with
69:29 like just an enormous amount of self assurance where they they you know where
69:32 they really sound like they have all the answers. And then you know what happens
69:35 is this these these ideas kind of saturate the media right because the the
69:39 media naturally prizes like definitive answers over open questions because you
69:43 know you want you know like when CNBC is like booking guests they want a guest
69:45 who's going to come on and say yes this is the way it's going to be X not like
69:48 you know I think that's a really good question and let's like debate it from
69:52 like eight different angles. And what I found is if you look back on those
69:54 predictions a few years later and you you can do this by the way if you pull
69:57 up like coverage of the internet from like 1993 through like 1997 or even
70:02 through like for that matter even through like 2005 or 2010 and you look
70:05 at like the kinds of confidence statements people were making in the
70:09 first 10 or 15 years like I would say like almost all of them were wrong again
70:15 generally like quite badly wrong and so I just I think the process I think with
70:19 massive with there's going to be a massive amount of technological change.
70:22 It's going to be like I don't know five or six layers of like structural change
70:27 that will play out over time and and again a lot we've talked about a lot of
70:30 this but like it the implications on like what are the definition of products
70:33 what are the definitions of companies what are the definitions of jobs what
70:36 are the definitions of industries how does this play out at the national level
70:39 how does this play out at the global level you know how does this inter by
70:41 the way how does this intersect with politics how does this intersect with
70:45 you know unions how does this intersect with you know war you know what's China
70:51 going to do um you know uh and So it's just like there's just there's there
70:54 just a tremendous number of unknowns like a very very large number of unknowns and I
70:59 think it's just like really really dangerous to prejudge these things and
71:03 so I'll just give I'll just give and it's just I'll just run this as a
71:05 thought experiment you know see what you think on this but it's like you know
71:11 like do do AI models the are AI models themselves like defensible like is there
71:16 a moat uh on AI models and on the on the one hand you'd be like wow it certainly
71:19 seems like there is or should be Because like if something takes you know
71:24 billions of dollars to build um and you need you know you need this like
71:26 incredible critical mass of like comput and data and there's only a certain
71:28 number of engineers in the world that know how to do this and you know they
71:32 are getting paid like NBA stars um and you know and then these companies have
71:35 to deal with all these like crazy you know political issues and press issues
71:39 and reputational stuff and regulatory and legal like all of that translates to
71:43 like you know okay probably at the end of this there's going to be two or three
71:45 companies that are going to end up with like you know 100% you know I don't know
71:51 whatever 5050 for 30 3030 or 90101 or whatever it is market share and then
71:53 they're going to have whatever profitability they have and it's going
71:56 to be a kind of a classic igopoly and or or maybe you know maybe one company's
71:59 going to definitively it'll be it'll be a monopoly and that and by the way those
72:02 outcomes have happened in software many times before and so may maybe that that
72:05 will be the outcome you know the other side of it is you know if you had told
72:10 me three years ago um you know that in the uh you know kind of Christmas of
72:13 chat GPT that like within basically a year to year and a half there would be
72:17 you know five other American companies that would have basically basically, you
72:21 know, exactly capable products. Um, and then there would be another five
72:23 companies out of China that would have exactly capable products and then there
72:26 would additionally be open source that was basically the same. Um, I would have
72:31 been like, wow, like it, you know, the thing that seemed like it was Blackmagic
72:33 all of a sudden, you know, has has become like commoditized really fast,
72:36 you know, which which by the way is exactly what happened, right? Like, you
72:40 know, within within a year of GPT3 coming out, there were their open source
72:43 GP3s running on a fraction of the hardware, right? That were available for
72:46 free. Um and then there were and then you know there were five you know now
72:49 now you've got you know in the game you know fully in the game you've got Google
72:52 and you've got Anthropic and you've got XAI and you've got Meta and you've got
72:54 you know all these other companies that are and then DeepSeek and you know Kimmy
72:58 and all these other Chinese companies. Um and so like even at the level of like
73:03 LLMs or you know AI models like you can squint and make that argument either
73:07 way. By the way same thing at the level of apps right it's like you know one
73:11 school of thought is you know the apps apps are not a thing because like the
73:14 model's just going to do everything. Um but another way of looking at it is no
73:18 actually like actually adapting the model as kind of the engine into a into
73:22 a domain involving human beings u where you need to like actually have it fit
73:24 for purpose to be able to function in the medical industry or the legal
73:28 industry or you know or whatever u or coding you know no you actually need
73:31 like the application level is actually going to matter enormously and maybe the
73:34 LLM's commoditizing maybe the value goes to the apps um and and and again you can
73:38 kind of squint either way on that one and I and I know very smart people who
73:41 are on both sides of that argument um and so I my honest answer on this is I
73:44 think we're in a process of discovery over time um which is you know in the
73:48 way I think about this kind of structurally is it's a complex adaptive
73:52 system the technology itself you know provides one of the inputs the legal and
73:56 regulatory process you know is another input um in you know actual individual
74:00 choices made by entrepreneurs um you know matter a lot um you know the
74:05 economics matter a lot availability of investor capital varies over time that
74:09 matters a lot um and this is a this is a complex system and so we we actually
74:13 don't know the the outcomes on this yet and and we need to basically be we need
74:16 to be open to surprises at the structural level uh of what happens and
74:21 of course as a as a VC this is very exciting because it means we you know
74:23 we're doing this now we should kind of make bets along every one of these
74:27 strategies um and kind of see and see how this plays out and I just say like
74:31 there may be like one I don't know there may be like one particularly brilliant I
74:34 don't know hedge fun manager or something who has this all figured out
74:37 but I I guess I would say if if if they exist I haven't met them yet.
1:14:39 The rapid evolution of AI models
74:42 So what I'm hearing here is don't over obsess with moes at this point because
74:45 we have no idea what it'll end up being and as much as it may feel like okay
74:48 there's no way OpenAI will lose this lead clearly we're seeing a lot of
74:52 competition GPT rapper point is really great a lot is such a derogatory term I
74:56 don't know year ago just like you're just GPT rapper now it's like the
74:59 companies that are the biggest companies fastest growing companies in the world
75:02 >> yeah well it's it's like a little bit like I don't know I mean even just like
75:05 with you know you know the you know this has been the you know the the holiday if
75:08 you know three years ago was the holiday of Chad GPD this last, you know, month
75:12 or whatever has been the holiday of of Claude, particularly Claude Code, right,
75:15 for for coding, but it's like, you know, it's pretty amazing because it's like,
75:17 okay, there was Claude, which is, you know, obviously a great accomplishment,
75:20 but then there's Claude Code, which which is an which is an app, right? It's
75:24 a cloud rapper, >> right? It's, you know, agent harness.
75:28 Um, and then um and then they did this amazing thing where they came out with
75:30 was it co-orker? >> Co-work. >> Co-work um and uh and remember they said
75:35 coowork, which is a club code wrote co-work in a week. >> Yeah. A week and a half. Yep. 100%.
75:40 Well, and that's and there's two ways of looking at that, which is like, wow,
75:44 that's really imp obviously that's really impressive that cloud code was
75:46 able to build co-work in a in a week and a half. That's great. That's amazing.
75:50 The other way to look at it is co-work was developed in a week and a half like
75:56 like h how much complexity could there be? How much of a barrier to entry can
75:58 there be in something that was developed in a week and a half? And so and and
76:02 then you know and then again it's this it's this it's this push and this pull
76:05 thing where it's like it's like wow it's incredibly val it's incredibly
76:08 functional incredibly valuable and people are like all over the world every
76:10 day now are like wow I can't believe what I can do with this is like the most
76:13 magical product ever but at the same time it took a week and a half right and
76:17 so right and so every other every other model company you know I'm sure you'd
76:21 have to expect is sitting there being like okay obviously we need to build you
76:24 know an Asian artist and then obviously we need to build a co-work you know
76:27 thing for for for regular people and obvious you know I I don't I'm not even
76:30 saying I know anything, but just like obviously they're all going to do that,
76:33 right? Um and so, you know, how defensible is that? And you know, in six
76:36 months, you know, and we've seen this happen before, like in is quad code
76:39 going to get lapped the same way that you know, GitHub copilot got lapped. You
76:42 know, the the history in the last three years has been everything that looks
76:45 like it's like the fundamental breakthrough gets gets basically
76:49 replicated and lapped very quickly. Like many of the smartest people I know in
76:51 the field when I when I really kind of talk to them kind of, you know, get a
76:53 couple drinks into them, they're like, "Yeah, they're basically, you know, one
76:57 theory is like there really aren't any secrets among the big labs." like the
77:00 big labs kind of all have the same information and they kind of have all
77:02 the same knowledge and they you know they're kind of they lap each other on a
77:05 regular basis but you know there's not a lot of proprietary anything at this
77:08 point and then and then you know again evidence of that is you know deepseek
77:11 you know came out of left field and basically was like a you know
77:15 re-implementation of a lot of the ideas under American big labs and you know and
77:20 had some original ideas of its own um but like you know wow it wasn't that
77:22 hard for you know some you know basically a hedge fund in China to do it
77:25 and so like how much defensibility is there but on the other side of it you've
77:28 got wow all these big labs are now paying you know individual engineers
77:32 like they're rock stars um and they're you know incredibly bright and creative
77:36 people um and you know maybe there's you know a dozen nent ideas in any one of
77:39 these labs that it's actually going to be a huge breakthrough that's going to
77:42 be hard to replicate and so again it's just like I think we just need to I
77:45 don't know my views I my view I need to put like a big discount on my
77:49 forecasting ability on this one like it for me it's much less interesting to try
77:52 to say okay as a consequence industry structure in five years is going to be X
77:55 the big winner in the category is going to be company Y the big you know product
77:59 killer app is going to be It's like I this is to say I don't think I can
78:02 predict that. Um I I think I I think a much much better use of my time is is
1:18:05 Indeterminate optimism in venture capital
78:06 being being very flexible and adaptable at a time like this.
78:09 >> So with all this in mind, do you feel like there's something you're paying
78:12 attention to more to help you decide okay this is where we want to place our
78:15 bet or is the answer essentially the strategy you guys have which is place a
78:19 lot of bets. You guys raised the the largest fund in history. Is that is that
78:23 the way you win in this world? >> Yeah. So for I mean for us yeah for for
78:26 us we have we obviously have a very very deliberate strategy. One one way to
78:29 think about this used the Peter Teal for you remember the Peter Teal formulation
78:33 of uh he said there's a two by two there's optimism and pessimism and then
78:37 there's determinant and is it indeterminate and indeterminate right um
78:43 and so um and he always argued that like there's he always argued that like
78:46 Silicon Valley is characterized by in too much what he calls indeterminant
78:49 optimism right and what he what he what he always described what he meant by
78:52 that is basically um I think the way he would describe it is an indeterminant
78:55 optimist who thinks the world is going to be better but can't explain are right
78:59 like some combination of things is going to happen to make the world be better
79:02 even if we don't know what those things are and and you know I think he he at
79:05 least historically would say like that's that's basically you know that that that
79:09 that risks at least being just like wishful thinking or delusional thinking
79:13 and what the world needs more is determinant optimists which are people
79:16 who are like no the world is going to be better because I'm going to do this
79:19 specific thing right and he would classify for example Elon you know he
79:23 would s sort of maybe say you know VCs are indeterminant optimists um and then
79:26 he would say you know Elon is the determinate determinate determinant
79:30 optimist where it's like no I'm going to build the electric car and I'm going to
79:33 you know solar and then I'm going to do you know Mars you right and I'm these
79:36 very concrete things and I I think there's a lot I think there's a lot to
79:39 Peter's framework but the way I would describe it is I I think maybe he and I
79:42 if you disagree with part of that it would be I think the indeterminant
79:45 optimism is a stronger phenomenon than at least I think he's historically
79:48 represented it as and I would put myself firmly in the indeterminant optimist
79:52 category and that's the strategy that we that we have at A6Z which is and and the
79:56 reason for that is It's not hopefully it's not so much wishful thinking. It's
80:00 more no what the indeterminant optimism of venture capital or the indeterminant
80:03 optimism of A6Z or Silicon Valley is very it's actually very specific which
80:07 is there are these extremely bright and capable people like Elon and many others
80:13 who are founders right and product and you know kind of product creators right
80:17 and and and each of those individual people is a determinate optimist like
80:20 each of them each of them individually has like a very strong view what they're
80:24 going to do but the great virtue of the capitalist system the great virtue of
80:27 the American economy the great virtue Silicon Valley is we don't just have one
80:30 of those and we don't just have 10 of those. We have a hundred and a thousand
80:34 and then 10,000 of those and and the way to optimize the outcome is to have as
80:37 many of those as possible be as good as possible. Run as hard as possible and
80:41 and then just the the nature of you know the nature of the future is like we just
80:44 don't know all the answers and that's okay and then and the right way to deal
80:48 with that is to run as many experiments as possible and have as many smart
80:50 people try to do as many interesting things as possible. Um and so yeah, I
80:53 would I would put myself firmly on the side of the indeterminate optimist. I'm
80:57 uh I'm wondering if the answer to the question of what you look for now more
81:01 and more is this determinate optimistic founder. Yeah. That has this massive
81:05 ambition and is actually working on achieving it. >> Yeah. Yeah. No, that's right. That's
81:08 right. I mean, look, the founders need to be deter determined optimist. Like
81:11 they need to have a very specific plan now. And look, the the critique the
81:15 critique always, you know, the critique from the founders is, oh, UVC's have it
81:18 easy because like you don't have to like you don't actually have to commit,
81:20 right? You don't actually have to like make you you don't actually have to
81:22 like, you know, you don't have to make the bed you lay in. You can like place
81:25 multiple bats. you can operate a portfolio, you know, you should have a
81:28 lot more sympathy for us as founders, you know, because we, you know, we only
81:31 get to make the one bet. Um, you know, and there's there's truth to that. You
81:34 know, the counter-argument on that is the founders get to run their companies.
81:39 We don't. So, so, you know, we we don't we don't get to put our hand on the
81:42 steering wheel. And so, you know, the great virtue of being a determined
81:45 optimist is you actually get to get to single-mindedly execute against that
81:49 goal. And and and you look, in the long run, who who does history remember?
81:52 History remembers Henry Ford, right? not you know whoever was the you know
81:55 whatever the seed investor who seated at Ford Motor Company and you know 10 other
81:59 car companies have failed right um and so you know the determinant optimist is
82:02 the per you know the founder is the founder and the company builder and the
82:05 engineer I mean these are the people who actually do the thing and you know
82:08 deserve 99.99999% of the credit but uh you know having said that I I do think there is a role
82:13 for having having some indeterminate optimist in the uh in the background
82:16 helping along the way and helping keep the whole the whole cycle going
1:22:17 The concept of AGI and its implications
82:21 >> do you think about AGI in shifting your investment thesis like as we approach
82:27 AGI and hit AGI as an investor, how do you think about your investment thesis
82:29 changing? >> Yeah. So, I've always kind of had a little bit of an is I've always kind of
82:35 struggled with the concept of AGI um because it at least well there there's
82:40 those defined terms which is where I kind of struggle with it which is
82:43 there's like the prosaic there's the there's the prosaic uh definition of AGI
82:47 and then there's like the I don't know cosmic definition and the way I would
82:50 describe it as well let's start with the cosmic one. So the the cosmic one is
82:54 basically it's the singularity, right? Um and so AGI is the is the moment where
82:58 you enter the singularity, which is to say that where the world fundamentally
83:01 changes and like the the rules of the old world are gone. We're now operating
83:05 in a new domain and then you know the kind of the full definition of
83:08 singularity is like it's a world in which you know human judgment is no
83:11 longer really relevant because the you know you get this self-improvement loop.
83:15 The AI the AI is improving itself and it's sort of racing you know so-called
83:18 takeoff scenarios. you can see at this takeoff thing where the AI is improving
83:21 itself and the machines are making decisions so much faster than people and
83:24 people are just sitting there watching the the machine do its thing you know
83:28 and I kind of described I don't really I don't really think that's I don't I
83:31 don't think we live in that world like whether you could call that utopian or
83:33 dystopian like I don't think we're lucky or unlucky enough to live in that world
83:36 we could debate that we can talk about that more but um the the the pros
83:40 definition of AGI that at least I think the industry participants have kind of
83:42 converged on and tell me if you agree with this is uh it's when the AI can do
83:46 every economically relevant task as good as a The way um the co-founder of
83:50 Anthropic put it is like a basket of the most valuable economic task. So it's
83:55 like 10 15 not every single economically valuable task. >> Okay. Got got it. Yeah. So it's maybe
83:59 even a slightly reduced slightly reduced definition. Um and by the way we're you
84:02 clearly getting close to that if we're not already there. >> And so on that one I kind of feel like
84:07 so I kind of feel like the cosmic one overstates what's going to happen. And
84:11 then I kind of feel like the kind of AGI definition that you just gave I think it
84:14 kind of understates what's going to happen. like it's almost too
84:18 reductionist and and the reason for that is I don't think there's any reason to
84:22 assume that human skill level is the cap on anything right and so the way we say
84:28 that is AGI always is you know the definition you gave the definition I
84:30 gave it's kind of in it's always kind of relative in comparison to a human worker
84:35 right and it's like I don't know like human skill level caps out at a certain
84:38 point but that's because of the inherent like biological limitations of the human
84:42 organism right like we're you know human I give you example human IQ human IQ Q,
84:46 you know, kind of what they call fluid intelligence or the the sort of G factor
84:50 of kind of uh, you know, fluid intelligence. Uh, IQ, I think, tops out
84:55 in in humans as a species, it tops out around 160, right? Where at at like 160,
84:59 it's like Einstein level, Einstein, Fineman IQ, >> in terms of IQ. Like, you just tops out
85:05 at 160. The the 160 IQ people are the ones who come up with new physics.
85:08 There's only a small handful of those. the generally speaking when we run into
85:11 somebody in the world who's like incredibly smart who's like a
85:15 best-selling author or like a you know one of the world's best I don't know
85:18 research scientists or one of the world's best doctors you know or
85:23 whatever um it would be probably 140 um is kind of the IQ that you're looking
85:27 for there. Um if you're looking for like a really good lawyer it's probably 130.
85:31 Um if you're looking for like a really good like line manager in a business
85:35 it's probably 110. um you know if you're looking for like an accountant like a
85:38 small business accountant who's good at doing the books for small businesses is
85:42 probably 105 right and so the the kind of scope of like impressive human you
85:46 know the the the ability of the human organism to do intellectually impressive
85:51 things you know it's sort of that 110 to 160 is kind of the spectrum and you know
85:54 good news is there's a lot of those people running around but like there's
85:58 not that many at 140 150 160 but it's like that's just that's like the
86:01 limitations of what can fit in here right and it's like there's no
86:05 theoretical limit on where this goes if you release the limitations of human
86:11 biology, right? And so can you have a you already have people running these
86:13 experiments to kind of do human equivalent, you know, kind of IQ uh uh
86:17 you know, for for existing AM model. And by the way, existing AI models right now
86:20 are kind of testing around the 131 140 level, which means they're going to get
86:23 to the 160 level and they're, you know, they're arguably on the mass high
86:26 starting to get to the 160 level now. But like I I think we're going to have
86:30 AI models relatively quickly that are going to be like 160, 180, 200, you
86:35 know, 250, 300, by the way. And I think that's great, right? Like I feel I feel
86:39 I feel as great about that as I do about the fact that we occasionally get an
86:41 Einstein, right? It's like would the world be better off or worse off with
86:44 more or fewer Einsteins? And the answer is of course the world would be better
86:46 off with more Einstein. And of course the world would be better off with
86:49 machines that have IQ, you know, more IQ like Einstein are greater than Einstein.
86:53 But like I think IQ's IQ of the machines is going to exceed that in the humans. I
86:56 think that's that's really good. Um, and then the performance, you know, again,
86:59 it goes back to like the AI coding thing is happening. The performance against
87:02 task is going to get better also. Like I I think, you know, this is where Line of
87:05 Stars in particular is like, yeah, okay, like this thing is starting to generate
87:08 better code than I can. Okay, so now we're going to have AI coders that are
87:11 actually better coders than the best human coders. I think that's great. I
87:14 think we're going to have AI doctors that are better than the best human
87:16 doctors. I think we're going to have AI lawyers that are better than the best
87:19 human lawyers, which actually is going to be very interesting to see. Uh, which
87:22 we can talk about, which I think is also great. Um, and so like I don't think
87:26 there's a I think we're used to living in a world where we just don't
87:29 understand how good good can get because we've been capped by our own biology and
87:33 we're going to get to experience what it's like when you have the capability
87:36 at your fingertips that's actually better than human in these domains. Um,
87:40 and so I I you see what I'm saying which is like I think this idea of like human
87:44 equivalent is just going to be like a footnote. It's like, oh yeah, that was
87:48 just on Tuesday, you know, in in 2026 is when they hit that and it kind of didn't
87:53 matter because the the next question was like, okay, what are we gonna what are
87:55 we gonna what do we get to do in a world in which we actually have machines that
87:59 are better than that, right? And so so so I think this is going to be much more
88:02 of an exploratory process for actually exceeding human capability than it's
88:05 going to be any sort of particular singular singularity moment or whatever
88:09 that happens just that just happens to coincide with the human threshold.
88:13 >> 200 IQ. I uh just like that frame of reference is such a uh mindexpanding way
88:17 to think about just how fast and how smart these things are going to get and
88:20 and quickly. >> Well, I don't know if you have this experience. I I have this experience all
88:25 the time. Well, two two experiences I have all the time. One is just like I'm
88:30 just like like I know I ought to be able to do this, but like I just can't like
88:35 it's going to take too long. You know, I I want to write this thing or I want to
88:37 like whatever. I want to have this theory on this thing or I have a plan or
88:41 whatever. And it's just like like I I don't have the eight hours or or by
88:45 the way the eight weeks or the eight years, right? And like I just don't know
88:50 enough yet and I'm just like I can't do the math in my head and my memory isn't
88:54 perfect and like I can't remember and I read you know after you had this you get
88:57 interested in something you read 10 books and then you're like I forgot
89:00 almost everything that I just read. Like I I wish I could retain it all but I
89:04 can't. It's just like I you just have this I I sort of live in this kind of
89:07 state of like endless frustration. So, it's like I like if I could just be
89:11 smarter than I was, like I'd be so much better at what I do, but I'm not. So,
89:15 so, so there's that. And I don't know how often you have this, but I have this
89:18 on a regular basis. It's just like, you know, I, you know, because of what we
89:22 do. Like, I know a bunch of people who I know for sure are smarter than I
89:25 am. And I know it because when I talk to them, I just find myself at a certain
89:28 point, you know, it's like for the first half of the conversation, I'm just
89:30 taking notes the entire time. And for the second half of the conversation, I'm
89:34 just like, Like, me. like this person is just smarter than I am
89:37 and they're just outthinking me and they're going to keep outthinking me and
89:40 I just can't and I'm just like all right god damn it like I gotta go home and I
89:43 gotta like have a drink because I'm just not, you know, I'm just not whatever
89:47 that is. I'm not that. And so we're just so used to having those limitations
89:54 um that the idea of having machines that work for us that don't have those
89:57 limitations. I I just I think that's much more exciting than people are
1:30:00 Marc's media diet
90:00 giving you credit for. >> Oh man, I could talk to you for for
90:03 hours, Mark. I'm thinking to close out the conversation, I want to ask about
90:08 your media diet and your product diet. You just talked about books, reading 10
90:11 books. I I think you famously read constantly. I saw a interview with you
90:15 where you're just like AirPods changed my life. I'm just listening to audio
90:18 books now all the time. So, in terms of media diet, what do you what are you
90:20 reading? What are you paying attention to these days in terms I don't know
90:23 podcasts, newsletters, blogs, things like that. And then any books in
90:25 particular? >> Yeah. Yeah. So, what I read is basically
90:29 I mean I would say read basically three categories of things. So like in terms
90:32 of like general media um it's basically I I sort of um I always describe it as I
90:36 have like a almost a perfect barbell strategy um which is I read X and I read
90:41 old books right so it's basically either like up to the minute what's happening right now
90:46 um or it's like a book that was written 50 years ago that has stood the test of
90:50 time and then you know we're presumably there's something timeless in it. Um,
90:54 and and then it's sort of everything in the middle I'm always like much more
90:57 skeptical about. And and it's particular it's kind of what I already said, which
91:01 is I think if you go back and you read old nobody ever does this. It's actually
91:03 really funny. Nobody ever does this. There's no market for it. But if you go
91:08 back and you read old newspapers, and by the way, you can you can do this.
91:11 Just read last week's newspaper, right? I guess we're taping on Friday. So read
91:15 last Friday's newspaper, right? And just go back and read it and be like, "Oh my
91:21 god, like none of this happened." like n that none of what they predicted played
91:24 out the way that they said that it would. None of this turned out to
91:28 actually be that like relevant or correct. Like they didn't understand
91:31 like you know they by the way they had no view of what was going to happen this
91:34 week that they couldn't know and so they were making predictions and forecasts
91:36 and so forth based on like not having any information but it's like wow like
91:40 you know like none of this happened like I wish I had never read this like oh my
91:43 god. Um and then you know it's kind of the same thing with magazines like go
91:46 back and read old magazines. Um and just like the the the the level of the you
91:50 know the just the endless numbers of predictions that they make. Yeah. And
91:52 and kind of you know the problem with you know newspapers at least they're
91:54 going dayto-day. The thing with magazines is like every it's like a week
91:58 or month you know kind of long cycle and so it's even you know by the time an
92:01 article even hits publication it's you know it's often out out of date. So I
92:04 just I just have like a big problem with kind of everything in the middle. Um and
92:07 so it's either it's either it's either of the moment or timeless. But then yeah
92:10 you mentioned like newsletters. I mean, so the the other thing and you know,
92:13 this is maybe obvious, but I think it's probably still underrated, which is the
92:17 actual practitioners in the field who are actually creating content, I think
92:20 probably is still like dramatically under underrated and I think this is a
92:23 huge part of like the Substack phenomenon and the newsletter phenomenon
92:26 and the podcast phenomenon is like direct exposure to the people who are
92:29 actually principles in the field who actually know what they're talking about
92:32 is probably still dramatically underrated. And I think again the reason
92:35 for that is like we we're we're used to being in this mass media kind of culture
92:38 in which basically everything is mediated, right? everything got filtered
92:41 through like TV interviews or like newspaper interviews or magazine
92:45 interviews and and you know obviously now more and more it's just no you
92:47 actually want like smart people who are actually working on something explaining
92:50 themselves and then you have you know you have new kinds of intermediation
92:53 like podcasts that that that kind of open that up for people to make that
92:57 possible um and so yeah like domain practitioners are um you know really
93:01 great I mean just to state the obvious and AI you know it's obviously your your
93:04 stuff but also like you know let Lex you know you know the fact that like Lex
93:08 Friedman can have you know the world's leading or you know whoever the you know
93:11 any of you guys, you know, there's a small handful of you guys who have
93:12 access to these people. You can have the world's, you know, kind of leading
93:15 experts in the domain actually show up and and by the way, it's, you know, it
93:19 looks the critique always is, you know, people talk their book, like if I'm
93:21 running a startup or whatever, I'm just selling, but it's like and there's
93:25 always a little bit of that. Um, but it's also, you know, my experience is
93:28 people love to talk about what they do and and you know, they they
93:31 fundamentally like want to express what they do and and and they want to explain
93:34 it and they want people to understand it and everybody kind of enjoys that and
93:37 they get to contribute to kind of human knowledge by doing that and they get ego
93:41 gratification by doing that. Um, and so I think there's just actually just
93:44 tremendous amounts of alpha in listening to the world's leading experts in the
93:46 space who actually just like show up and talk about what they're doing. And of
93:49 course like the world is a wash in that today in a way that it wasn't as
93:53 recently as 10 years ago. So I yeah I do as much of that as I can too.
93:56 >> And there's also just this culture in tech Silicon Valley in particular of
93:59 sharing of not trying to keep these secrets. Everyone on LinkedIn is always
94:03 like how is this free like it's just the way it works. >> Yeah. It's somebody said Silicon Valley
94:08 is a company town but the the the company is Silicon Valley >> right and but and again at the level
94:13 this goes again there's one of these great n equals one at the level of n
94:15 equals one is somebody you know and I've run startups before run companies
94:18 before. um at the level of n equals one of like running a company that's just a
94:22 giant pain in the butt like because you know your secrets are
94:24 walking out the door and your employees are walking out the door and the whole
94:27 thing sucks. But you know the other side of it is you also benefit from that
94:29 right because you get to hire people with all these skills and experiences
94:33 right and you you're in this you're in this ecosystem that that adapts right
94:36 and channels talents and and and skill and knowledge and people into into the
94:39 new fields and so you know so that you know there's kind of the push and pull
94:42 of that at the level of just being an individual individual CEO um at the
94:46 level of of just being in the ecosystem to your point like yeah it's it's an
94:49 absolutely magical phenomenon and by the way like you know one of the one of the
94:52 you know for all the for all the issues in Silicon Valley um you know I think AI
94:57 I did the comment once I AI is the ninth major technology platform in the history
95:01 of Silicon Valley, right? That, you know, Silicon Valley is Silicon Valley
95:05 is still called Silicon Valley. We haven't made Silicon here in decades,
95:08 right? Uh we used to actually, you know, it's called Silicon Valley because they
95:11 used to make chips, right? They used to have the like the actual fabs were in
95:13 Silicon Valley and then they and they designed them and they made the chips.
95:17 Um and and so and that was you know wave one starting in the 19 actually that was
95:20 like actually no actually more or less like wave three or whatever but like it
95:23 was you know that was when the the indust the the area was named like in
95:26 the 1950s but now we're on like wave nine right um and and the the company
95:31 town phenomenon where the company is the industry like the the the again the
95:35 indeterminate optimism the nobody had nobody had to sit and plan and say okay
95:39 in the 1990s Silicon Valley is going to do the internet in the 2000s they're
95:42 going to do the smartphone in the 2010s they're going to do the cloud in the
95:45 2020s they're going to do AI It it just the the the the right the indeterminant
95:49 optimist optimism of ecosystem flexibility of the ecosystem met that
95:52 the the the Silicon Valley could could morph um into all these categories and
95:56 and again maybe a testimony to indeterminate optimism. >> This reminds me of the meme of how we're
96:01 all just rappers over sand. Everything we're building is just rapper wrapper
96:03 rapper rapper. >> The rapper thing is hysterical. Yeah.
96:06 Yeah. I'm a I'm a software company now. I'm I'm a chip rapper, right? Um uh
96:10 Yeah. I'm a I'm a I'm a I'm a business application. I'm a database rapper.
96:14 >> Um Yeah, exactly. I'm a sand Yeah. You and I are you, we're all now sand
96:16 rappers. >> Sand rappers. >> Perfect. >> Okay, one more question. Along the media
1:36:18 Favorite movies and AI voice technology
96:21 diet, I asked your partner Ben Harowitz uh what to talk to you about? Uh the Z
96:25 and A16Z if people don't know him. And he said that you're really into movies
96:27 these days. >> Yeah. >> And so I don't know any movies. Any
96:31 movies you're really into these days? Any movies you've absolutely loved
96:33 recently? >> Yeah. So the movie that blew my socks off uh last year, which I think is the
96:39 best movie of the decade for sure and maybe of the last like 15 years, is this
96:43 movie. Unfortunately, it's one of these things. Not a lot of people have seen
96:45 it, but I would highly encourage it. It's called Edington. >> Not heard of it.
96:49 >> Have you not heard of it? Okay. So, Ed, you're going to really enjoy it. So, I
96:53 won't I won't spoil too much of it. So, at at at at the surface level, this the
96:58 following spoils nothing. At the surface level, it's set in a small town in New
97:01 Mexico called Edington, which is a small town of about 600 people. Um, and um
97:07 there's a sheriff uh who's played by Waqen Phoenix, who's like an old crusty
97:11 basically right-winger. And then there's a um uh there's a mayor uh played by
97:16 Pedro Pascal who's basically a young hip progressive. And uh and then the movie
97:22 starts I think in March of 2020. And so it starts when COVID first hits and then
97:27 it sort of as it plays out over the next few months it it then it intersects and
97:30 it it sort of extends into the summer of 2020. So you know kind of the the George
97:34 Floyd moment and then the you know the the protests and riots and kind of
97:37 everything. So sort of the convergence of COVID and then the um and then the uh
97:41 and then and then the uh the all the all the BLM stuff and and and and then um it
97:46 and then and then there's a third kind of element to it which is um there's a
97:49 company which is basically a loosely disguised version of meta if you read
97:52 the backstory of it which is building an AI data center on the outskirts of town.
97:56 So they kind of pull that in uh as sort of a thing that looms larger and larger
98:00 over time. And then um the thing it really is great at is it really shows um
98:03 you know this is a small town in New Mexico and so everybody in the town gets
98:07 kind of fully wrapped up in all the co stuff and they get fully wrapped up in
98:09 all the BLM stuff and they get fully wrapped up in all the like you know tech
98:13 anxiety stuff but they're all experiencing it basically through the
98:17 internet right which which is which is you know what what actually happened
98:21 right and so so it it's it's so so the reason I love the movie so much is one
98:24 one is it's the first movie that directly grapples with 2020 of what
98:28 happened in 2020 and it just like fully fully engages and grapples with like all
98:30 the dynamics that were playing out in the country. But the other reason is
98:32 it's the first movie that does a really good job of showing what it what it what
98:36 it was like especially in that era to live in a world in which there were
98:39 things happen in the real world and people were kind of experiencing events
98:43 online, you know, like in a way that was like very central in their lives, right?
98:46 Um and so it does like a really good job of pulling in like smartphones and
98:50 social media um in a way that um uh in a way that movies really really really
98:53 struggle with. And then the whole thing comes together in an incredibly
98:56 entertaining way. Um, and so, and I won't even say I I I won't even say I
98:59 completely agree with the movie or whatever, and I think the director of
99:01 the movie and I would probably disagree about a lot, but he really tries hard to
99:06 like really grapple with like what is actually like to live like a human being
99:10 in the 2020s in America in a way that I think many other filmmakers who are very
99:13 talented have just been very scared of touching. And and this guy, for some
99:17 reason, he's just like, "Yeah, I'm just going to find all the third rails and
99:19 I'm just gonna like grab them." >> I can see why that's your favorite movie
99:21 of the year. >> It's great. It's great. It's great. Everybody should see it.
1:39:24 Marc's product diet
99:25 >> Oh, man. Okay, final question I want to ask about your piet uh your product diet. Are
99:31 there any products you use that maybe are less known that you love that you
99:34 want to recommend? You can, you know, mention products you're investors in if
99:37 if you use them constantly. >> I mean, we have, you know, we have so
99:39 many that it's really hard to, you know, I always feel it's like, you know, who's
99:42 your favorite children? So, it's it's really hard to to to uh to uh you know,
99:46 to to to pull out specific ones. Um, but I'll, you know, I'll talk about a few.
99:51 Um, I mean, just I'll just observations. So, one is my my 10-year-old. Um I my
99:54 10-year-old my 10-year-old right now is 100% obsessed with Replet. Um and and by
99:59 the way, it was not from me. Do you have kids? >> I do. I have one two and a half year
100:01 old. >> Two and a half. Okay. So, you haven't run into what I'm running into now,
100:05 which is whatever it is you do is not cool, right? Like two and a half. Whatever
100:10 daddy does is like the coolest thing in the world. I can tell you by the
100:14 time he's 10, whatever you do is like deeply uncool, right? And and I'm highly
100:18 aware of that. Um, and so like if I mention, oh yeah, we work on XYZ, you
100:22 know, he's like, okay. Um, but when he discovers something, then then it's
100:25 cool. Or when his friends tell him about it, it's cool. And so he he he through
100:29 no inter interference on my part uh discovered Replet about uh about three
100:33 months ago and discovered vibe coding and is like completely obsessed with
100:36 vibe coding games and all kinds of all kinds of things and like literally was s
100:39 do it for hours and so I'm seeing that phenomena play out. Uh which is super
100:44 fun. Um uh that's one. Two is I am just completely in love with all the AI voice
100:48 stuff. Um I think it's just absolutely amazing, hysterical. Uh my favorite
100:53 party trick at dinner parties now is to pull out uh Grock uh with Bad Rudy,
100:58 which is if you've seen it's it's the it's a foul mouse raccoon uh avatar on
101:05 the uh in in the Gro app. So um I think that's super fun. We have this company
101:08 Sesame that had, you know, they they went viral last year for this, uh, you
101:13 know, the this just incredibly like, you know, intimate, emotional, you know,
101:17 kind of voice experiences. Um, so I think the voice stuff is fantastic. I'm
101:20 also super fascinated by all the voice input stuff. Um, and so um, you know,
101:26 you know, most recently that company recently sold, but um, you know, the all
101:30 the I think like the pendants, the wearables, like all that stuff is going
101:34 to be big. The meta glasses um, I you know, I think there's going to be a
101:36 whole wearables revolution here. Um, I I love the voice input stuff. Um, I have
101:40 this app on my there's this app on my phone now called Whisper Flow. Um, which
101:45 is voice transcription. Um, which works like staggeringly well. Um, uh, it's
101:50 like incredibly it's like a voice transcription function, but you can
101:52 actually talk to the AM model while you're doing voice transcription. So,
101:55 you can kind of it kind of understands when you're telling it, no, no, you
101:57 know, I want bullet points over there and I want this and that. And it
101:59 understands that you're not telling it to type in the words I want bullet
102:02 points. It just actually understands that you want bullet points. And so like
102:05 that's a great example of a super useful thing. And so I I think the voice mode
102:09 stuff is going to be is going to be uh is going to be really great.
102:12 >> Uh subscribers of my newsletter get a year free of Replet and Whisper Flow. So
102:17 there we go. Uh uh what's the what's the most memorable thing your son built with
102:19 Replet? >> Oh well so he's gotten super into Star Trek. Um and so so far it's been he's
102:24 like writing like Star Trek simulators. Um >> so like all the you know all the by next
102:29 generation they actually had >> Next generation. Okay. I was going to
102:30 ask which >> Well, he like we actually we like them all. We watched the new Starfleet
102:34 Academy last night which actually is quite is actually quite good. Um but uh
102:37 we we watched the original, you know, we watched we watched them all, but it was
102:40 in next generation where they actually developed an actual design language for
102:42 the computers >> because if if you watch the original series, they just had like basically,
102:47 you know, knobs with lights and they didn't really, you know, they just like
102:49 were like, you know, around on set trying to pretend they were doing
102:52 it. But by next generation, they actually had designed, they actually had
102:56 a UI design language. So, one of the one of the fun things you can do v coding is
102:58 you can say give me a Star Trek next generation, you know, user interface
103:01 for, you know, whatever this that or whatever. And it actually uses the they
103:05 call it this I'm a nerd now. They call it LCARS um design language and um it'll
103:09 you know it'll actually build you like Star Trek Next Generation British
103:13 um using that design language but you know with your choice of like a
1:43:16 Closing thoughts and recommendations
103:17 Star Trek game for example. Um and so he's he's going crazy for that kind of
103:19 thing. >> That sounds extremely delightful. You guys should uh open source or release
103:24 that. Mark, I like I said, I could talk to you for hours. Uh, you got things to
103:29 do. Uh, anything you want to leave listeners with before we wrap up?
103:32 Anything you want to double down on or just leave listeners with?
103:34 >> Yeah, so a couple things. So, one is we got super lucky last week. Uh, Py
103:38 McCormack uh wrote the best piece ever written about us actually? Um, which he
103:42 released um and so it's the best explanation of what we do uh and how we
103:45 think. And so I I would definitely recommend that. Um, and then you know
103:48 we're putting a lot we have a you know great team of folks now. We're putting a
103:52 lot of effort ourselves into video um in you know in content um and so I
103:55 definitely recommend our YouTube channel which I I think has a lot of great stuff
103:58 and is going to be very exciting in the next year. >> Awesome. We'll link to that. I think
104:02 it's just YouTube.com6Z something like that. And you guys have
104:04 great stuff. >> Mark, thank you so much for being here.
104:07 >> Awesome. Thank you for having me. I really I really appreciate it.
104:10 >> Bye everyone. >> Thank you so much for listening. If you
104:13 found this valuable, you can subscribe to the show on Apple Podcasts, Spotify,
104:19 or your favorite podcast app. Also, please consider giving us a rating or
104:22 leaving a review as that really helps other listeners find the podcast. You
104:27 can find all past episodes or learn more about the show at lennispodcast.com.