Latest / In the Loop AI / Shaun Maguire: This Is Far Bigger Than The Internet Boom

Transcript

Shaun Maguire: This Is Far Bigger Than The Internet Boom · In the Loop AI

0:26
Speaker2: Picture a guy who treats the frontier like his backyard. That's Sean McGuire,
0:30
Speaker2: and he was gracious enough to join us on the show today to discuss energy breakthroughs,
0:33
Speaker2: factory floors humming with humanoid robots, brain machine computers, and space exploration.
0:38
Speaker2: In fact, Sean's even invested in a company who shoots satellites into outer
0:41
Speaker2: space to beam light rays back at Earth and turn night into day.
0:44
Speaker2: It's crazy stuff. And he walks us through exactly which of these sci-fi visions
0:48
Speaker2: are already sneakily here today, which will upend entire industries in months,
0:52
Speaker2: and which have the opportunity to reshape entire civilizations over the next
0:55
Speaker2: couple of years. Some of the timelines even left me speechless.
0:58
Speaker1: And that is what the Limitless podcast is all about.
1:02
Speaker1: So Bankless Nation, welcome to Limitless, where we explore the frontier technologies
1:07
Speaker1: that are poised to reshape our world.
1:09
Speaker1: I'm David Hoffman, joined by my co-host, Josh Kale. And our very first guest
1:13
Speaker1: is Sequoia Capital Frontier Tech investor, Sean McGuire.
1:17
Speaker1: Now, if you are hearing this on the Bankless feed, don't worry,
1:20
Speaker1: Bankless isn't going anywhere.
1:22
Speaker1: And if you found us on the Limitless feed, congrats, you're ahead of the curve.
1:26
Speaker1: Limitless is a brand new podcast.
1:28
Speaker1: Out of Bankless, focusing on
1:30
Speaker1: frontier technologies and getting ahead of a rapidly accelerating future.
1:34
Speaker1: Bankless will return to focusing on crypto and markets with Ryan and me at the helm.
1:39
Speaker1: And Limitless is where Josh, Ijaz and I, and also sometimes Ryan, will chase the future.
1:43
Speaker1: How AI and other frontier technologies are going to impact our lives and what
1:47
Speaker1: we need to know and do about it.
1:49
Speaker1: If you're familiar with the bankless cadence, you'll find yourself right at
1:51
Speaker1: home with Limitless. On Monday, we're going to do deep dive interviews with
1:55
Speaker1: founders and investors that are building the future.
1:57
Speaker1: And then on Thursdays, we have the weekly AI roll-up, our fast-moving recap
2:01
Speaker1: of the week's biggest AI headlines.
2:02
Speaker1: Now, I'm pretty biased here, but I think the AI roll-up is truly the most accessible,
2:06
Speaker1: informative, and engaging AI news show out there.
2:09
Speaker1: So if all of this sounds like your jam, hit subscribe on your favorite podcast app or YouTube.
2:14
Speaker1: There's a link in the show notes to go find Limitless. And if you don't mind,
2:17
Speaker1: we would definitely appreciate a quick five-star review and also sharing it
2:20
Speaker1: with your friends to help more curious minds stay ahead of the curve.
2:24
Speaker1: Your support, of course, powers everything we do.
2:27
Speaker1: And so thank you for coming on the ride with us as we push into brand new frontiers.
2:32
Speaker1: Now with that, let's get right into the debut episode on The Limitless Podcast
2:35
Speaker1: with Sean McGuire from Sequoia Capital.
2:38
Speaker1: Bankless Nation, today on the podcast, we have Sean McGuire.
2:40
Speaker1: He's a partner at Sequoia Capital.
2:42
Speaker1: Sean's got one of the more interesting backgrounds that I've ever seen.
2:45
Speaker1: Starting out working as a physicist, working on quantum gravity,
2:48
Speaker1: then spent time working at DARPA, where he was deployed to Afghanistan,
2:52
Speaker1: Then shift gears into startups and investing.
2:54
Speaker1: He co-founded a cybersecurity company called Expanse, which was later acquired by Palo Alto Networks.
2:59
Speaker1: And then he's also been an early backer of companies like Visa, Watershed, SpaceX.
3:04
Speaker1: At Sequoia, he's focusing on early stage investments, especially in hard and
3:07
Speaker1: frontier technologies, which is why we wanted to bring him on the show today.
3:12
Speaker1: Sean, welcome to Bankless.
3:14
Speaker0: What's up, guys? Thrilled to be here. First of all, did not start with quantum gravity.
3:18
Speaker0: Started probably with an F in Algebra 2. So I've had a lot of failures before the successes.
3:24
Speaker0: And unfortunately, I was on early background. Visa, I think you meant Vize.
3:28
Speaker1: Vize, excuse me.
3:29
Speaker0: Just all good. Visa would have been pretty legendary, and hopefully Vize becomes even bigger.
3:36
Speaker1: Yeah, I guess Visa is just the wrong generation. It was a little bit earlier than all of us here.
3:41
Speaker1: So maybe with that, since I need a little bit more help kind of understanding
3:44
Speaker1: your background, just expand on your perch a little bit for listeners who just
3:47
Speaker1: might need more context about, you know, what your deal is, who you are,
3:51
Speaker1: and especially as we get into more modern times with all of the crazy technologies that we have today,
3:55
Speaker1: just explaining a little bit of your context and what interests you and how
3:59
Speaker1: your arc has been, I think would be pretty helpful.
4:02
Speaker0: Sure. Thrilled to be here. I've been a big fan of.
4:04
Speaker3: The pod for
4:05
Speaker0: A long time. Appreciate it. So I'll try to keep this pretty brief,
4:09
Speaker0: but I would say kind of my childhood was characterized by just like not listening to what I was told to do.
4:19
Speaker0: Instead of paying attention in school or like even going to class a lot of the
4:24
Speaker0: time, I just stayed home or would do, or even in school, would focus on my own learning.
4:29
Speaker0: And I've basically been obsessed with computers, the internet.
4:34
Speaker0: Like then physics, then math.
4:38
Speaker0: Basically my entire life, like computers really started at the age of seven.
4:41
Speaker0: And so I kind of sacrificed not learning English very well for learning a lot
4:47
Speaker0: about computers at a young age.
4:50
Speaker0: And, you know, I took it pretty seriously. I was really into building computers,
4:54
Speaker0: really into like honestly understanding like how computers are made,
4:58
Speaker0: understanding silicon at a pretty young age and understanding a lot of industries.
5:04
Speaker0: I've just kind of been driven by curiosity my whole life.
5:07
Speaker0: I was really obsessed with the chemical industry as a kid.
5:10
Speaker0: I'm happy to talk about the chemical industry today to me. It's honestly one
5:13
Speaker0: of the most fascinating industries in the world.
5:15
Speaker0: It's five trillion a year in sales. No one knows anything about it.
5:18
Speaker0: It's bigger than oil and gas. It's where everything comes from.
5:21
Speaker0: But it's just like I'm giving that as an example that I have a hundred anecdotes like that.
5:25
Speaker0: It's just kind of being really interested in different things and spending time
5:29
Speaker0: going pretty deep at a young age.
5:31
Speaker0: And you Now, as an investor, I get to kind of invest in those things.
5:38
Speaker0: Basically, I get to invest in things that were childhood passions.
5:41
Speaker0: And so I can make chemicals instance.
5:43
Speaker0: I'm an investor in a chemicals marketplace called Node, K-N-O-W-D-E.
5:47
Speaker0: And almost every investment I've made at Sequoia can be traced to some deep childhood obsession.
5:55
Speaker0: Like with SpaceX, I was completely obsessed with space from a very young age.
5:58
Speaker0: I started off at the age of nine being obsessed with the solar system, but
6:01
Speaker0: then like starting to learn about rockets and you
6:05
Speaker0: know like taking that all the way to having a failed space launch company in
6:09
Speaker0: my early 20s but kind of i think that's what people need to understand about
6:14
Speaker0: me is that the investments i make now like it's not coming from two years of
6:19
Speaker0: thinking about something it's coming from you know 20 to 30 years i.
6:24
Speaker1: Think that that background is really helpful because the reason why we want
6:27
Speaker1: to get you on today is just really just to talk about all the frontier technologies
6:30
Speaker1: that really seem to be maturing seemingly all at once.
6:34
Speaker1: People often joke that the world has not been the same ever since Harambe was killed in 2016.
6:39
Speaker1: But I think I want to actually take that a little bit more seriously and fast
6:42
Speaker1: forward to the world has not been the same ever since ChatGPT 3.5 came out in 2022.
6:47
Speaker1: Because now just a few years later, Elon is catching rockets flying out of the air.
6:52
Speaker1: We have actual robots walking among us. I just saw a video today of a robot
6:57
Speaker1: boxing league coming out of China.
6:59
Speaker1: So that's pretty crazy. Driverless cars are actually in production, driving people around.
7:03
Speaker1: And then now also, of course, we have people talking about actual AGI coming in in two years or less.
7:08
Speaker1: So it just feels like this current moment of time seems especially unique.
7:13
Speaker1: And it seems like you've been thinking about a lot of these things as they have been maturing.
7:17
Speaker1: From your vantage point as an investor in Sequoia, how does this just period
7:21
Speaker1: of history feel for you right now?
7:24
Speaker0: Well-articulated, I mean, this period feels insanely exciting.
7:30
Speaker0: I truly feel like unbelievably privileged and lucky to be an investor in this time.
7:37
Speaker0: I have been very passionate about the history of venture capital,
7:40
Speaker0: you know, for honestly, for over 20 years.
7:44
Speaker0: And I think that if you look in VC, there were kind of like,
7:50
Speaker0: depends how you define it, but like two to four really great periods before
7:53
Speaker0: today. And I think some examples were the early computing era.
7:58
Speaker0: It depends how broad your interests are, but early internet era, then the dot-com era.
8:07
Speaker0: I would say, in my opinion, the early mobile era was incredible,
8:13
Speaker0: with giant companies being created very quickly.
8:16
Speaker0: And now we're in this kind of new era where you have both AI and a resurgence
8:20
Speaker0: of hardware happening simultaneously. And so, I truly feel incredibly lucky
8:25
Speaker0: to be an investor right now.
8:27
Speaker0: And I think that, to build on that, I think the AI stuff is more obvious to the public.
8:34
Speaker0: Like, it's talked about all the time.
8:37
Speaker0: Like, you're, you know, it's kind of incredible.
8:41
Speaker0: And NVIDIA has become, like, a hardware company has become such a popular, hot company.
8:47
Speaker0: But to take a bigger look at this.
8:52
Speaker0: I wrote a manifesto at Sequoia about three years ago. It was like a hardware manifesto.
8:56
Speaker0: And basically what I said in it was, if you look at Sequoia,
9:02
Speaker0: 53-year-old firm now, I was 50 years old when I wrote this.
9:05
Speaker0: For the first 25 years of the firm, Sequoia made almost all of its money in hardware.
9:09
Speaker0: In the last 25 years, we've made almost all of our money in software.
9:13
Speaker0: And the question I asked was like, is hardware dead, long-lived hardware?
9:18
Speaker0: Or was there some weird kind of quirk of history or secular trend of why,
9:23
Speaker0: you know, hardware was not as popular, not as great of a way to make money,
9:28
Speaker0: and that might come back.
9:29
Speaker0: And I'm like, it was a straw man question. I'm strongly in the latter camp.
9:33
Speaker0: And if you want, like right now, I can kind of drill into some of the,
9:37
Speaker0: like, insights I have there.
9:39
Speaker0: But I really think we're entering this dual kind of golden phase of both hardware
9:47
Speaker0: returning in mass at the same time that AI,
9:50
Speaker0: which has both the hardware element
9:52
Speaker0: and the software elements happening simultaneously. So it's pretty wild.
9:56
Speaker1: I think that actually is where we want to go first in this agenda,
10:00
Speaker1: not specifically through hardware, but also discussing manufacturing.
10:03
Speaker1: There's a bunch of subjects that we want to go one by one, starting with hardware and manufacturing.
10:06
Speaker1: We also want to talk about energy, space exploration, everything Elon Musk is
10:11
Speaker1: doing, hardware, human connections, even chemistry, if we can find time.
10:15
Speaker1: But let's start there about manufacturing, because as you said,
10:19
Speaker1: the first 25 years of Sequoia returns were all hardware investments.
10:22
Speaker1: And I think this question of manufacturing is also especially timely,
10:26
Speaker1: given that we are in the midst of this Trump's Liberation Day maneuvers,
10:29
Speaker1: trying to bring back manufacturing back to the United States.
10:33
Speaker1: Maybe just talk about the importance of hardware manufacturing,
10:37
Speaker1: generally speaking, for software, and then also how important it is for here
10:41
Speaker1: inside the United States.
10:43
Speaker3: Yeah, I mean, I'll start with,
10:46
Speaker0: I mean, whatever one's view of politics are,
10:51
Speaker0: like, America found itself in a position over the last 25 years where very little
10:58
Speaker0: of advanced technologies are made in America anymore.
11:02
Speaker0: And this makes it much harder.
11:05
Speaker0: This is one of the reasons why it's been hard to create hardware companies in America.
11:10
Speaker0: I think it was much bigger than that. It's not like the primary thing,
11:12
Speaker0: it's just where we were in Moore's Law and where we were in different big platforms.
11:18
Speaker0: But whenever the hardware supply chain is in a place, it's much easier to innovate
11:24
Speaker0: on the final end product.
11:27
Speaker0: And if you go back, so I'll give an analogy from history.
11:32
Speaker0: I was incredibly lucky as a child, like genuinely just insanely lucky.
11:37
Speaker0: I met this man named Arnold Beckman.
11:39
Speaker0: I met him, he was being pushed in a stroller and he asked me a bunch of science
11:42
Speaker0: questions. It's kind of a crazy story. I was like eight years old,
11:45
Speaker0: or sorry, I was nine years old.
11:47
Speaker0: And this guy, I think Arnold is probably the most underrated person in the history
11:53
Speaker0: of American technology, quite literally. Like I think the most underrated person.
11:57
Speaker0: I'll spend a couple of minutes on this. I think it is like very important anecdote
12:02
Speaker0: to understand kind of the challenges of manufacturing in America today.
12:06
Speaker0: So, basically, Beckman was a professor at Caltech.
12:08
Speaker0: He was a chemistry professor. He invented the... So, he was doing advanced chemistry,
12:13
Speaker0: but he invented the portable pH meter to help farmers measure the acidity of
12:18
Speaker0: oranges and when it was the right time to pick oranges.
12:22
Speaker0: And that ended up becoming a company called Beckman Instruments,
12:25
Speaker0: which ended up, at its height, employing about 2,000 chemistry PhDs.
12:29
Speaker0: And they made many, many different instruments, primarily for the chemicals
12:34
Speaker0: industry and then for early pharma industry.
12:38
Speaker0: And so anyway, so he was doing that. And he was known as one of the apex minds
12:42
Speaker0: at Caltech, and there was an undergrad named Bill Shockley.
12:46
Speaker0: Shockley was a brilliant guy, and he was attracted to Beckman,
12:49
Speaker0: and so he became one of Beckman's kind of protégés.
12:53
Speaker0: And then Shockley went to MIT to do his PhD.
12:57
Speaker0: Then he went to Bell Labs, invented the silicon transistor. When he was at Bell
13:00
Speaker0: Labs, Shockley's a much more common name than Beckman.
13:04
Speaker0: Shockley got the Nobel Prize for inventing the silicon transistor.
13:06
Speaker0: He went, this is the part of the story that people don't know.
13:11
Speaker0: So Shockley ended up going actually back to Caltech after Bell Labs for a couple
13:15
Speaker0: of years to be visiting faculty, kind of figure out what to do next,
13:20
Speaker0: and kind of back to work with his mentor, Arnold Beckman.
13:24
Speaker0: And this is where they decided to start Shockley Semiconductor.
13:27
Speaker0: Something that people don't know has kind of been lost in history is Shockley
13:31
Speaker0: was actually not an independent company.
13:33
Speaker0: It was a wholly owned subsidiary of Beckman Instruments of this chemical instrumentation company.
13:39
Speaker0: And so I started Shockley. Beckman was the chairman.
13:42
Speaker0: You know, they hired brilliant people like Bob Noyce and Gordon Moore and all
13:47
Speaker0: these people that became the Traders 8. They left to start Fairchild.
13:51
Speaker0: Then, you know, Intel spun out of that.
13:54
Speaker0: Beckman continued to be mentor to, like, the Intel founders as they built the company.
14:01
Speaker0: Okay, so here's where we get to the core point.
14:05
Speaker0: For probably the first 30 years or so of building silicon, the hard thing was making wafers.
14:15
Speaker0: Like, it was making very pure wafers and other chemical processes.
14:20
Speaker0: Like, to make a silicon wafer has about 20 chemical steps.
14:23
Speaker0: And then a bunch of the other hard problems early on were all chemical problems.
14:28
Speaker0: In the last 20 years, the hard problem has become lithography.
14:31
Speaker0: It's etching out these very fine, you know, features.
14:35
Speaker0: But for 20 to 30 years, the hardest thing was chemicals.
14:39
Speaker0: And having this guy, Arnold Beckman, that had this insane chemicals company
14:46
Speaker0: with 2,000 genius chemistry PhDs that were always making new equipment,
14:50
Speaker0: having that sit literally right next to Intel meant that whenever Intel was
14:56
Speaker0: running into a problem, kind of scaling the next generation of, like,
15:03
Speaker0: silicon, trying to push, you know, Moore's Law forward.
15:07
Speaker0: They would basically just share in real time, like, hey, we're running into this problem.
15:11
Speaker0: They'd share it with Beckman and his chemists, and they'd be like,
15:15
Speaker0: oh, you know, this might be able to solve it, or if not, they would come up
15:20
Speaker0: with some new instrument that could solve it.
15:22
Speaker0: And having this coupling of the people that were pushing silicon forward,
15:26
Speaker0: sitting right next to people that we're pushing chemicals forward and chemical
15:29
Speaker0: instrumentation and sensors, et cetera, is what led to being able to scale the
15:34
Speaker0: whole system lightning speed.
15:37
Speaker0: And the thing that we've lost in America in the last 25 years is we've lost
15:41
Speaker0: this supply chain. We've lost this coupling, you know, and so you have.
15:45
Speaker3: A couple very unique instances,
15:48
Speaker0: You have Elon, basically, that has gone and into these kind of totally new areas,
15:54
Speaker0: and built his own supply chains, kind of himself, for his own companies.
15:59
Speaker0: And so he's able to have this tight coupling because he's basically doing everything
16:03
Speaker0: within his own companies.
16:07
Speaker0: But for almost every other company, like, if you are a drone company and your
16:14
Speaker0: limiting factor is making motors.
16:16
Speaker0: Like, the motors are being made in China. At least the best ones are being made in China.
16:20
Speaker0: And, you know, it's much harder to talk to the engineers. I mean,
16:23
Speaker0: and now with what's happening in the world, like, we can't even buy motors from
16:27
Speaker0: them for a defense-grade drone.
16:30
Speaker0: But you don't get to go talk to the people building the components and learn
16:34
Speaker0: what's coming one year from now.
16:37
Speaker0: And you don't get to tell them what are your problems and have them go start
16:40
Speaker0: putting them in their roadmap right now.
16:41
Speaker0: And kind of getting that back, getting the whole supply chain back,
16:48
Speaker0: the whole, like, all these different components, it is not something you can do in five years.
16:54
Speaker0: Like, it's not even something you can really do in a decade.
16:58
Speaker0: Maybe in individual verticals, like, if we decide, you know,
17:03
Speaker0: making drones is a core national priority, then we can probably bring the supply
17:09
Speaker0: chain just for drones into America in less than 10 years.
17:13
Speaker0: But if you wanted to do leading-edge chip fabrication.
17:19
Speaker0: I think America would really struggle, even with 100% max effort,
17:24
Speaker0: to catch up to where the cutting edge will be in 10 years over the next 10 years,
17:29
Speaker0: because the supply chain is a few orders of magnitude more complex than for drones.
17:35
Speaker0: And so this is kind of... It was one of the biggest mistakes,
17:39
Speaker0: I think, in business history or in national history for America to let its entire
17:44
Speaker0: supply chain kind of go, starting in the late 90s, to China.
17:50
Speaker0: And I personally think it's very important that we bring it back.
17:55
Speaker0: And I have a bunch of theories and philosophies about how we should do this.
17:59
Speaker0: I think my view is kind of different than how most people are thinking about it today.
18:04
Speaker0: But if your view of the world is that there will never be conflict again,
18:10
Speaker0: like we're at the end of history, then sure, globalization is great.
18:14
Speaker0: Outsource your supply chain, just buy from everyone.
18:16
Speaker0: But if you believe that we're not at the end of history, then I think it's pretty
18:21
Speaker0: important to make at least critical components in America.
18:26
Speaker0: But I think we can't be naive about how hard this will be. And you have to do
18:30
Speaker0: it kind of systematically.
18:32
Speaker2: Yeah, that's what I was super interested in talking to you about,
18:35
Speaker2: is where those constraints actually lie when it comes to manufacturing here.
18:38
Speaker2: What does the optimistic bull case look like for bringing that back onshore?
18:42
Speaker2: And you said that you made this point a little while ago that I really liked,
18:45
Speaker2: which was that every software evolution is preceded by a hardware revolution.
18:49
Speaker2: And I guess I was really interested on where we are on that spectrum of progress
18:53
Speaker2: and where we are on the spectrum of progress and how we're able to get to the
18:57
Speaker2: place where we have this optimistic case for manufacturing.
19:00
Speaker2: It's like the example was like because of the iPhone we got the App Store,
19:03
Speaker2: because of the GPU we have AI mining.
19:05
Speaker2: Are we software constrained or are we hardware constrained? And where are those
19:08
Speaker2: limitations and how do we break through them to bring the manufacturing back to the United States?
19:12
Speaker0: Man, thank you for asking this. What a layup. This is my favorite question.
19:18
Speaker0: I would say there were two parts to it.
19:21
Speaker0: On the point that hardware revolutions precede software revolutions,
19:27
Speaker0: I really strongly believe this.
19:30
Speaker0: I think it's almost a definitional point.
19:33
Speaker0: You know, if you want to have the App Store, which enables creating a bunch
19:38
Speaker0: of, you know, creating Uber and a bunch of these big companies,
19:41
Speaker0: you have to have the iPhone.
19:43
Speaker0: To have the iPhone, you have to have Qualcomm, you have to have Broadcom,
19:47
Speaker0: you have to have 20 years of building out networks, of building out comms,
19:51
Speaker0: technology, almost all of that was on the hardware level.
19:55
Speaker0: If you want to have deep learning, you need to have the GPU,
19:59
Speaker0: that's 20 years of progress.
20:01
Speaker0: If you want to have cloud computing, you have to have very cheap commodity.
20:05
Speaker0: Kind of the whole point of what is a cloud, it's basically the idea that you
20:10
Speaker0: can put a bunch of cheap commodity hardware in one place and then put distributed systems.
20:17
Speaker0: They basically use algorithms on top of this commodity hardware to kind of handle
20:24
Speaker0: failure, handle things breaking in the hardware in a fault-tolerant way,
20:30
Speaker0: which gives you a giant cost advantage over hosting your own hardware.
20:36
Speaker0: And basically, we had to have 20-plus years of memory and CPUs,
20:42
Speaker0: et cetera, getting cheaper to do that. You know, part of why VR is not better
20:48
Speaker0: right now, I mean, this is what people talk about.
20:50
Speaker0: The hardware is not that great, and it's gotten a lot better,
20:53
Speaker0: and the Vision Pro is a lot better, but we're still not there.
20:57
Speaker0: I, with AI, I personally, and I think this is much more obvious now,
21:02
Speaker0: but I've been saying this for many years, I still think we're wildly hardware limited right now.
21:09
Speaker0: And so even if we have AGI in two years, the number of units of AGI we would
21:17
Speaker0: have would probably be very small.
21:19
Speaker0: In a hypothetical world, let's say it takes a gigawatt data center to have one
21:23
Speaker0: AGI. you know, we won't have that many gigawatt data centers at that point.
21:29
Speaker0: And for any gigawatt data center, you can probably do the same thing with, call it 10 megawatts.
21:37
Speaker0: Like 100x more efficient, not even on the algorithm side, purely on the hardware level.
21:43
Speaker0: Like if you start using more silicon photonics, like optimized chips,
21:49
Speaker0: just way better interconnects, these types of things.
21:53
Speaker0: You know, you can probably get like a 100x improvements simply just by better hardware.
21:59
Speaker0: Anyways, I think that AI is going to be limited by hardware for the next five to 10 years.
22:05
Speaker0: And don't get me wrong, I think we're going to make unbelievable progress on
22:07
Speaker0: the software side as well in parallel.
22:10
Speaker0: This is one of these beautiful periods where like the software and hardware
22:13
Speaker0: are co-evolving with each other.
22:16
Speaker0: But the software basically always evolves faster than hardware.
22:20
Speaker0: So anyways, on a much bigger level, I think we basically hit the end of a bunch
22:29
Speaker0: of software revolutions more or less at the same time.
22:35
Speaker0: And so it didn't really leave many places to go.
22:38
Speaker0: We hit the end of the mobile revolution roughly at the same time that I think
22:42
Speaker0: we hit the end of the traditional cloud revolution. I'm not talking AI.
22:47
Speaker0: AI is the one area of software that I think is kind of at the beginning of a
22:52
Speaker0: giant, giant revolution.
22:56
Speaker0: Powered by hardware though and i will just say even in ai hardware like what
23:01
Speaker0: we're seeing right now wouldn't have been possible if it weren't for nvidia
23:07
Speaker0: buying melanox you know or melanox existing,
23:11
Speaker0: independently if not if nvidia was still able to use it this thing melanox is
23:15
Speaker0: an internet connect company it's actually an israeli company sequoia was the
23:18
Speaker0: seed investor it's acquired by for eight by NVIDIA five or six years ago, I can't remember when.
23:24
Speaker0: And basically, Mellanox is what lets NVIDIA have good, full systems,
23:29
Speaker0: not just single GPUs, but full systems and ultra-high bandwidth in these systems.
23:35
Speaker0: And it's probably the real reason why NVIDIA has a moat over AMD and Intel today was Mellanox.
23:43
Speaker0: But there was a lot of progress on the hardware side that's enabling the AI revolution.
23:48
Speaker0: We're seeing right now, kind of on the first question, which was how do we bring
23:54
Speaker0: back manufacturing to the West? What would that look like?
23:57
Speaker0: I think we should all learn from the best examples we've seen.
24:01
Speaker0: To me, the two companies in America doing manufacturing the best right now are SpaceX and Tesla.
24:08
Speaker0: I'm wearing a SpaceX hat right now. Nice. Amazing company, my favorite company in the world.
24:13
Speaker0: I'm sorry, I'm not supposed to take favorites, but come on, SpaceX.
24:16
Speaker0: Likewise, badass. It's just an incredible company.
24:20
Speaker0: But basically, look, I think SpaceX is one of the only...
24:25
Speaker0: I think it's maybe the only company I can think of, along with actually Mellanox,
24:32
Speaker0: that is outperforming China across the board.
24:36
Speaker0: Even Tesla is facing very real competition from Chinese auto companies,
24:42
Speaker0: and there's certain areas where Tesla's ahead, There are certain areas where
24:45
Speaker0: the Chinese EV companies are ahead.
24:47
Speaker0: But SpaceX is very far ahead of every Chinese rocket company.
24:53
Speaker0: SpaceX should not take this for granted. They will never take it for granted.
24:56
Speaker0: They work at maximum speed.
24:58
Speaker0: But I think we should learn from the fact that these two companies were able
25:03
Speaker0: to build supply chains, build giant manufacturing companies in America that
25:08
Speaker0: are the best in the world in their categories.
25:10
Speaker3: And I think, basically,
25:12
Speaker0: The lesson that I would take away is, you know, they started off with,
25:18
Speaker0: initial vehicles that were like, that had a market, but were like a low enough
25:25
Speaker0: level of complexity that a company could do it.
25:27
Speaker0: So like the Roadster for Tesla or the Falcon 1 for SpaceX, which was,
25:31
Speaker0: you know, Falcon 1 was not a great rocket, but proved that they can do it.
25:35
Speaker0: And then it became the Falcon 9, which was like an okay rocket.
25:39
Speaker0: Before the Falcon 9 could be reused, like it was still not at the state of the
25:45
Speaker0: art, you know, in the history of rocketry. Like, there were probably some better
25:49
Speaker0: rockets than the Falcon 9 Expendable.
25:52
Speaker0: But by starting with these things and vertically integrating along the way,
25:57
Speaker0: it put these companies in a position to where then, when they got to,
26:00
Speaker0: like, the Falcon 9 Reusable, or call it the Model S, they had actually intersected
26:06
Speaker0: and become, like, best in the world in their classes.
26:09
Speaker0: And they controlled their own supply chain.
26:12
Speaker0: And so then, as the companies scaled, they could keep investing in the supply chain.
26:16
Speaker0: You know, SpaceX announced recently that with Starlink user terminals,
26:21
Speaker0: they're now even making their own PCB, you know, doing their own PCB assembly.
26:28
Speaker0: Which is a much easier problem than, say, like making a leading edge chip.
26:33
Speaker0: But we haven't been doing much of this in North America recently.
26:37
Speaker0: And so, like, the fact that you start off as a rocket company and it leads all
26:40
Speaker0: the way to be doing, like, PCB assembly is just very bullish for manufacturing in America.
26:47
Speaker0: And then once you have a big factory doing PCB assembly, you know,
26:51
Speaker0: the people that work at SpaceX doing that, someday some of them will leave,
26:54
Speaker0: and they'll actually know how to do this, and they'll probably start doing...
26:58
Speaker0: Maybe they'll start a company that just does PCB assembly for other American
27:02
Speaker0: companies, and just these things catalyze entire supply chains and then ecosystems beyond that...
27:10
Speaker0: And so anyways, for me, I think a giant mistake would be America trying to intersect
27:18
Speaker0: Chinese manufacturing in areas that are already mature or already on exponential trends.
27:26
Speaker0: I think we have to go try to find what are the next areas, like what are the
27:30
Speaker0: things that are relatively new where we can go kind of invest in them now and
27:34
Speaker0: get ahead and then build vertically integrated supply chains there.
27:39
Speaker0: And we have like roughly two today in, you know, electric vehicle manufacturing
27:43
Speaker0: with Tesla and SpaceX with Space Launch.
27:47
Speaker0: But I think we need like 10 of these.
27:51
Speaker0: And honestly, I think that if America had, I think it's as simple as if you
27:55
Speaker0: have roughly 10 SpaceX's and Tesla's that are vertically integrated,
28:01
Speaker0: primarily manufacturing in the West, that alone will basically make all of manufacturing,
28:06
Speaker0: That's like enough to revitalize all of manufacturing in the West.
28:10
Speaker0: And so to take this like one step further, I am unbelievably bullish on a 20-year
28:17
Speaker0: time frame on silicon photonics.
28:19
Speaker0: So almost everything that we use today, you know, with silicon are silicon electronics.
28:25
Speaker0: So, you know, a GPU is, it's electronics.
28:30
Speaker0: Like the computation is happening via electrons, not photons.
28:35
Speaker0: You know, almost every sensor we use is an electronic sensor.
28:40
Speaker0: So, like, a lot of the sensors in the iPhone, etc., are, like, electronic sensors.
28:45
Speaker0: Almost every... Anyways, like, CPUs are electronic. Almost everything is electronic.
28:53
Speaker0: Silicon Electronics is so far along in its manufacturing curve that getting
28:59
Speaker0: to the leading edge is very hard.
29:02
Speaker0: I mean, like, hundreds of billions of dollars minimum, plus 10 years at a minimum,
29:09
Speaker0: to even have a chance of getting to the leading edge. And I think even there,
29:14
Speaker0: it probably wouldn't catch up.
29:15
Speaker0: Silicon Photonics, in my opinion, is just now, like, just now getting to the
29:22
Speaker0: point where the first real commercial applications have started to happen.
29:26
Speaker0: And once you have first real commercial applications of something,
29:29
Speaker0: that's when you get this feedback loop where you start to get revenue,
29:34
Speaker0: which causes investment, and that drives the cost down.
29:38
Speaker0: Because now you're building fab facilities that are optimized for photonics, not for electronics.
29:42
Speaker0: And then as the cost goes down, you can make other products.
29:45
Speaker0: And then the supply chain gets better, and we get better at making reliable
29:49
Speaker0: lasers that go on chip, which are a prerequisite for Silicon Photonics.
29:54
Speaker0: I mean, I could talk for a long time about Silicon Photonics,
29:57
Speaker0: but I think Silicon Photonics will be as important in the 21st century as Silicon
30:03
Speaker0: Electronics was in the 20th.
30:05
Speaker0: And I think that no one is ahead right now.
30:08
Speaker0: Like, this is something where America can easily win in Silicon Photonics manufacturing
30:13
Speaker0: if we're to kind of do it the right way and be smart about it right now.
30:18
Speaker0: That said, both China and Taiwan, they feel the same way.
30:25
Speaker0: And I think they've identified in the last, really just in the last 18 months,
30:29
Speaker0: silicon photonics as critical technologies for the next two decades.
30:35
Speaker0: And so they're now just starting to very heavily invest in it.
30:37
Speaker0: And I'm going to take this a little further.
30:42
Speaker0: These things are of national strategic importance. And so with TSMC,
30:47
Speaker0: until about 18 months ago, TSMC was willing to, so when you make a chip,
30:54
Speaker0: there's many steps to it.
30:56
Speaker0: You know, one step is fabricating the, you know, the device.
31:02
Speaker0: Another step is packaging it, which is kind of when you take the chip or whatever
31:07
Speaker0: and you kind of more or less like put the hood on it and put all the wires that
31:11
Speaker0: come in and out of the chip and you test and make sure that everything is kind
31:15
Speaker0: of like, of working properly and.
31:18
Speaker0: A few other steps in there. Until about 18 months ago, TSMC was willing to package.
31:25
Speaker0: So packaging is incredibly hard. Packaging is probably harder than fabricating the chips.
31:30
Speaker0: And most people don't understand this, but packaging is probably harder than fabricating the chips.
31:36
Speaker0: And TSMC was willing to package silicon photonics chips made by other fabs.
31:43
Speaker0: Like if you had a chip made by Global Foundries, TSMC was still willing to package
31:48
Speaker0: it, because TSMC is the best in the world at packaging chips.
31:52
Speaker0: But in the last 18 months, they basically stopped packaging silicon photonics
31:57
Speaker0: chips made by other companies.
32:00
Speaker0: So you have to now do everything with TSMC, which basically makes a strong incentive
32:06
Speaker0: for the leading companies to only work with TSMC, which makes it hard for any other company,
32:13
Speaker0: to kind of get inroads in silicon photonics.
32:16
Speaker0: And so I think this is like highly symbolic of how strategic Taiwan and TSMC
32:23
Speaker0: view, which are deeply intertwined, view silicon photonics.
32:26
Speaker0: For instance, I think that's smart of them. Like, I don't blame them.
32:30
Speaker0: I think that's brilliant. Like, great leadership by Taiwan.
32:32
Speaker0: They're playing the game the right way.
32:35
Speaker0: But it's just, I think, real confirmation of my point of how important silicon photonics is.
32:41
Speaker0: And this is one area, like, if America can stay at the cutting edge of space,
32:47
Speaker0: stay, like, you know, get to the cutting edge of silicon photonics, you know.
32:52
Speaker0: Stay at the cutting edge of EVs.
32:58
Speaker0: If Tesla and then others can basically get to the cutting edge of humanoid robotics.
33:06
Speaker3: Is something that, like,
33:07
Speaker0: I strongly believe is coming. And I think Tesla... Tesla's very well-suited
33:13
Speaker0: to manufacture these things at scale.
33:16
Speaker0: And, like, I hope the space... I hope there are many winners in America.
33:18
Speaker0: I hope it's, like, a highly competitive space with many winners.
33:24
Speaker0: But manufacturing things at scale is really, really hard.
33:27
Speaker0: I think Tesla's very well-suited for making humanoid robots.
33:31
Speaker0: Just as a, like, I'll kind.
33:34
Speaker3: Of end this
33:35
Speaker0: Monologue in one second, but something that I think is oftentimes missed when
33:42
Speaker0: people look at hardware companies, well,
33:45
Speaker0: basically, I think hardware companies, so with software companies,
33:49
Speaker0: software companies usually get harder to build.
33:53
Speaker0: The bigger you get, like, it gets harder to acquire customers once you,
33:57
Speaker0: because there's not, it's not that hard to replicate the code of a software company.
34:01
Speaker0: Once you prove that a market's good, competitors will come in,
34:04
Speaker0: and then it becomes harder to acquire customers and all these things.
34:07
Speaker0: With hardware companies, it's really hard to get them to work.
34:11
Speaker0: But once they work, they actually generally get easier with scale.
34:14
Speaker0: There's many reasons for this. It's insanely hard to get the supply chain working.
34:19
Speaker0: It's very hard to make the thing.
34:21
Speaker0: It's very hard to work out all the kinks in the system.
34:26
Speaker0: There's usually a baseline number of things you have to build to get the system to work.
34:31
Speaker0: But once you have a working... If you look at basically every hardware company
34:36
Speaker0: in history, almost every hardware company has produced many of their own products organically.
34:42
Speaker0: You know, Apple has so many products. Tesla has many products.
34:46
Speaker0: NVIDIA has many products. Broadcom has many products. Qualcomm has many products.
34:49
Speaker0: Like, any hardware company has lots of products.
34:52
Speaker0: And so, if you use Apple as an analogy, you know, Apple isn't fabricating their
34:57
Speaker0: own devices, but they are designing them.
35:00
Speaker0: The Vision Pro reuses a huge number of components from the iPhone.
35:06
Speaker0: So like the, you know, face ID is reused.
35:10
Speaker0: A bunch, you know, the accelerometers are reused. The magnetometers are reused.
35:15
Speaker0: Like I think I may be wrong on this, but I think the Vision Pro has about a thousand sensors in it.
35:19
Speaker0: And some of those are like, you know, 10 of the same sensor.
35:23
Speaker0: But I think my estimate is something like 950 of the thousand were reused like
35:31
Speaker0: ported over from other Apple products and there's called 50 that were like net
35:36
Speaker0: new designed for the Vision Pro,
35:38
Speaker0: this is such an advantage of hardware companies like once you get going of getting
35:43
Speaker0: to reuse other products and this is something where for Tesla with Optimus like
35:47
Speaker0: I think quite there's there's There's a lot of things that get to be reused from,
35:52
Speaker0: and even if not of specific components,
35:56
Speaker0: the processes or supply chain to make them is basically the same supply chain as making cars.
36:00
Speaker0: And so, it's like, by starting off, by getting really good at making cars in
36:04
Speaker0: America, it can lead to making many other products, and even reusing a lot of
36:08
Speaker0: the components in there.
36:10
Speaker0: This is kind of like, my core thesis is, we have to have about 10 SpaceX Teslas,
36:15
Speaker0: and that's how you really bring manufacturing back to America.
36:20
Speaker2: Man. And there it is. I personally, I love the monologue.
36:22
Speaker0: That was great. I think you're really.
36:24
Speaker2: Good at explaining things on first principles. So I kind of want to apply that
36:27
Speaker2: to the next category, which is how do we power all of these things that you just spoke about?
36:33
Speaker2: It's how do we power the robotics, the electric cars, the actual manufacturing
36:36
Speaker2: and factories that are required to build all of these things?
36:39
Speaker2: Do we have the energy capability now to even do that?
36:43
Speaker0: Or are there key.
36:43
Speaker2: Unlocks that need to happen on the way before we could start making things at
36:47
Speaker2: the scale that you're describing?
36:49
Speaker0: I mean, another great question. I'm very happy that America has woken up to,
36:57
Speaker0: like, just how fundamental energy is.
37:00
Speaker0: And this was kind of lost from the conversation for a decade or so,
37:05
Speaker0: as, like, climate change took priority over kind of growth.
37:12
Speaker0: And I understand that philosophy. but GDP is just extremely correlated with power production,
37:24
Speaker0: electricity generation, and we really stopped investing in power generation in America.
37:34
Speaker0: I don't mean fully stop. We've had a lot of growth, but we have not been investing
37:39
Speaker0: in it anywhere near the same, to the same level as China.
37:44
Speaker0: I think probably, like,
37:48
Speaker0: if America were to fail in, like,
37:54
Speaker0: manufacturing resurgence, I think it would be because of just power constraints,
37:59
Speaker0: just not having low enough power costs and not having enough just power generation.
38:04
Speaker0: So, I think this is the most important question.
38:08
Speaker0: And how do we get there?
38:10
Speaker0: I mean, look, so obviously we need to start...
38:15
Speaker0: We need to invest in things that will be good long-term power generation sources.
38:19
Speaker0: But I think the only thing that lets you, like, really have lots of power generation
38:25
Speaker0: in the short term is using fossil fuels.
38:27
Speaker0: And it's kind of, this is maybe an unpopular view, but I basically view the
38:34
Speaker0: world as like there's peacetime and there's wartime and kind of different,
38:39
Speaker0: you should have different optimization during peacetime versus wartime.
38:43
Speaker0: You know, obviously there's hot wars, things like World War I,
38:47
Speaker0: World War II, where a lot of American industry got devoted towards making,
38:52
Speaker0: you know, equipment for the wars.
38:56
Speaker0: But then you have the Cold War, which is something slightly different where,
39:00
Speaker0: you know, this is, again, an unpopular view, but you have a different set of rules during war.
39:10
Speaker0: And personally, I'm glad that America fought the Cold War.
39:16
Speaker0: I think there were many blunders during the Cold War.
39:20
Speaker0: You know, like, many aspects of the Vietnam War were giant mistakes and stupid.
39:27
Speaker0: But I think it would have been a much worse outcome for the world if.
39:31
Speaker3: America would have just
39:31
Speaker0: Ceded the world to the Soviet Union and basically let communism spread.
39:37
Speaker0: So I'm generally very happy that we did this.
39:40
Speaker0: I personally think we're going through a very similar thing right now with China.
39:45
Speaker0: It's Cold War 2.0, and...
39:49
Speaker0: China is not playing by their, like, you know, they have a wartime attitude.
39:57
Speaker0: And I think it would be foolish for the West to not simultaneously have a wartime attitude.
40:05
Speaker0: And kind of an end byproduct of this is, I think that we need to be investing
40:11
Speaker0: like crazy in fossil fuels right now.
40:14
Speaker0: Because I think the stakes are too big. Like, I think, like,
40:20
Speaker0: liberal democracy is at stake. That's my personal view.
40:23
Speaker0: I'm happy, like, I'm happy. A lot of people disagree with me on that.
40:27
Speaker0: And I don't think it's as clear. I think that China has been very strategic
40:31
Speaker0: with not making this as obvious as it was during the Cold War.
40:35
Speaker0: I think they learned from Cold War 1.0 and some of the mistakes made by the Soviets.
40:40
Speaker0: But I think basically in the short run, we have to really deregulate fossil
40:46
Speaker0: fuels. Granted, we should do it as clean as possible, and this is something
40:53
Speaker0: where there's a pretty big narrative violation.
40:55
Speaker0: If you look at the CO2 emissions of a clean coal power plant versus a dirty
41:01
Speaker0: coal power plant, you can have wildly more efficient coal power plants.
41:07
Speaker0: And hopefully we don't use too much coal, but things like natural gas,
41:12
Speaker0: we should be using huge amounts of gas right now.
41:16
Speaker0: But at the same time, we have to be investing in much better,
41:19
Speaker0: longer-term technology.
41:21
Speaker0: And to me, that's like solar is unbelievably important.
41:27
Speaker0: Solar is basically at a point right now where the price per watt,
41:33
Speaker0: like nothing will ever catch up to the cost of solar.
41:36
Speaker0: Or sorry, nothing in the next, nothing that is even close to working today.
41:41
Speaker0: Like fusion could get there, but with uncertain timeline.
41:45
Speaker0: Nuclear, I don't think, can catch up to solar is my kind of personal opinion
41:51
Speaker0: on this full systems level given where solar is now.
41:54
Speaker0: If we were to go back in time, if we had a time machine, we went back 30 years
41:57
Speaker0: and we deregulated nuclear 30 years ago and let nuclear go down an exponential
42:01
Speaker0: cost curve starting 30 years ago, I think it would probably be way ahead of solar today.
42:07
Speaker0: And I don't think solar would have even gotten to where it is now because we
42:12
Speaker0: wouldn't have needed it. nuclear would have been so good.
42:15
Speaker0: But I still, I think America should be investing very heavily in nuclear,
42:19
Speaker0: even though it won't be as low of a cost as solar.
42:25
Speaker0: One, I think China has basically a lock on large-scale solar manufacturing for
42:33
Speaker0: at least the next decade or so.
42:36
Speaker0: And they have, they basically control the supply chain of a lot of the.
42:41
Speaker0: The components that are required to make solar panels.
42:45
Speaker0: And so just from a strategic perspective, I don't think we can only rely on solar.
42:52
Speaker0: And so I would be investing in nuclear.
42:54
Speaker0: To have nuclear be even reasonable cost, you basically have to build at least 100.
42:59
Speaker0: You have to build a bunch of nuclear power plants using the same design,
43:03
Speaker0: basically the same construction crew, et cetera, where you get economies of
43:06
Speaker0: scale and people kind of learn what they're doing on the first one.
43:11
Speaker0: And then do it more efficiently by the fifth and then tenth one.
43:15
Speaker0: I would love to see America, you know, start building another hundred nuclear
43:19
Speaker0: power plants, if nothing else, for the strategic significance.
43:25
Speaker0: But in the immediate term, I think we need to really invest more in natural
43:29
Speaker0: gas and other fossil fuels.
43:31
Speaker3: Awesome.
43:32
Speaker2: Yeah, I was super curious in the dynamic between fossil fuels, nuclear, solar.
43:36
Speaker2: Which one will be more important on what timescale? It sounds like solar is
43:40
Speaker2: the best medium term prior to getting nuclear, which leads me to a fun question
43:44
Speaker2: that I'm just super interested in because it seems too good to be true.
43:47
Speaker2: You're invested in this company called Reflect Orbital, which is literally planning
43:51
Speaker2: to take rockets and send satellites up into space and reflect them down onto
43:54
Speaker2: solar farms at night. Is this possible?
43:57
Speaker2: What are the dynamics of this? Is this where solar is headed to?
44:01
Speaker0: So it's a complex thing to answer.
44:04
Speaker0: So, one of my lessons of being an investor now for basically a decade and being
44:11
Speaker0: in startups for 20 years is, so I'll use an analogy from Varda.
44:18
Speaker0: So did you guys know Delian's company, Varda?
44:21
Speaker2: Absolutely, yeah.
44:23
Speaker0: So, I love Delian, he's a character, smart guy. I passed on the Varda seed.
44:29
Speaker0: And I have no idea if they'll make it as a company. Like, it's always impossible
44:34
Speaker0: to predict these things.
44:35
Speaker0: But I didn't like their initial idea.
44:39
Speaker0: Their initial idea was trying to make, you know...
44:42
Speaker0: So I love the general idea of build a space company.
44:46
Speaker0: Because I just thought the timing for it was very good, that launch costs over
44:50
Speaker0: the next decade are going to drop dramatically,
44:52
Speaker0: and mass orbit is going to increase dramatically, and that, like,
44:55
Speaker0: a really smart team will figure out some way to, like, by assembling a really
45:00
Speaker0: smart team, they'll figure out some way to profit from that.
45:03
Speaker0: But their initial idea was making this thing called ZBLAN.
45:06
Speaker0: ZBLAN is ultra-pure, like, fiber-optic cables.
45:09
Speaker0: It's especially good for really long-range, like, data transmission.
45:14
Speaker0: Like, I just personally don't think ZBLAN is that important.
45:18
Speaker0: I think it'll be a shrinking market, especially with Starlink.
45:21
Speaker0: Like, Starlink is just the best way to send data over long distances.
45:27
Speaker0: And so I passed Von Varda because I basically felt like they had the right macro
45:31
Speaker0: idea around like building a space company now, assembling the talents.
45:35
Speaker0: But I felt like they had the wrong micro idea of ZB LAN as a starting point.
45:41
Speaker0: And a lesson for me from that was like, if you think that the macro on something
45:45
Speaker0: is unbelievably good and the team is insanely good,
45:50
Speaker0: even if the, and it's a seed investment, even if the micro idea is,
45:55
Speaker0: you know, you're not fully certain about it,
45:57
Speaker0: then you should still make the investment and that an unbelievable team will, like,
46:02
Speaker0: take advantage of the macro tailwind.
46:04
Speaker0: And so for Reflect, it's a very similar thesis for me.
46:07
Speaker0: The, again, I think.
46:12
Speaker0: Almost everyone's underestimating what the space economy is going to look like in a decade.
46:18
Speaker0: Like, in 10 years, people's minds are going to be blown.
46:22
Speaker0: You know, like, we are certainly going to have some infrastructure on Mars.
46:29
Speaker0: You know, like, Elon's talking about, you know, there's a Mars launch window every 26 months.
46:34
Speaker0: The next one is end of next year. SpaceX is trying to send some,
46:41
Speaker0: like, uncrewed by human, but crewed by Optimus starships to Mars,
46:47
Speaker0: then I hope they make it. I think they have a real shot, but I don't think it's guaranteed.
46:51
Speaker0: But I think in 2028, in the next window after 2026.
46:56
Speaker3: I think it's
46:56
Speaker0: Basically guaranteed that SpaceX sends some starships to Mars.
47:02
Speaker0: And I think in the window after that, which I think would be like early 2031,
47:07
Speaker0: I think it's basically guaranteed that SpaceX starts to send some like,
47:12
Speaker0: forward cargo and stuff to Mars, a useful cargo, whether it's solar panels,
47:18
Speaker0: you know, sending stuff that will be infrastructure for humans to go eventually.
47:25
Speaker0: And by the time you get to 2035, I think we're going to have...
47:32
Speaker0: I just think people are going to be shocked at the amount of masks going to
47:37
Speaker0: orbit, the amount of money made in space, You know, Starlink is one of the fastest
47:44
Speaker0: growing products in history in terms of revenue.
47:47
Speaker0: If you fast forward 10 years, I'm very confident Starlink will have over $100 billion of revenue.
47:55
Speaker0: I think there will be other space businesses that have lots of revenue.
47:58
Speaker0: And there's only like five credible teams, in my opinion.
48:04
Speaker0: There's a bunch of teams, but there's like five that I think are heads and shoulders
48:08
Speaker0: above the rest. that are really kind of building towards this feature right now.
48:13
Speaker0: And whatever, I can't tell you exactly what they'll do that will be valuable,
48:20
Speaker0: but I think just learning how to actually do things in space and assembling
48:26
Speaker0: teams that are absolutely cracked, that know how to do things in space right now,
48:32
Speaker0: I think it's like buying land next to the railroad stops.
48:38
Speaker0: Like, when you knew, like, you already knew that they're being constructed.
48:41
Speaker0: You knew the railroad lines were being constructed, and you knew where the stops
48:44
Speaker0: were, and you go and buy the land at the stops.
48:47
Speaker0: And you maybe don't know exactly how you're going to develop it,
48:49
Speaker0: but you sure as hell know it's going to be really valuable.
48:52
Speaker0: And so that's kind of how I view, honestly, that's how I view Reflect.
48:56
Speaker0: Like, I bought gold, like, A-plus real estate.
49:02
Speaker0: It's like I bought the real estate right next to the San Francisco station at
49:07
Speaker0: the end of the Trans-Pacific Railroad.
49:12
Speaker1: Listeners broadly, the base case for maybe not just listeners,
49:15
Speaker1: because the bankless audience tends to be pretty optimistic and curious about their future.
49:19
Speaker1: But the average person in the United States, when they hear about how much investment
49:24
Speaker1: or how hard we are going to invest in getting into space, their first reaction
49:30
Speaker1: might be, how does that benefit my life? And we have space.
49:34
Speaker1: We have, God, I'm forgetting the name of the satellite, the Wi-Fi,
49:37
Speaker1: Elon's Wi-Fi satellite. Starlink.
49:38
Speaker0: Starlink.
49:39
Speaker3: Okay. And we have Starlink,
49:41
Speaker1: Which is a very awesome, very strong example. Like I've run into businesses
49:45
Speaker1: in like, you know, United States national parks where there's no Wi-Fi and they
49:48
Speaker1: have a Starlink so they can actually hold, maintain their business.
49:51
Speaker1: It's great. But that's just like one example.
49:54
Speaker1: I think understanding that Elon is trying to like reduce the costs of getting things into state.
50:00
Speaker1: He's trying to bring it down to zero, trying to get it as close to zero as possible.
50:03
Speaker1: Like all technology, people are still having a hard time imagining how that
50:07
Speaker1: benefits them, like how, how that impacts their life.
50:10
Speaker1: And I know maybe Sean, like the takeaway from what you just said is like, Oh, I don't really know.
50:14
Speaker1: I just know that some teams are going to figure it out. Therefore I'm buying
50:17
Speaker1: the real estate, but are there any, like any things you could leave the listener
50:21
Speaker1: with to go explain to the rest of the world?
50:23
Speaker1: Like why like fundamentally cheap space travel will unlock so many things that
50:28
Speaker1: will actually turn into quality of life improvements for the average person
50:32
Speaker1: who's left on planet earth
50:33
Speaker0: I'll say a few things first of
50:37
Speaker0: all this is not directly into your question it's an analogy to
50:40
Speaker0: the past but like what's you
50:44
Speaker0: know people have this mental model that space is very hard
50:48
Speaker0: and expensive and space is very hard and will always be hard but space didn't
50:55
Speaker0: make much progress for about 30 to 40 years and when a field is stalled for
51:01
Speaker0: like 40 years it gives people bad intuition when the field starts to go through
51:05
Speaker0: exponential progress because like you're,
51:08
Speaker0: overweighting the stalled period if you look at the cost per
51:11
Speaker0: kilogram to orbit it's basically stalled for like 30 years around 100 000 a
51:15
Speaker0: kilogram spacex is going to push that the starship when it's fully reusable
51:21
Speaker0: down to like 20 a kilogram this is like a 50 000 x reduction in cost,
51:29
Speaker0: whenever in history, something like that has happened.
51:36
Speaker0: And it's not just the cost, it's also the amount of stuff you can put in space.
51:41
Speaker0: It's been very limited in the past, but we're going to be able to put up,
51:45
Speaker0: call it a million times in 10 years, like a million times more stuff per year
51:51
Speaker0: than we were able to put up, you know, 20 years ago.
51:57
Speaker0: And when those two things happen, like cost goes down by 50,000x and amount
52:03
Speaker0: of stuff you can send up goes up by a million.
52:07
Speaker0: Literally, the only analogies I can think of that are semi-appropriate are early railroad era,
52:16
Speaker0: where you went from needing to take a horse-drawn carriage across America,
52:23
Speaker0: where you had a very high chance of dying or getting a vicious disease along the track,
52:31
Speaker0: to you could just very easily and comfortably send extremely large amounts of
52:37
Speaker0: stuff very cheaply across America and all the businesses that unlocked and what
52:42
Speaker0: it did to the value of real estate in the West.
52:45
Speaker0: And the other analogy is early shipping and the invention of the ship and then
52:50
Speaker0: discovery of new places and what that did to the global economy.
52:55
Speaker0: I really can't stress how... This is basically the same thing.
53:00
Speaker0: We are going to have railroads to space. We're going from horse-drawn carriages
53:04
Speaker0: to space to railroads to space. That is happening right now and over this next decade.
53:09
Speaker0: And in terms of what the businesses will be, I think that,
53:13
Speaker0: again, I would look at when NASA was at its peak and was still able to take
53:20
Speaker0: risks before the Challenger shuttle disaster.
53:26
Speaker0: They invented, like, when you have this clear goal of, like,
53:30
Speaker0: we need to win the space race, we want to put humans on the moon,
53:33
Speaker0: we wouldn't spend a lot of money to do it.
53:35
Speaker0: There were so many spin-out technologies that came as a result that deeply benefited
53:40
Speaker0: just regular people, whether it's breakthroughs in radio technology,
53:46
Speaker0: or, you know, like the microwave oven, or all sorts of materials that were used
53:55
Speaker0: in many different products, you know, on and on and on.
54:00
Speaker0: As SpaceX builds a city on Mars, whatever you think of the value of that city,
54:08
Speaker0: the technologies that are going
54:10
Speaker0: to go into building that city will be unbelievably valuable on Earth.
54:15
Speaker0: And there's something very ironic that when you try to solve the even harder
54:19
Speaker0: problem where you're more constrained, it's actually in some ways easier to
54:23
Speaker0: solve some hard problems.
54:26
Speaker0: And if you want, I can get into why I think that is. But here's some of the
54:29
Speaker0: things that are going to come from building a city on Mars.
54:32
Speaker0: We're going to have better farming tech than we've ever had before.
54:36
Speaker0: It's basically going to force us to
54:37
Speaker0: figure out how to actually do vertical farming, which no one has like...
54:41
Speaker0: You haven't had SpaceX-level engineers go and try to solve that problem.
54:45
Speaker0: We're going to probably... Like, I think that they'll...
54:50
Speaker0: There's going to be huge breakthroughs in power production. I don't know exactly
54:54
Speaker0: what will make sense on Mars, whether it would be like nuclear or solar,
55:01
Speaker0: but there will be huge breakthroughs in that.
55:04
Speaker0: There will be, like, I actually think there will be huge breakthroughs in electronics fabrication.
55:10
Speaker0: So we'll basically have to build autonomous fabs or very, very low human labor fabs.
55:17
Speaker0: I'm not saying that will happen in 10 years, but I think 30 years from now,
55:21
Speaker0: you know, the goal is to have Mars City be fully self-sustaining,
55:26
Speaker0: which would mean that it has to have a fab, and that will have to be like a
55:29
Speaker0: basically autonomous fab.
55:31
Speaker0: And that fab will probably be even better than the fabs we have on Earth.
55:38
Speaker0: And anyways, when you try to build a fully autonomous city on Mars,
55:45
Speaker0: like it forces you to rethink from first principles almost every technology
55:50
Speaker0: that we consume today, every input down to like, how do you have...
55:54
Speaker0: I'm going to say something really crazy that is the truth.
55:59
Speaker0: We don't even have clean food in America today.
56:05
Speaker3: Like our food, dear America,
56:09
Speaker0: I love you. This is tragic. Our food supply is literally poison.
56:14
Speaker0: Our food is filled with arsenic, lead, and forever chemicals.
56:17
Speaker0: And it's in the soil all over the country.
56:20
Speaker0: And we can't get rid of it. or we haven't built technology yet to figure out how to get rid of it.
56:26
Speaker0: I really only started to understand this over the last six months or so.
56:29
Speaker0: A founder that I work with, who's a very smart guy, like, started measuring,
56:33
Speaker0: like, doing detailed chemical analysis of soil all over America.
56:37
Speaker0: And it's basically, it's really all contaminated. And...
56:42
Speaker0: If on Mars we start figuring out how to make food kind of with,
56:45
Speaker0: in a fully closed system, we can use that same technology to make safe food in America.
56:52
Speaker0: Like, we're going to basically get to the point where we don't even need soil to make food.
56:55
Speaker0: Even though the soil is contaminated, it's okay. We'll be able to,
56:58
Speaker0: you know, make food without soil, you know, and measure what's in it and get back to basics.
57:05
Speaker0: And I know that sounds crazy, but like, the path towards having clean food in
57:10
Speaker0: America might actually be Mars.
57:14
Speaker0: And I hope we figure it out sooner than that, but that's the sad reality of
57:19
Speaker0: where some of our industries are today.
57:21
Speaker3: Yeah, I love
57:22
Speaker2: That as the accessible bull case for Mars. I definitely share your optimism,
57:25
Speaker2: sometimes over a little bit too much in terms of timeline and what we're actually capable of,
57:30
Speaker2: and find us in the minority, because it is so challenging to wrap your head
57:33
Speaker2: around the fact that we actually can place things on another planet within the
57:36
Speaker2: next decade, which is super exciting.
57:37
Speaker2: But there is one part of this that is slightly disturbing in the sense that
57:41
Speaker2: a lot of this has to do with one single person, one single company. And this is SpaceX.
57:46
Speaker2: This is Elon. And this is true in a lot of other categories where they are the
57:49
Speaker2: leader in autonomous robots.
57:51
Speaker2: They are the leader in robo-taxis. They are the leader in a lot of these categories
57:54
Speaker2: that we're talking about that are the future.
57:56
Speaker2: And you are someone who's uniquely positioned to talk about this because of
57:59
Speaker2: your exposure to, what is it?
58:00
Speaker2: I think it's SpaceX, a boring company, XAIX. You are mega Elon supporter,
58:05
Speaker2: mega Elon bull, like bull case for Elon.
58:08
Speaker2: And also why there's no clear first or second best in these industries.
58:13
Speaker2: And where can we find more people like this to build these important projects?
58:18
Speaker0: In my opinion, Elon is by far the best entrepreneur in the world today.
58:22
Speaker0: And almost certainly of all time.
58:26
Speaker0: I think it comes from many places. One is his like radical first principle thinking.
58:31
Speaker0: Another is his, like, psycho-level work ethic.
58:36
Speaker0: You just can't... I don't think anyone should...
58:40
Speaker0: People can't comprehend how hard the guy works unless you've seen him do it.
58:44
Speaker0: Like, most humans would drop dead trying to work as hard as he does.
58:49
Speaker0: And I think he ramped up to that capability over many decades.
58:54
Speaker0: But I actually, so I think Elon is the most important entrepreneur in the world.
59:00
Speaker0: But I actually think most of his companies would be fine today if,
59:06
Speaker0: you know, in the worst case scenario, God forbid, something happened to him, such as if he got sick.
59:11
Speaker0: You know, I think that it would severely impair the long-term potential of the companies,
59:19
Speaker0: but it's kind of like how tragically when Steve Jobs passed away,
59:23
Speaker0: you know, from Apple, like, they had already designed the iPhone,
59:27
Speaker0: they already had a roadmap for the next 10 to 20 years, and other people were
59:32
Speaker0: able to execute on that roadmap.
59:34
Speaker0: Well, I think that Apple hasn't innovated anywhere close to as much as it would
59:38
Speaker0: have with Steve still there.
59:40
Speaker0: But I want to be clear, I actually think that, God forbid, in the worst case,
59:45
Speaker0: if something happened to Elon, I think that Optimus would still happen.
59:49
Speaker0: I think that autonomous vehicles are mature enough, like that would happen.
59:52
Speaker0: I think that Mars would still happen.
59:56
Speaker0: Starship would definitely still happen.
59:59
Speaker0: Starlink would get to absolutely insane scale. The direct-to-sell program would really still happen.
1:00:07
Speaker0: So I just, I actually, I'm kind of of these two minds.
1:00:10
Speaker0: Like, I think humanity would not move nearly as fast as it will with him.
1:00:17
Speaker0: And like overall innovation be way lower.
1:00:21
Speaker0: But actually, I wouldn't underestimate just how far along and mature these different
1:00:26
Speaker0: product lines are today.
1:00:27
Speaker0: And so sorry, the first question was like, or so what was the second part of
1:00:31
Speaker0: the question? What did I miss? How do we find more Elon's?
1:00:34
Speaker3: Yeah, where are the other,
1:00:36
Speaker2: Where are the second best companies? Where are the companies that are also going
1:00:38
Speaker2: to take us to Mars? Is it just one and will it just be one? Or are there other people capable?
1:00:43
Speaker2: And what does it take to build a company like that so that there is more people
1:00:46
Speaker2: working on these hard problems?
1:00:47
Speaker0: I don't think you can create an Elon overnight. I think Elon is...
1:00:54
Speaker0: Elon, as an individual, is the product of 50 years of pain, suffering,
1:01:03
Speaker0: work, compounding all of this.
1:01:07
Speaker0: I really believe that. I also think that something that is underappreciated
1:01:14
Speaker0: about how effective he is today is that over the course of the last 20, 30 years,
1:01:20
Speaker0: he's been able to assemble a group, a collective of, call it,
1:01:25
Speaker0: 20 people around him that are some of the most talented humans on the planet.
1:01:30
Speaker0: That it's like, you only come across one of these people every,
1:01:34
Speaker0: or call it a couple of these people every year.
1:01:37
Speaker0: And it takes a decade of testing someone and testing what they can do to really
1:01:42
Speaker0: learn their true potential and learn that they're reliable.
1:01:45
Speaker0: So if you go look at SpaceX, there are people like Mark Giancosa,
1:01:51
Speaker0: or a Boring Company where there's Steve Davis,
1:01:55
Speaker0: you know, or basically every one of his companies has five or so people that
1:02:02
Speaker0: are just out of this world, talented.
1:02:06
Speaker0: And you can't, and all these people have a mind meld with each other.
1:02:10
Speaker0: They know how to work well.
1:02:11
Speaker0: They know how, like, they don't even need to speak to know what needs to get done.
1:02:17
Speaker0: And you can't rush that. It literally takes, like, 20 years to kind of build
1:02:23
Speaker0: an organism, almost, that is self-sustaining and, like, maximally evolved to be effective.
1:02:31
Speaker0: So, I think to have second Elons,
1:02:37
Speaker0: if you wanted to have one emerge over the next decade, it would have to be someone
1:02:42
Speaker0: that's already like 10 plus years,
1:02:45
Speaker0: into doing really incredible things, to where their own personal rate of learning
1:02:50
Speaker0: has been on this maximum trajectory for 10 years.
1:02:54
Speaker0: And they're kind of like, the leverage they have in the world is on a maximum trajectory.
1:03:02
Speaker0: Leverage, meaning, you know, like, building these insane teams,
1:03:05
Speaker0: building trust with these people, acquiring, kind of, credibility,
1:03:11
Speaker0: acquiring capital, you know, building manufacturing capability, etc.
1:03:16
Speaker0: And so, some candidates, I mean, there are people like John and Patrick Collison,
1:03:20
Speaker0: who are unbelievably talented humans, and they're still quite young.
1:03:27
Speaker0: And, like, I think they love Stripe so much.
1:03:33
Speaker0: And I think they see that, I think, for them, like, Stripe will be the vehicle
1:03:36
Speaker0: through which they do most of their other projects over the course of the lives,
1:03:41
Speaker0: rather than how Elon has started a bunch of different companies.
1:03:44
Speaker0: But, like, I could see John and Patrick, 10 years from now, really surprising
1:03:49
Speaker0: the world with, like, what they're, the scale of what they're able to do.
1:03:54
Speaker0: And leveraging the learning they've had over the last 10 years and the position
1:03:59
Speaker0: they're in now to push an order of magnitude or two beyond where they are now.
1:04:05
Speaker0: But I think you basically have to look at other really, really great founders
1:04:10
Speaker0: or companies that are already quite far along right now to find the people that
1:04:15
Speaker0: are even eligible to kind of be the next Elons in 10 years.
1:04:19
Speaker0: And what happens a lot of times is people, it's really hard to sustain,
1:04:24
Speaker0: like, look at Warren Buffett's net worth, and this is a weird analogy,
1:04:28
Speaker0: look at Warren Buffett's net worth, and it's like only,
1:04:33
Speaker0: only in the last 20 years or whatever that he's actually become so insanely rich.
1:04:38
Speaker0: If you go back to the 90s, he was a billionaire or whatever,
1:04:41
Speaker0: but it was like single-digit billionaire.
1:04:45
Speaker0: We can't underestimate compounding. And Elon is one of the only entrepreneurs
1:04:52
Speaker0: that hasn't stopped his rate of compounding because he's still willing to work
1:04:58
Speaker0: at this maximum intensity.
1:05:01
Speaker0: And it's very hard to do that. It's very, very, very hard to keep going and
1:05:06
Speaker0: work 100-hour weeks every single week and embarrass yourself.
1:05:13
Speaker0: By definition, to be at maximal rate of compounding, the way you really learn
1:05:17
Speaker0: is by making mistakes and to still be willing to embarrass yourself,
1:05:21
Speaker0: take enough risks that you make mistakes, test starship in public where inevitably
1:05:26
Speaker0: a lot of them are going to blow up and the media is going to write bad articles.
1:05:28
Speaker0: Like, it's just exhausting.
1:05:31
Speaker0: And most people get burned out from doing that.
1:05:33
Speaker0: And so, like, to have another Elon or more of them, we need,
1:05:38
Speaker0: it would require some of the people that have compounded at incredible rates
1:05:41
Speaker0: early on to keep going and push through.
1:05:44
Speaker0: And I think that's where Elon's personality of, like.
1:05:49
Speaker0: You know, growing up in South Africa and dealing with, like,
1:05:52
Speaker0: a lot of challenges at an early age and, like, the infinite grit,
1:05:56
Speaker0: and also just probably some singular aspect of how he was born,
1:05:59
Speaker0: how he was raised, you know, like, who his mother is, like, I think all these
1:06:04
Speaker0: things play in, it's like, you need...
1:06:06
Speaker0: You need someone to do enough early on to where they're compound at the maximum
1:06:11
Speaker0: rate, and there's only maybe like one person a year that breaks through,
1:06:16
Speaker0: you know, to like level two of that or whatever.
1:06:20
Speaker0: And to go to level three, you also need kind of these very unique personality
1:06:25
Speaker0: and kind of crazy, challenging life and upbringing.
1:06:29
Speaker0: And so, anyways, I hope we get more. But we should all appreciate like how lucky we are.
1:06:34
Speaker0: And I'm sorry to like get into this because obviously Elon is a very controversial
1:06:40
Speaker0: figure right now but and so you have people like attacking Tesla dealerships and all that and here.
1:06:48
Speaker3: I am like I
1:06:48
Speaker0: Honestly think we should all whatever you think of the man's politics like we
1:06:54
Speaker0: should celebrate his work ethic and like celebrate what he's done to push technology
1:06:58
Speaker0: forward in the world it's like it's really really important Yeah.
1:07:03
Speaker1: I think most listeners are going to be familiar with basically every company
1:07:06
Speaker1: and project that Elon has, except for maybe his most recent foray into robotics.
1:07:12
Speaker1: That's, I think, the most opaque one. That's just the newest one on the scene
1:07:17
Speaker1: that I'd like to dive into a little bit more.
1:07:19
Speaker1: As we're talking about space and colonizing Mars, I would imagine the order
1:07:23
Speaker1: of operations is to not put humans on Mars, but first put robots on Mars.
1:07:27
Speaker1: And we can even see the robots that he's debuting here on Earth,
1:07:31
Speaker1: and it's all very iRobot-y.
1:07:33
Speaker1: It turns out they actually are just looking like actual humans.
1:07:36
Speaker1: Maybe you could just kind of ground us and the listeners. I'm pretty unfamiliar,
1:07:40
Speaker1: Josh is far more familiar than me, about just like the state of robotics as
1:07:45
Speaker1: an industry, why they actually ended up looking exactly like humans.
1:07:48
Speaker1: Is that the way things are going to be?
1:07:50
Speaker1: And just overall, where are we in the robotics arc of technology development?
1:07:55
Speaker0: So on the question of why are robots looking like humans, I think there's two
1:08:01
Speaker0: primary answers for that.
1:08:03
Speaker0: One is that the world has been designed for humans.
1:08:09
Speaker0: And that may sound simple. That's actually incredibly important.
1:08:16
Speaker0: You know, like, humans can go upstairs, no problem.
1:08:20
Speaker0: If you're a wheeled robot, going upstairs is hard. if we were to go start building
1:08:26
Speaker0: new cities for robots, you probably wouldn't put stairs in there,
1:08:29
Speaker0: and then you can have wheeled robots everywhere.
1:08:31
Speaker0: And wheeled robots are easier than robots that have actuating legs.
1:08:36
Speaker0: But we've already built cities for humans. They go upstairs since there's stairs everywhere.
1:08:43
Speaker0: Warehouses are built for humans. And so the height of where things are and the
1:08:48
Speaker0: layouts, they work for humans.
1:08:53
Speaker0: So basically, if you deviate too much from the human form factor,
1:08:59
Speaker0: it actually starts to become, like, all sorts of things become really hard to
1:09:03
Speaker0: do, just simply because of how the physical world was designed,
1:09:07
Speaker0: almost from, like, an architecture perspective, and from a, you know, from a form factor.
1:09:12
Speaker0: And then the second thing is, I think, less important, but it's the psychological thing, which is.
1:09:21
Speaker0: I think humans just react better when robots look like humans.
1:09:29
Speaker0: If you take a robot and you make it some very different form factor,
1:09:37
Speaker0: even the face, if you don't have the eyes, if you make it just a fully black
1:09:43
Speaker0: face, it's fine we can start to...
1:09:47
Speaker0: Like if you put in like three eyes or whatever instead of two,
1:09:51
Speaker0: I think that's kind of scary, honestly, for some people.
1:09:55
Speaker0: And it's just like, it's unnecessary.
1:09:57
Speaker0: Like why, if you don't need to do that, why to just kind of conform to what
1:10:05
Speaker0: people are already comfortable with?
1:10:06
Speaker0: But I think the first reason is way more important.
1:10:09
Speaker2: There is one thing that is absent from the portfolio that I saw that is one
1:10:12
Speaker2: of Elon's ventures that I'm super excited about and also a bit dystopian, which is Neuralink.
1:10:16
Speaker2: And it is the connection between this new form of intelligence that we have
1:10:21
Speaker2: and our analog biological meatspace selves.
1:10:25
Speaker2: And I'm curious of your takes on just brain-machine interfaces in general,
1:10:28
Speaker2: the practicality or even realistic version of what those look like.
1:10:34
Speaker2: Will we actually get there? Are they critical? Who are they critical to?
1:10:38
Speaker2: And kind of what is the order of operations that we get to this point that Elon
1:10:41
Speaker2: is guiding to, which is basically a merging with artificial intelligence is
1:10:45
Speaker2: we are pretty slow. We are building this thing that's very fast.
1:10:47
Speaker2: Is there a peaceful coexistence of those two things through companies like Neuralink?
1:10:53
Speaker0: You're asking great questions. So Sequoia were not investors in Neuralink,
1:10:58
Speaker0: but I have personally known a lot of the, like, I mean, I followed the company
1:11:04
Speaker0: from the very beginning.
1:11:05
Speaker0: I have some very close friends that were there, basically, from the beginning.
1:11:09
Speaker0: Like, I knew most of the founding team as social friends, and some of whom were
1:11:14
Speaker0: actually Burning Man campmates back when I used to go to Burning Man a long time ago.
1:11:19
Speaker0: I haven't been in a long time. But believe it or not, like, a lot of the early
1:11:25
Speaker0: team was from my burning aid camp. We were called the phage.
1:11:27
Speaker0: It was a bunch of scientists, especially a lot of biologists.
1:11:32
Speaker0: It was like Matthew McDougall, who's the chief scientist of the company,
1:11:36
Speaker0: was in my camp for a long time.
1:11:38
Speaker0: I think that Elon's philosophy with starting companies has...
1:11:45
Speaker0: People will not believe me when I say this, but I really think he's basically
1:11:49
Speaker0: said, what are the biggest existential risks for humanity?
1:11:52
Speaker0: And then once he finds one, he goes and starts a company to actually try to address that risk.
1:12:01
Speaker0: And in the case of SpaceX, it's like, God forbid if some astronomical event
1:12:08
Speaker0: happens that destroys life on Earth, such as an asteroid hitting our planet.
1:12:15
Speaker0: We need to have humans somewhere else and have a fully self-sustaining city in that place.
1:12:22
Speaker0: So another one is Tesla, where he saw the challenges of climate change early on.
1:12:31
Speaker0: And, you know, I think there are many other challenges too, but he saw it as a challenge.
1:12:38
Speaker0: And I think he also had very strong intuition for what was happening with both
1:12:43
Speaker0: solar production and with the ability of batteries to create a Tesla.
1:12:50
Speaker0: With OpenAI, which he basically started,
1:12:56
Speaker0: I think he was, I mean, in his words, he was afraid of other people building
1:13:05
Speaker0: AGI or superintelligence that he,
1:13:08
Speaker0: you know, that were not maximum truth-seeking and were not taking safety as seriously as he wanted.
1:13:16
Speaker0: And so he decided to create an AI company or AI nonprofit to go try to be the
1:13:23
Speaker0: first to superintelligence.
1:13:25
Speaker0: I'm sorry, I said AGI and superintelligence.
1:13:27
Speaker0: They're different, slightly different things, but let's say superintelligence, not AGI.
1:13:33
Speaker0: But to try to be the first to superintelligence and to do it in a maximally
1:13:36
Speaker0: truth-seeking and safe way.
1:13:38
Speaker0: In the case of Neuralink, I think once he started working on AI.
1:13:44
Speaker0: Felt like the timescale for superintelligence is closer than most humans would
1:13:50
Speaker0: have thought a decade ago.
1:13:52
Speaker0: And that we should have a company that,
1:13:58
Speaker0: or sorry, on superintelligence, even in the case of trying to do it in the safest way,
1:14:05
Speaker0: I think the numbers he's said in podcasts and things like that are even doing,
1:14:12
Speaker0: even in the best case, he thinks it's something like 80% chance that superintelligence
1:14:19
Speaker0: is safe for humans and that everything happens very smoothly.
1:14:23
Speaker0: And, you know, hopefully, he's wrong. Hopefully, it's 100%, 99.999% or whatever.
1:14:30
Speaker0: But I think he said about 80%.
1:14:33
Speaker0: And I think Neuralink is a hedge against that.
1:14:38
Speaker0: And granted, when you started the company, I don't think anyone knew the exact
1:14:42
Speaker0: timescale of when we'd get to superintelligence.
1:14:45
Speaker0: Now, I think most people in the field would say we're probably going to get
1:14:52
Speaker0: superintelligence meaningfully,
1:14:53
Speaker0: before we have high bandwidth, like, readily manufacturable brain-machine interfaces.
1:15:01
Speaker0: But if you go back, like, seven years, and if you would have thought it was,
1:15:05
Speaker0: like, 20 years for superintelligence, we maybe could have intersected that,
1:15:11
Speaker0: you know, kind of, or even got there ahead of time with brain-machine interfaces.
1:15:16
Speaker0: And there's actually a hedge against.
1:15:19
Speaker3: Call it, slightly dangerous superintelligence,
1:15:23
Speaker0: Because, and it's like a radical idea that a lot of people would disagree with,
1:15:27
Speaker0: but you can basically merge with that superintelligence.
1:15:33
Speaker0: And I'm not, I don't want to put words in his mouth. Like, that may not be how
1:15:37
Speaker0: he was thinking about it. But I think that no matter the level you want to take
1:15:42
Speaker0: this, I think that's the furthest extreme.
1:15:44
Speaker0: And I don't mean merging the Borg. I mean having a symbiotic relationship where
1:15:53
Speaker0: humans get leverage from this and are
1:15:57
Speaker0: better able to understand the superintelligence and what its goals are.
1:16:03
Speaker0: Anyways, I think that basically everything, you know, the last one of The Boring
1:16:07
Speaker0: Company, maybe that was not an existential... I think that's probably the only
1:16:12
Speaker0: one that's not an existential threat to humanity.
1:16:15
Speaker0: But traffic is really damn annoying.
1:16:20
Speaker0: And we can have much, much higher quality of life by solving traffic.
1:16:26
Speaker0: And then with X, I think that was existential to humanity.
1:16:32
Speaker0: Maybe not like X... maybe not to human life, but definitely to democracy.
1:16:38
Speaker0: And I'm strongly in the camp that freedom of speech was on the way.
1:16:43
Speaker0: Like, the way out the window.
1:16:46
Speaker0: And so I personally think it was necessary that something had to be done with
1:16:50
Speaker0: X, which I think has had a cascading effect across other platforms.
1:16:54
Speaker0: In other words, I think this is the lens through which he started these companies. And...
1:17:01
Speaker0: It's definitely a narrative violation from like the perception that Elon is
1:17:07
Speaker0: just like a greedy guy or whatever he is.
1:17:12
Speaker0: I really believe he started these companies because he felt like these are the
1:17:17
Speaker0: most pressing problems facing humanity.
1:17:19
Speaker0: And he wanted to play a role in trying to make these things happen well.
1:17:24
Speaker1: I think identifying all of these industries that are critical for humanity is
1:17:29
Speaker1: definitely the main motivations for the topics that we discussed in this podcast today.
1:17:33
Speaker1: We talked about a lot of just frontier technologies. We covered a lot of ground,
1:17:36
Speaker1: but I want to ask this one final question as we wrap things up here.
1:17:40
Speaker1: A lot of the things that we discussed today are like kind of scary.
1:17:43
Speaker1: Super intelligence, like space travel.
1:17:46
Speaker1: There's just like a lot of kind of scary things out there, like human robots
1:17:51
Speaker1: walking around the world. I think if you go and pick a movie out there that
1:17:54
Speaker1: has human robots walking around the world, it's actually not a good movie. It's not a happy movie.
1:17:58
Speaker1: And so I think maybe rightfully so, the average listener or average person out
1:18:04
Speaker1: there is looking at all of the topics that we are talking about today.
1:18:07
Speaker1: And they have experienced fear as an initial gut reaction.
1:18:12
Speaker1: Maybe you could assuade that reaction or maybe give your take about why you're excited,
1:18:18
Speaker1: why this is good, why people ought to be excited, and why people ought to invest
1:18:21
Speaker1: their time understanding each of these verticals that we've talked about today,
1:18:25
Speaker1: because whether we are scared or not, they are coming.
1:18:29
Speaker0: Well, space exploration, we should not be afraid.
1:18:33
Speaker0: I mean, sure, like, the astronauts that go on early missions,
1:18:38
Speaker0: like, they're very brave, and things may happen to those individuals.
1:18:41
Speaker0: But for humanity, I think that this is only a giant net positive.
1:18:46
Speaker0: And is a prerequisite to really understanding the universe we live in and,
1:18:53
Speaker0: like, making progress towards answering questions like, where do we come from?
1:18:57
Speaker0: On The Boring Company, traffic sucks.
1:19:02
Speaker0: Like, no existential risk. On The Humanoid Robots, I personally only see that as...
1:19:14
Speaker3: Upside.
1:19:15
Speaker0: On superintelligence, I do think that superintelligence is one that I'm scared of, to be honest.
1:19:22
Speaker0: And I'm scared of it not being developed properly.
1:19:29
Speaker0: So I'll give you an example. I'm Jewish.
1:19:36
Speaker0: Jews are a very small minority in the world, less than 20 million Jews.
1:19:40
Speaker0: We're always outnumbered, so things like Wikipedia, Jews are always going to be outnumbered.
1:19:45
Speaker0: And when AI models use Wikipedia as a source of training data,
1:19:54
Speaker0: that honestly scares me.
1:19:56
Speaker0: There's a lot of things on Wikipedia that are just not accurate.
1:20:02
Speaker0: And especially when you have some controversial topic and one group is a minority,
1:20:06
Speaker0: it's very easy to have bias.
1:20:09
Speaker0: On those topics, and, like, here, I don't think there's anything that can stop AI progress.
1:20:19
Speaker0: I think it's a human incentive.
1:20:22
Speaker0: Like, if America were to, if the West were to stop all AI research,
1:20:26
Speaker0: China's not going to stop.
1:20:27
Speaker0: You know, like, the West stopped a lot of cloning research 20 years ago, and China did not stop.
1:20:36
Speaker0: You know, like, COVID was almost certainly a lab leak. You know, like, China was...
1:20:44
Speaker0: Like pushing the limits on virus evolution.
1:20:50
Speaker0: And I would much... Like, we're going to have superintelligence someday.
1:20:56
Speaker0: And I think it will be much safer if there's multiple kind of competing superintelligence,
1:21:04
Speaker0: some of which hopefully are good and developed, hopefully all.
1:21:10
Speaker0: But like having some developed by people that I like... And when you have the
1:21:17
Speaker0: diversity of these things, I think it helps a lot and they can keep each other in check.
1:21:22
Speaker0: And, you know, if they're all developing and they have roughly the same level
1:21:26
Speaker0: of performance, it may, like humans, we make it to the point that humans can't
1:21:29
Speaker0: even find the mistakes in the superintelligence.
1:21:32
Speaker0: One superintelligence will be able to find the mistake in another.
1:21:35
Speaker0: And, you know, that's actually very useful for keeping these things honest and
1:21:40
Speaker0: pushing the development in a safe way.
1:21:43
Speaker0: And I just, like, cat is out of the bag. It's going to happen.
1:21:46
Speaker0: And, but I think, like, the...
1:21:49
Speaker0: We do not want only one person to have this capability.
1:21:53
Speaker0: I think I'm, you asked me to be positive, and I'm not being positive,
1:21:56
Speaker0: but, like, nuclear weapon, like,
1:21:58
Speaker0: when nuclear weapon research, when you get to the point in science where the
1:22:03
Speaker0: jump from the science to a working product is not that far,
1:22:10
Speaker0: call it a trillion dollars in today's dollars.
1:22:14
Speaker0: If it's valuable enough, a government will make that jump.
1:22:20
Speaker0: And nuclear weapons are an example.
1:22:24
Speaker0: And I'm really glad that America got nuclear weapons before the Nazis.
1:22:33
Speaker0: And then when multiple people have them, I think it's very important that we
1:22:39
Speaker0: kept it to only nation states having these tools, because nation states have
1:22:44
Speaker0: a strong incentive to not use them.
1:22:48
Speaker0: Terrorists don't have that strong of an incentive. If you don't control a large
1:22:52
Speaker0: landmass, if you don't have a city that you're responsible for,
1:22:55
Speaker0: et cetera, the mutually assured destruction theory breaks down.
1:23:00
Speaker0: With superintelligence is going to happen,
1:23:04
Speaker0: I think there's like a 95% chance, in my opinion, that it is very positive and everything goes great.
1:23:11
Speaker0: But the 5% chance that it's not that great, I think it'd be very bad to only
1:23:16
Speaker0: have one group have superintelligence.
1:23:21
Speaker0: And it is for sure going to happen. And so the biggest mistake we can make is
1:23:25
Speaker0: like, try to stop progress in the West.
1:23:29
Speaker0: And I don't know, with all of that.
1:23:32
Speaker3: Humans are resilient.
1:23:33
Speaker0: We're going to figure it all out. It's all going to be fine,
1:23:36
Speaker0: and, you know, I don't know what else to say.
1:23:39
Speaker2: That's a nice place to leave it. I think that you've just laid out a very nice,
1:23:43
Speaker2: pragmatic, optimistic outlook of the future.
1:23:45
Speaker2: I think there are certainly some controversial things, and if people find them
1:23:48
Speaker2: disagreeable, I think that's good.
1:23:50
Speaker2: I would encourage everyone to go check out your X feed, because I think at one
1:23:53
Speaker2: point you had a chain of like 12 posts in a row that I favorited.
1:23:56
Speaker2: I think one of the things I admire is your ability to just break down things
1:23:59
Speaker2: into very clear first principles, but doing so for ideas that are often controversial
1:24:04
Speaker2: and not broadly accepted.
1:24:07
Speaker3: So I think
1:24:07
Speaker2: It's great to open up the conversation. I would encourage everyone to go check
1:24:09
Speaker2: it out. And just thank you very much for joining us today. It's been a pleasure.
1:24:13
Speaker0: Last thing I'll say, I mean, thank you. Other than thank you for having me.
1:24:17
Speaker0: Look, I'm going to be wrong all the time. I'm going to say things people disagree with.
1:24:22
Speaker0: I like, I think it's, I think we lost the ability in the West to have reasonable
1:24:28
Speaker0: disagreements in the last decade or so.
1:24:32
Speaker0: And I'm sure I'm wrong on a lot of stuff. I welcome the conversation and I hope others do the same.
1:24:38
Speaker1: John, thanks for coming on today.
1:24:39
Speaker3: Thank you guys.