Transcript
Hook
0:00 · But more more concretely, there are so many AI labs today that have all the cash they need.
0:04 · They have all the compute they need and they're still not able to ship anything soda.
0:08 · And then you start seeing people leave and so on. And my diagnosis it's is [music] it's the culture. If you stop taking the actions that demonstrate the mission alignment to what you've said to your [music] team and to your the world matters to you, then your culture starts to fray.
0:26 · Before we get into today's episode, I just have a small message for listeners. Thank you. We would not be able to bring you the AI engineering, science, and entertainment content that you so clearly want if you didn't choose to also click in and tune into our content.
0:39 · We've been approached by sponsors on an almost daily basis, but fortunately enough of you actually subscribe to us to keep all this sustainable without ads. And we want to keep it that way.
0:49 · But I just have one favor to ask all of you. The single most powerful, completely free thing you can do is to click that subscribe button. It's the only thing I'll ever ask of you and it means absolutely everything to me and my team that works so hard to bring the week. If you do it, I promise you we'll never stop working to make this show even better. Now let's get into it.
Introduction
1:15 · [music] We're in Periodic Labs with Anshumita, CEO founder of Amp.
1:20 · Welcome.
1:20 · Thanks for having me.
Why AI Compute Is Being Wasted
1:21 · At Google, uh if you utilize So, there's two types of utilization usually, right, that you're measuring in these clusters. One is node allocation and the other is MFU. So, node utilization is usually like what percentage of cards in the data center are just like used. And that if it's not at like 95% There's no excuse.
1:41 · There's no excuse, right? Like I think 95% at Google, which is where my co-founder Seb uh came from, uh he built the Borg X Borg GQM scheduler at Google.
1:50 · And there I think 95% is considered an outage. So, 96% node utilization is should be standard. And and most single-down clusters are not running at that. So, that's one. And then MFU utilization should be I would say the best in class today somewhere between 60 and 70%. I think it's a leadership question, right? Is is there an And and fundamentally it's an alignment question, which is are the people who are funding the cluster and then deploying the cluster actually aligned?
2:19 · And sometimes theoretically they are, but in practice the number of people in the chain, the supply chain between like the capital and all the way to whoever's managing the cluster and then whoever's measuring what the output is are just so many, you know, degrees of separation away that like the you know, have you ever heard that sort of you know, radian metaphor, which is at the beginning of of of an arc, if you have two arcs that are two lines that are just off by a few degrees, that it spreads out, right? At scale. And I think what's happening is a lot of cluster implementations and infrastructure a lot of frontier labs and other teams, that's what's happening is they're they they initialize the plan, which is kind of like not North Star with a team that wants to do good, but then they're required to scale so fast instead of iteratively that the wastage just compounds really fast at scale. And so, I I think we know the answer, which is just do iterative bring-ups, you know, if you spend time with people who've been in the semiconductor industry or the DSA industry for a long time, this is not new. And I don't think AI should be an excuse. Like sure, something What is new? Okay, we have a lot of new capabilities, but that doesn't mean just abandon common sense. Common sense should always be in fashion.
3:30 · [laughter] You know, AI scaling doesn't change the In fact, if anything, AI scaling should be putting a premium on the value of common sense and infrastructure because the the margin of error now is so much lower and the cost of wastage are so much higher. And the cost of wastage by the way is not just economic. I'm obviously I'm I'm an investor or I'm an investor by background. Over the last few years now we're running an AI infrastructure business called, you know, Amp and I think that it's okay to say this time is different on the capabilities front. Like we are genuinely getting capabilities at of of the of a kind we haven't had before.
4:04 · That doesn't give you an excuse to say this time is different for everything, especially infrastructure. So, look, I love the hacker mindset and the hustler mindset and that's great for the startup mindset. But you you remember this moment where Zuck went from saying uh move fast, break things to like move faster, stable infrastructure.
4:23 · Move faster, stable infrastructure. I think now we need to move fast with like responsible infrastructure.
Responsible Infrastructure and Data Center Backlash
4:29 · Yeah.
4:29 · They're going to say like where is the impact, you know, there was a really in in our class yesterday Scott Nolan, who's the founder of General Matter, came by at Stanford to speak about energy bottlenecks. And he had a phenomenal idea. He said, if you look at the marginal unit economics of compute per hour, let's let's call it like $4 an hour.
4:49 · If you're having to bring up a new data center in a new community, why not just say we're going to charge 450 an hour and that marginal impact or that marginal increase, we just literally take that and give it to the local community as cash. I can tell you as a customer of that compute, I would love that. I'd be happy to pay an additional 50 cents per hour.
5:08 · at scale.
5:09 · Wow, yeah.
5:10 · Because if that means the public benefit is so clear to the communities that the data centers are coming up in, I'm going to feel like that compute is much more reliable. You know, up to 20% of all data centers this year in in the US, my understanding is are at risk.
5:24 · Of community backlash?
5:27 · Of not getting the community support they need to get brought up.
5:31 · Wow.
5:32 · That's a huge number.
5:32 · Yeah.
5:32 · Now, we I think we should dig into what that number is. I think it's a little bit of overstated. Um these things can get over reported. But it They don't just care about jobs. They care about all the other stuff around it, right? Like they they care about power grid, they care about environment.
5:45 · permitting and so on. And and imagine I think if you said but there's a new AI deal if we're bringing up a data center in your community, we're actually going to reduce the cost off your electricity bill. Okay, now we're talking.
5:58 · Yeah.
5:58 · Right? The The community is going, "Okay, now now this is a deal. I feel like a partner in this."
6:03 · Yeah.
6:03 · Right now, that's not happening.
6:05 · Mhm.
6:06 · There will be audits.
6:07 · There will be investigations. And when the when the regulators come, I don't know when it's going to be. The folks who are moving fast and breaking things in the name of AI progress better be prepared.
6:19 · That's certainly not how we're procuring compute. Oh, we're trying as much as we can to work with partners who have long-term track records. Many of whom, by the way, are not like AI providers. I think this whole idea of neo clouds being somehow this new category is a lot of marketing speak. There are really good, reliable, trusted data center providers in America who've been around 20-plus years.
6:38 · Yeah.
6:39 · I love those folks. They know how to Sure, have have they Are they sponsoring happy hours at NeurIPS? No.
6:46 · [laughter] Are they legibly bitter lesson built?
6:48 · No. Are they hanging out in my you know, in like situationally aware parties? No. But they're adults.
6:54 · Yeah.
6:55 · I trust them.
6:56 · They can run land. They can run power.
6:57 · power shell. They have credit histories.
6:59 · We sit down, we have conversations. Many of them live in Silicon Valley. They've, you know, they've had to deal with the boom and bust cycles of the internet. And I love those folks, you know? They They are stable infrastructure partners and thinkers. And I think there's a lot of short-term thinking going on in the compute layer. And it's going to catch up to us. It It's not going to be good.
AMP Grid: Making FLOPs Flow Like Megawatts
7:19 · You talk about aligning incentives. Um and you know, I would think that aligning incentives means you have the full stack in one company, which is XAI and OpenAI, right? So, you as a standalone infrastructure layer, why are you somehow more aligned to your portfolio companies than people who just own the whole thing?
7:40 · In systems design, right? there's there's two regimes of architecture, right? You have integration and then you have pooling and utilization, right? So, or rather the way to increase utilization often is you can do systems integration where you collapse a lot of process into one node, or you can pull out a process from a node and share that amongst various that resource amongst several different nodes. And so, we see the the Amp Grid, which is the the system we're building here, which is basically a compute grid. You know, we're doing trying trying to do for compute what the electric grid did Yeah, what the power grid did for electricity.
8:17 · This is a pooling and utilization layer across clouds. And so, we're actually the opposite of a full stack integration. It's like a It's much more horizontal. And it's, you know, it's multi-cloud, it's multi-silicon. Um the goal is to try to make mega flops flow like megawatts. And that is very hard to do today for many reasons.
8:39 · Like there's stranded pools of compute all over the place and and there's no fungibility. And so, right now we do it at the level of scheduling, and we often do it at the economic layer. But as we start to announce what we're working on, it's extraordinary like how many folks are coming out of the woodwork saying, "Hey, I'm actually working on a way to make compute fungible at this part of the stack and that part of the stack."
8:59 · And as a grid, we'd like all of these folks to participate on the grid. This you know, people will often ask me, "Andre, you're a neo-cloud?" And I go, "No, actually, neo-clouds are suppliers.
9:09 · Oh, sometimes you laugh at how you venture capitalist from I go, "No, actually, they are they are demand like sort of off-takers of the grid." We see ourselves as what's called an independent system operator. So, if you study the history of the electric grid, once it became legible to a lot of factories and industrial sort of participants that, "Hey, actually turns out pooling is a good idea. We should pull our generators instead of all having a half a generator running at half capacity in our backyard." There was a need for an independent entity who could coordinate all these parties.
9:36 · Transmission line, power generation facilities, transmission lines, factories. And that neutral coordination mechanism is very critical. In order if if you study like the history of grids, the most enduring ones were those that never owned their own assets.
9:53 · They were ones that had or often started with long-term anchors who were uncorrelated sources of demand, a steel factory or shoe mill or whatever in a particular town who weren't competitive where the steel factory wanted to spike up at night, the shoe mill wanted to spike up during the day.
10:07 · So then you pull and you share, right?
10:09 · So each of you is guaranteed some base load, but then you you kind of schedule your spikes to drive a peak utilization across the town. The gold standard, so to speak, historically has been these utility companies like PJM Interconnection in the Northeast of America where they over many, many years became this what's called an ISO, an independent system operator of the grid. So that's how we see ourselves.
10:31 · Yeah.
10:31 · Um economically, that's what we are.
10:33 · From a technical perspective, we started at the scheduling layer because Sab and Mihai who run engineering here built that scheduling They they did that at Google. Um and uh And you have infra shops from Discord as well. I I I don't know I don't know if Discord is like the primary identity of what whatever. I'm just No, Discord was Choosing a well-known name.
10:53 · Well, I I I So I was running the developer platform there. The internal infrastructure I was not responsible for. That was actually a guy by the name of Mark Smith who was extraordinary. And yes, Discord did pull So Discord is actually a counter example. I had the chance to learn a lot about fully full stack infra there because thing, yeah.
11:09 · Um it's the it's the other architecture which is um Discord built its own WebRTC so voice and video infra.
11:18 · Mhm.
11:18 · So like Discord did not use Yeah, did not for communication Discord did not use third-party infra. It was all built in-house. And then the way you maximize utilization was you pull demand from the world's 200 million plus monthly active gamers, right? And so so that's that's how those stacks were were constructed. You know, again in system design the two concepts that keep coming up over and over again are abstraction and and composition.
11:42 · Right? And bundling and unbundling abstraction and composition, [laughter] like verticalization and and horizontalization. So in that sense Amp is an independent system operator of the grid. We pull demand we pull supply from a number of partners we trust at about 1.3 gigawatt scale over 4 years.
12:01 · And then we pull demand from some of the world's best, you know, research labs and so on. We're sitting at one, you know, periodic labs who need extraordinary long-term demand. And the idea is that you know, each of them is guaranteed base load on the grid, but they can spike up and down flexibly on for compute with much shorter timelines as needed.
12:19 · That that was roughly the design of the program I came up with at A16Z called oxygen. The same that was the same design of the the GQM Borg X Borg GQM implementation at Google that Mihai and Sevin built. Which was that how do you allow uh teams inside of Google on the internal infrastructure to be guaranteed capacity uh for their base workloads, but when they need to spike up on research, how could they ensure that that was sufficiently there?
12:48 · And of course the big innovation that was discover not discover but it kind of implemented in the space this infra space maybe 3, 4 years ago at Google was the idea of interruptible demand. Right? Where you just queue up a bunch of jobs and through this like sort of credit system there can be a bidding mechanism.
13:05 · It's it's dynamic prioritization basically and and jobs can get interrupted based on somebody else who's saying, you know what? I have 10 tokens, 10 credits I want to spend on this job. Another like team lead, research lead is like genie 3 or whatever is only worth five, you know, credits and nano banana 2 is worth 10 credits. And so the nano banana job gets priority. I That's a That's that's up example.
13:26 · It's very real. Brain Marketplace was real.
13:28 · Yeah.
13:29 · And uh we've we've covered this on the pod with David who was Oh great. Okay. Awesome.
13:33 · Uh and the the criticism is that well, actually sometimes you need central commands to go all in on the thing. Uh and actually sometimes capitalism via credits doesn't work.
13:45 · Not not it's not a criticism of Amp. I'm just saying like this is a thing that has been tried uh internally within Google and it led to Google missing GPT.
13:52 · Like we structured ourselves essentially very similarly to Google. We we are structured as a holdings company. So, you know, Alphabet Holdings is Alphabet Holdings and then they've got these subsidiaries called Google and other bets other bets and so on. We've got um you know, Amp Holdings and we've got our infrastructure business and then we've got a capital business called Foundry that incubates new frontier AI labs and invest in them as venture capital um like periodic, you know, we put a few hundred million dollars into Anthropic from our fund earlier this year. So, wherever we feel like teams are making progress, especially researchers and so on who push the frontier inside of existing labs like DeepMind, I find, you know, there comes a point where they feel misaligned with the dictatorship of Alphabet Holdings and at that point sometimes the dictatorship doesn't want them anymore and they're like, "Thank you. You've done your job here. You've kind of helped us through the zero-to-one phase and for whatever reason we're going to deprioritize your amazing like omnimodel or whatever it is and instead we're going to prioritize coding and uh I think that's a tragedy, but I get it.
Foundry, Frontier Labs, and Research Hoarding
14:48 · Like they're, you know, Sergey and Demis are running their own business there, but that doesn't mean we should the rest of us should sit around waiting for that progress to get unlocked for the rest of the world and humanity. I mean, if if you think about how much extraordinary research has happened inside of DeepMind over the last 10 years. I I I mean, Demis and Sergey and those guys did such a great job, but at the end of the day, so much of that has never seen the light of day.
15:12 · Or they they're like papers only, but they never actually shipped into production or I mean, what's worse is the paper is actually not even being published anymore cuz there's a six-month embargo inside of DeepMind, right? Like we've heard about this where a paper comes out and then I think he has a 6-month embargo window where if anybody on the business team says this could be interesting, it's embargoed for life.
15:30 · Exactly.
15:30 · So the stuff that gets published is the stuff that's not good enough. So there's a adverse [laughter] selection problem, basically. Yeah. At this point It's It's a common complaint at NeurIPS, by the way. Like it's like, "Well, why would I look at the papers that are the trash of GDM?"
15:42 · [laughter] Again, I think it's a tragedy. I mean, I get it. They're running their business, but the rest of the space I I I think there's negative externalities of uh research being hoarded. And so that's a There's a market failure. And And somebody needs to unlock that research. And uh we can't do it on our own. We only have 1.3 gigawatts of compute. That's nothing. That's about $40 billion of cloud spend.
Gigawatt-Scale Compute and End-of-Life Prediction
16:02 · Uh we're going to need a little more.
16:03 · said That's a new number I haven't I haven't come across that uh that gigawatt number. That's huge.
16:07 · Yeah.
16:07 · And And to be clear, we haven't secured all of it. That's how much demand we have started to secure. I think publicly we haven't actually confirmed how much we have for this year.
16:16 · Where do you want to get to?
16:17 · I think the steady state would be that we have a base load pool of 1.3 gigawatts at all times of base load capacity. For spike capacity, right now my estimate is we need roughly 6 gigawatts over the next 4 years for all our teams to feel like they were able to keep moving the frontier whatever they're working on, whether it's like superconductor discovery over here.
16:38 · There's a new investment we're working on right now, which is in the end-of-life prediction space in in healthcare. It's extraordinary how much you can you can give people You know, this was actually my graduate school work. I went to graduate school for bioinformatics at Stanford Med.
16:50 · Yes.
16:51 · And I know we Econ MCS bio.
16:53 · Uh so my I was this really weird cat where like I was never satisfied with my major options. So at one point I was an econ major, then I was a CS major, then I was a MCS major called mathematical computational science. And they decided they were going to end that major.
17:07 · So I took all that coursework and I applied it to grad school my graduate degree in bioinformatics, uh which was the master's program. And then I thought I was going to do a PhD.
17:15 · I never ended up doing it. I got I I dropped out and went work at Cliner, but I was lucky enough to apprentice with this professor at Stanford Med. His name is Nigam Shah and he was working on end-of-life prediction. Stanford is one of the only research facilities in America that has a longitudinal patient data set that's larger at scale. I think it's at least 12 million patient lives. The only larger data set is at the VA, the Veterans Affairs, you know, of America.
17:45 · And to do research, like do any deep learning and so on on that data set, it was called the stride data set at that time, you had to be a Stanford Med School affiliate, which is why I went and enrolled in the bioinformatics department.
17:55 · Wow.
17:55 · Um and look, deep learning was early.
17:57 · Nigam Shah had the visibility like the the vision to see that like you could do end-of-life prediction to help palliative care. In you know, in America, the like over 30% of all Medicare Medicaid spend, at least at that time, was spent on end-of-life care.
18:12 · And what's, you know, we we grew up in Asia, so we all, you know, at at least I I I won't speak for you, but I have a very different relationship with death than I find folks who grew up in America do.
18:24 · In America, spiritually and culturally, especially in Western societies where Christianity the Christian tradition sort of frames death as this terminal point. Um there's often a judgment day and so on. The the way we view death is with a finality. In Indian culture, in Hindu culture, you know, death is one Buddhist as well.
18:48 · You're a Buddhist, yeah. So it it's one it's one step in a journey of many lives, right? And so, uh I grew up in the city called Chennai in the south of India, and when people die, you dance on the street. You know, there there's like a procession where your your body is carried to to be cremated and your family like like celebrates. And there's drums and so on. It's to say it's thing.
19:11 · And, uh it's because the idea is that you you're going to be reincarnated. You know, you've been liberated from the responsibilities of this life, and now you're on to your next. It's a new It's like going off to a new college or whatever, right?
19:23 · And so, it was so alien to me when I got here as an undergrad that it the medical system works backwards from that assumption that we have to view death as this terminal thing and delay it, postpone it. It's a bad thing.
19:40 · And so, at the time, clinical decision decision support in in the United States was this very primitive field. Even to this day, physicians in the United States often will tell you when you have a terminal disease, this is your we we've diagnosed you, which is great. Our Our ability to diagnose you is extraordinary. You have somewhere between 6 months to 6 years to live.
20:03 · What do you do with that information?
20:04 · Yeah.
20:05 · It The error bars are so high that that then you you you when in times of uncertainty, we default to culture.
20:12 · And when the culture is let's This is a bad thing. I've got to prolong my life, then you start doing things like And and and just to to sort of from a systems perspective, what's going on there is physicians often feel like they need to provide such higher bars because there's always some uncertainty in end-of-life diagnosis. And if you provide the wrong diagnosis or recommendation to your patient, you can be sued for medical malpractice.
20:41 · And then your license can be taken away. It can be catastrophic for your career. In contrast, if in countries where that's not the case, what you often observe is that patients like physicians are quite prescriptive with their recommendation. They say, "Hey, this is your condition. The literature says that you probably have this much time on Earth left. My expert opinion is that you are an outlier or whatever. And they try to be more prescriptive.
21:11 · And that empowers a patient. Right?
21:13 · Because then a patient can say, "I trust my doctor. They said on average I have 6 months to live. But if I do these things, I may have a shot because of my particular predispositions or or my genetic history or whatever." And that empowers you to go about your life in a actually more scientific way than leaning on religion, culture, spirituality, and so on.
21:35 · In contrast, here because of that medical malpractice sort of thing looming over your your your head, a physician never gives you a clear recommendation.
21:45 · Right.
21:45 · So instead you you say, "Okay, doc.
21:48 · Well, let's try it all." And then you start a whole regime of [panting] drugs and therapies. And then you often spend weeks and weeks in the hospital. And that deteriorates your quality of life. And when that deteriorates your quality of life, like you instead instead of spending your last few days doing the things you love with your family, you're spending it on a hospital bed.
22:10 · And that ends up being 30% of Medicare Medicare and Medicaid. So it it's worse for the patients. The doctors feel terrible. The American taxpayers paying a huge amount of money. And so this is why Nigam Shah, who is this professor at Stanford, said, "Honestly, if there's I I kind of sat down with him. I was this young guy, you know, I was 21. And I was like, "I I want to work on a big problem." He's like, "The big problem is end-of-life care."
22:32 · And so we tried to do deep learning to say So we start trying to learn deep learning on these stride patient data sets to say, "Could you have an AI system make a recommendation that is orders of magnitude more precise about how much time you have left once you've been diagnosed with a terminal condition than a human?"
22:49 · And then if we can get that precision to be high enough, then you can empower the patient. And it turns out the tech works. They get Once you get the data set, like RL works. Honestly, even regression models work. You don't need to get that fancy. Half the time we were just running doing like very simple neural nets.
23:06 · Today, I what we can do with RL is extraordinary. The problem remains then and now is regulatory. Because you actually can't shift the burden of the wrong clinical diagnosis from the physician to the AI system. And so at that time I got quite disillusioned 10 years ago. For 12 years ago where cuz I felt I just didn't have the resources to influence regulation. Today I'm very lucky. I'm in a different place. You know, I've I'm a lot older and so I've been spending a lot of time on my next incubation, which is how can we unlock the like patient empowerment by training AI models to do end-of-life prediction much with much more precision.
23:48 · Oh, wow. And you're you're still focused on this I I haven't been able to get it this out of my mind a single day for the last 14 years. This is the hill I want I would like to die on. This too, I would say. You know what? I actually prefer not to die.
24:03 · Yeah, exactly.
24:04 · But [laughter] but I I think two bipartisan issues I think two issues that should be bi- bipartisan in America are how do we empower patients to make the right clinical decisions at the end of their life such that we're reducing the taxpayer burden with science. It's just good old science and AI can help here. And the second is, you know, net positive data centers. Cuz I think that's the biggest critical bottleneck on training and good enough AI models to help people at the end of their life. So So they sort of two sides of the of the same scaling bottleneck curve, but those two we formed Amp as a public benefit corporation.
24:38 · My wife and I, who you've met, you know, you've met with Yeah.
24:40 · Um her passion is is education. You know, her family is a long line of educators and so on and of physicists. And so this this class is my attempt to stopping the black sheep of the family and be a an educator. But if I'm not educating the thing I would be doing is working, you know, on on these two problems whether on the political spectrum or as a researcher back at at in some lab.
25:03 · Um and my hope is if anyone's listening to this podcast if you know if if they're passionate about either of those two topics, I'd love to hear from them. We we we can share the contact in the show notes, but uh we're looking for people to join both of those missions on the on the political side as well as on the medical side on the research side.
Frontier Systems, Output Maxing, and Alignment
25:20 · You know, you said uh this is a discipline that you want to form. You call it it's called various variously called frontier system. It's variously called one person frontier lab.
25:31 · Uh what is the ideal name or shape of this? Like the what is the mission?
25:36 · Of the class?
25:37 · Uh of the discipline that you're I guess exploring, right? Like I I the class is called frontier systems, but like for me maybe one phrase is like uh you're just anti-waste, right? Which is wasting wasting GPUs, wasting in human and Medicare. But like is there is there a broader theme that I'm that maybe you can encapsulate more succinctly?
25:57 · Yeah, yeah. The from an engineering perspective it's basically it's output maxing.
26:00 · [laughter] You know, it's the it's the it's the department of output maxing.
26:04 · of what we have.
26:04 · Exactly.
26:04 · I'm a huge believer in optimal outcomes. You know, I I think both in America and other countries uh we are losing our appreciation for nuance and this is the thing of AI is the same case, right? Oh, the bitter lesson holds. Okay, fine.
26:19 · But that doesn't mean you just like throw 500 GB300 500,000 GB300s at your like you know, sub-optimal model scaling and you waste a bunch of compute. It also doesn't mean that, you know, the most optimal have like 50 different architectures where there isn't enough standardization. Like one of the reasons Anthropic has had extraordinary sort of velocity is cuz they picked the transformer architecture and said this is simple, let's double down on it, right? And now luckily there's enough investment going in the space that we can afford other architectures, but at the time investment was just too fragmented into other architectures. So, that arguably unlocked scaling.
26:54 · So, I think there's a philosophy I think we all owe it to ourselves to do output maxing with a new capability called AI on a global level. I think if I was starting a new department at Stanford, depending on how fuzzy or technical I wanted to be, I'd probably call it the Department of Alignment. Uh [laughter] you know, like It's an overloaded term.
27:12 · But it is but alignment really is a hard problem.
27:17 · Mhm.
27:17 · And I think when you unlock it, full stack alignment is super hard in any organization and in any system.
27:23 · Like in a in a venture capital firm, if you can have full stack alignment between your limited partners and you know, the founders who are creating the value and ultimately the public that owns the IPO stock, that is a gift that keeps giving. And when you study the history of these systems, when they start off, they usually start out small scale where the feedback loop is actually so tight that there's alignment. And then the more you try to scale, the more division of labor happens, more specialization happens, and at each step you add abstractions.
27:51 · And wherever there's an API interface, there's like loss. There's communication loss. And so I think a really cool thing would be for us to figure out is there a way for us to have our cake and eat it, too, as an engineering discipline, is there a way to actually scale up and scale out without losing any alignment, without you know, lossy transmission?
28:13 · I mean, standards.
28:14 · So, standards is one way. The other way is you just have net new capability. So, like sup you know, what we're trying to do here is discover new superconductors. A room temperature superconductor would be a lossless transmission mechanism for energy. I mean, we would have flying cars.
28:28 · Yeah.
28:28 · Yeah, right within a few years of having a new room temperature superconductor. So, I think those are the two. You either have to standardize on protocols or or API specs that allow lossless communication, or you can come with a whole new capability that unlocks so much abundance that standardization doesn't matter, cuz you just unlock net new capacity.
28:46 · Yeah.
28:46 · This So, that This is what I spend my days thinking about [laughter] about these days.
Compute Markets, SF Compute, and Non-NVIDIA Chips
28:50 · I mean, no. I think every infra person at who wants scale and wants to output max does eventually end up thinking about this. Uh we don't have time to go into it, but we have done an episode with SF Compute that is trying to standardize the futures contract for compute.
29:05 · Right.
29:06 · Uh I don't I don't know how that's going, by the way, but like at some point this this will be part of think Evan is awesome and SF Compute is the kind of effort that I hope we can accelerate because what what often happens is these exchanges are really hard to get uh it's hard to bootstrap them.
29:23 · Right? Because they often require there there's many inefficiencies between parties. There's trust boundary inefficiencies in infrastructure because you don't trust, you know, one part of the stack doesn't trust another part of the stack to give them visibility. There's capital markets inefficiencies.
29:36 · There's operational efficiencies. So, if you can inject like a a single shock to the system of a ton of compute demand or supply, then you can accelerate uh these new flywheels. And so, my hope is one day, you know, or or soon, if if SF Compute needs extra like has excess capacity, they just hook it up to the grid and they get flooded with demand from us. And on the other side if they have a ton of demand, but they don't have supply, they again hook up to the grid and it's a two-way protocol where they can just hook up to our capacity.
30:11 · And I don't think we're too far from that, you know, today.
30:14 · Our working implementation of it is mostly through a group of labs, um universities, and a few like uh sort of trusted parties who are who all feel like they're in alignment to borrow an over sort of used word. Um but our hope is to just have it be an open protocol that anyone can hook up to on Hook up for demand or hook up for supply? And Primarily demand, it sounds like. Like you wouldn't want to offer demand.
30:39 · Both, yeah. Unfortunately, what's happened the last 6 weeks is, you know, we thought we'd have a bunch of excess capacity by the end of this year.
30:48 · It's all gone.
30:49 · It's exploding.
30:49 · Yeah, it's all gone. And so, I have my text messages are full of friends, I mean, we know many of these people.
30:55 · These are founders who've raised billions of dollars in San Francisco going on, "Any chance you have like 15 nodes in the next few weeks?"
31:03 · What is the scope for non-Nvidia, right? The you have Lisa Su coming and Reiner Pope as well. And so, there is a lot of demand for more performance, alternative architectures and all that. At the same time, this hurts your standardization.
31:22 · I don't think so. So, actually, Reiner's a great example, right? Reiner's a CEO and founder of Matrox. I actually had him by for office hours in the class earlier today. And there was an insight he brought up that I hadn't considered before, which is when they decided to pick the standard for their data center. They picked the Nvidia reference architecture. So, the Matrox chips just plug in to any site that has an Nvidia bring-up plan.
31:53 · And you know, the It's just software then. It's it's not the hardware.
31:57 · Well, from an input and output IO perspective, it's the same footprint as an Nvidia rack.
32:04 · That makes sense.
32:05 · Where they have done innovated a bunch from what I can tell is on systems co-design. Which is where a lot of the gains are to be had. And so, he, you know, he picked He was like, "Andre, you know, [laughter] there's just so much work to do when you're building a new chip company.
32:20 · Can't fight on every just can't fight on every front. Cuz so, my my question to him was, "Well, you're working on this new chip. Their tape-out is next year. You know, what who are you going to partner with to host the chips?" And he said, "Whoever will host them, that's just not that's not my focus." And I I "But how did you like you decided for you know back to earlier systems design question he he decided that like like he didn't want to be a full fully integrated chip provider. The bottleneck they're focused on is the the logic die.
32:49 · And they he feels they can crank out a ton of performance gains through co-design there. But then that means you delegate you know to our question earlier like it you he's like that the the data center provider is a different part of the stack. And so then he's dependent on that part of the ecosystem to host his chips to get the performance gains to the customer. So now you have another abstraction.
33:12 · And you have you might have loss. So I asked him how how do you prevent loss?
33:15 · And back to your point he said, "I just picked the Nvidia standard cuz I I didn't want to like I I I wanted to piggyback off of an existing protocol. And that what's great about Nvidia is that reference architecture is known. It's open. They've published it. So Jensen's actually enabled someone like Rein Reiner to build a chip company like Matrox. And I don't see them as competitive. The compute demand is so high.
33:39 · Like I don't I think Nvidia's not able to meet the demand of production. So we just need more chips.
33:44 · And I think it's very smart what Matrox has done which is say we're just going to we're not going to innovate on the data center design cuz actually thank you Jensen you've done all the hard work. Where we can innovate is somewhere else. And I I think that's that's very healthy. I think that's how we unlock new bottlenecks.
33:59 · And my view is these you know chip teams like Matrox who have arrived at the insight that co-design is the way the primary bottleneck for them is trust boundary.
Trust Boundaries, Co-Design, and Researcher CEOs
34:13 · To do co-design well you need visibility into the next model generation as soon as possible cuz it takes two years to tape out. So if by the time I bring my chip to market your model architecture has changed I'm host. Now when he was inside Google he was sitting next to the Gemini team he He on Palm or whatever.
34:31 · His co-founder was the uh was one was one of the Palm guys, I think.
34:35 · Yes.
34:35 · Yes, exactly. So, when you're inside the trust boundary of Google, then your systems co-design loop is super tight. When you leave as a founder, one of the biggest risks you take is now you're outside the trust boundary. And so, what I love doing is helping chip teams who can help us unlock more capacity for the independent ecosystem access to trust.
34:54 · Because when I if if I've been like involved with a lab from day one, and I was lucky enough to work with Anthropic, and then I'm on the board of Mistral, and at Black Forest Labs get started. I think at this point I'm on six or seven different teams.
35:08 · Only six?
35:10 · I feel like it would My mental number was going to be 13, but yeah, it's No, I you know, I go deep with one at a time.
35:16 · You were founding CEO of Arena?
35:18 · No, that wasn't an That wasn't Administrative CEO.
35:21 · It was an administrative 5-month gig where Waylon and Anastasia were graduating from their PhDs. And they didn't need a product team. So, I helped recruit the head of engineering, product, and design. But Anastasia has always been the CEO of that company. I I I played a pinch-hitting job. I'm an intern. I was CEO intern for 5 [laughter] months.
35:44 · Um I I interviewed him. He's like he's very very well-spoken. I mean, I think he's a debate former debate champion, uh but also very quantitative and mathematical, which is such a unicorn.
35:54 · you look at his output, he's an output maxer. Like, by the time he was graduating from his PhD, which he only graduated last year, he had published more work with a citation count than like people twice his age. But at the same time, he'd already started a project called LM Arena that was being used by millions of people as a side project.
36:16 · And time and time again, what I've realized is venture capitalists suck at seeing human beings as like dynamic agents. Where where They want to put you in a box. This is your thing.
36:28 · So the first time I got introduced to Anastasios, somebody had told me, like, oh, he's amazing, but you know, he's a researcher.
36:36 · Yeah.
36:36 · I was like, what what what what do you mean he's a researcher? He's not a CEO, not a founder. Not a CEO, exactly. I was like, are you are you crazy? Have you met Dario? Dario is a scientist. He's gone from zero to like what will soon be a trillion-dollar company in 4 years. Being a CEO, nominally speaking, is not that hard.
36:59 · Being a good CEO is hard. Being a great CEO actually requires a level of performance that scientists who have already published at the top of their field have accomplished. It is super hard to be a competitive scientist. To publish in academia over the last 20, 30 years, to make it to the top of your discipline at a place like Berkeley, you were a star athlete. Like you you are an athlete of the mind.
37:27 · And you perform at the highest levels.
37:28 · And to get there, whether you're, you know, Anastasios or Waylon at Berkeley, or you are Robin who with Blackforest and created Stable Diffusion, or if you're like Guillaume at Meta who created Llama before he started Mistral, like the amount of human leadership you have to demonstrate to get the resources, like get the trust of the organization, publish it, put it up I mean, I would just fund researchers all day, [laughter] right? If who who who have contributed already to the field. If if if they've put Soda out there, they're they're star athletes already.
38:02 · If they haven't done Soda, look, they they can still be good CEOs, but then I find the failure mode is that they just don't want to be CEOs. They primarily want to publish. And that's okay, too.
38:13 · You know, one of the things we do with the Ampere Grid is we donate excess compute we have to nonprofits, like university labs. We carved out like a couple thousand H100s, but I do think there's extraordinary research being done on university campuses. You know, my father-in-law is a physicist. He's a professor. Extraordinary work in physics, and we need that. But if you want to be a CEO, what you need to be willing to do is be super confrontational you know, outside of science. Like within the scientific community, some of the best researchers are very confrontational about their convictions, right? This architecture is right.
38:49 · To be a great CEO, you basically have to be willing to be confrontational up and down the stack.
38:53 · To your own team.
38:54 · To your own team, hiring, recruiting, customers. Well, I would say Yeah, pretty much to everyone.
39:00 · Yeah, yeah.
39:01 · I see I I feel a little bit of that in my own work, but like yeah, I can't imagine the stakes that Dario has had to go through. It's It's pretty good.
39:08 · I I don't think the stakes are that different from how you're feeling it, right? Stakes Stakes are personal scaling vectors, right? Like the stakes that seem so low to you, like having this podcast, where you can talk to somebody and just have a conversation. I mean, you're an extraordinary communicator, right? Like already in this conversation, you've pulled more out of me than most people.
39:26 · [laughter] You know, and I've been on 12 podcasts in the last 2 weeks.
AI Coachella and First-Principles Thinking
39:29 · we've just seen each other enough that you There's some base trust. And I think And I know that you you know that I've done my homework, and like I I know that trust is a big deal for you, so Right.
39:37 · Yes.
39:38 · I I think trust is about consistency, and you and I have seen each other in the community for years, right? Like I don't know the first time we met was at NeurIPS in New Orleans. I don't know if you remember that.
39:49 · I remember it.
39:50 · god. Ra- Raiko had set up this Yeah.
39:53 · You know, Raiko's amazing, and he set up this luncheon, and Yeah, I was like, who's this Discord guy? I'm like, okay, but No, you weren't like No, you made some investment. You were much less polite.
40:00 · You were like, who's this VC?
40:02 · [laughter] You like No, I Was I Oh my god, I'm so sorry. I was I didn't know who you were.
40:06 · I'm so sorry. But like you weren't you weren't The introduction was bad. I I was I I didn't know who you were.
40:11 · The See, this is the thing about context, right? Like Um but then I I think I like heard your accent. Yeah. And I was like, are you are you Singaporean? And you're like, yeah, and I said I went to high school JC in Singapore. And then the ice broke.
40:23 · Yeah, yeah, yeah.
40:24 · But this is the you know, there there are some in the scientific community sometimes the stakes are very high for people who haven't had the emotional We know what it's called EQ.
40:37 · Mhm.
40:38 · coaching and mentorship. Right? Which is like to have scientific impact you often need to be a extraordinary emotional like emotionally in tune person with the folks you're trying to influence. And so what comes so naturally to you is actually a super high stakes thing to other people.
40:54 · And so I wouldn't assume that Dario is more stressed out than you. You know, these these things are like you'll be surprised how similar and small sometimes the problems are to you that some of the world's biggest leaders are facing. And that's what I've learned from this class. You know, the the guest speakers are Sam, Satya, Jensen, they are Coachella.
41:13 · Yeah, they they are Coachella. Right? So we got to get all the headliners and they're I'm very lucky that some of these people have either mentored me over the years or I've done business with them and when you you know, take the performative stuff out and any assumptions you may have about about these people that you read in the press or on Twitter and we're all just humans. We're all trying to get along and what's so special about this moment is AI is forcing like scaling the bitter lesson is forcing a lot of people to revise their assumptions for how the world works and go back to first principles or go back and educate themselves. So the people kind of people you know, I was um I won't name who this person is, but I was at an event last week in Texas and ran to somebody who said, "Aonde you know, I I came across the class. What do you think about real-time action prediction models?"
42:04 · And I was you know, don't know how happy it made me feel when they asked me that question. I know they've done the work.
42:09 · They've challenged the assumption they you know, they didn't ask me, "What do you think of world models?
42:14 · They said what do you think of real-time action prediction models?
42:19 · World models, don't get me wrong, are cool and everything, but you and I both know that that is a layer of abstraction that is sometimes not usefully precise enough.
42:26 · Yeah.
42:27 · Right? Arguably There's like four different kinds of world models.
42:29 · Exactly.
42:29 · [snorts] We've done the part with general intuition, by the way, which is very focused on Oh, cool. Yes. I love him. He's He's great. And this what I love about people who've done that level of work, they realize they're not in competition with people who the rest of the world thinks they're in competition with.
42:45 · Yeah. Because they're not in the category, they're in the specific thing they're trying to do.
42:49 · They're focused on their mission and they have a systems understanding of the bottle they're trying to to like, you know, solve. And when somebody else says, "I'm working on a real-time, you know, action prediction models, too."
42:59 · Pim goes, "Oh, I love that person. I want I can learn from them." But the minute they're like, "Oh, that person's a world model person." It's like, "Ah, like which type of world model person?"
43:08 · But mostly they're just trying to figure out if they're wasting their time cuz we don't have enough time. So, you know, Pim, for example, is super loves this other company I work with we've talked about called Black Forest Labs, you know, and he's mentioned me multiple times that he's so he thinks what Flux is doing is really cool. You know, Andy Blatman came by and spoke in the class.
43:28 · And what I find over and over again is for people who do the work, who can be usefully precise enough about like what is actually going on in the world of frontier research, the sense of camaraderie is still real and alive. But it gets lost sometimes when you have to like abstract the technical complexities in like business terms and then the VCs are like, "How are you different from that world model company?" I say, Yeah.
43:50 · Where do I even start to explain this stuff? And then The misalignment Yeah, I think like people listening get a sense of like what it is like to operate at a real level like you're like yourself rather than like like the journalist level where you have to sort of put everyone in like a rough category and create a narrative of competition and who's winning today, who's behind.
Leading vs Winning in Frontier AI
44:09 · Yeah.
44:11 · This idea of winning is so weird to me.
44:15 · You do want to win. You want to compete you you want to compete with us.
44:18 · No, I think you want to lead. No, I think you want to lead. Yes, so you want to push the frontier. You want to push the state of the art. You want to do something that hasn't been done before. You want to capture value. But you don't want to capture so much value that like people think you're on a line with your mission or trying to do what's best for the world.
44:33 · You want to capture enough value that you can keep innovating, right?
44:37 · And I think that people want to lead. They don't really this idea of winning and losing like you know, again, I love Jensen. He's a he's a leader. The mindset that he talked about on Dwarkesh's podcast, right? He's like, "I didn't wake up with a loser mindset."
44:51 · [laughter] I think that was awesome, right? Because he's he's an engineer. Dwarkesh has done the work. So there's at least even though that to me it was very obvious they're talking about the same thing. They just passed each other. Like they said to on like, you know, basically Jensen has this like five-layer cake abstraction of how the industry works.
45:09 · And Dwarkesh had I I think from that podcast had more of like a pre-training, mid-training, post-training systems loop concept.
45:16 · It's just a factor of who he talks to, right? Like it's pretty clear.
45:18 · It's the abstraction, the mental models, the it's the whole dude, so much of the problem in the world is reasoning by analogy.
45:27 · Mhm.
45:28 · And then the assumptions that are held invisibly.
45:31 · Yeah, I I I've said like this is actually the best time in human history for first principle thinkers because everything you think will happen is actually now coming true.
45:39 · Correct. [laughter] And the venture capital community is like notorious for this where people look in times of uncertainty, they like cling to axioms that end up being true from the previous era. And they they kind of like proclaim them with confidence as if they're truths, but they're not. And it's very important to see the distinction between a heuristic and an axiom.
46:03 · Yeah.
46:04 · An axiom can be proven like from internal consistency point of view.
46:10 · A heuristic is a where you kind of a shortcut. And my god, the number of people I have had to put up with over the last few years who proclaim like use heuristics as axioms to judge people to judge which companies are going to succeed or I mean the number of people that are like oh yeah, yeah, yeah, Anthropic, they're just training models right now. But this one continue.
46:34 · They go to be the B SAS.
46:35 · Yeah, they like which which over the fullness of time if you squint at it maybe, but the way you arrive there is so important that you can you just you can dismiss people. Here's what happened, right? What happened is Anthropic basically achieved takeoff in October of last year. That training run Whatever F37?
46:54 · I forget the numbers now, but whatever that checkpoint was We saw it at Cognition.
46:58 · Yeah, right? You probably the to those of us in the community, especially once post training was done and it was released in December.
47:04 · Yeah, can I sneak a sneaky question in there? I don't know if you have a perspective, maybe you don't. I just the the number one question is how did Anthropic crack coding, right? Because Claude one, Claude two, okay, like it was part of it, but it wasn't a big deal. And the leading hypothesis, it's a lucky dice roll that was then compounded, right? Like it was like mildly better, but then they saw it and they were like, okay, let's really invest.
How Anthropic Cracked Coding
47:29 · I had this very annoying teacher.
47:31 · Yeah.
47:31 · I I went to this boarding school called Rishi Valley in India, which is like this uh bird preserve, like 350 acres of of bird preserve in in rural India. And there was no technology for 7 years.
47:43 · Uh there was this teacher, I won't name them, but they would have this I hated it every time he said this to me. He was like, luck favors the prepared mind. Which is like a common saying, but the way he delivered it like always grated me cuz he's always try Like I was always one of those kids who got like a good grade without trying very hard. Cuz like high school and middle school is not that hard if you if you're generally like paying attention so on.
48:08 · And there was this one time where But but but I I would get an 80% grade, and he would keep pushing me to say like the reason you didn't get the 95 plus percent is because you're not that lucky. And I would say, "What do you mean?" Cuz I I would think that I deserved that grade, and I would sometimes argue with him.
48:25 · And he'd say, "You didn't have a prepared mind. If you want to get lucky again, there was basically one time where I got like 95 or 96 on this on this subject.
48:34 · And I Now that I I felt entitled. I was like, "Okay, I'm going to keep doing this." And I didn't. And then he was like, "Luck favors a prepared mind. You got lucky last time, but you got to stay prepared." And I didn't understand what he meant. Now as I'm older, I'm like, "Okay, these adults actually knew a thing or two."
48:47 · Anthropic has been the most prepared company for 4 years. And so then when the right like context data comes in, the right developers start sending in, you know, the right context diffs, sure, you could say you got lucky, but if you ask me, they're pretty damn prepared with paranoia for like 4 years.
49:07 · And if you remember, it was so hard for them to get going early on that they had to do so much more with so much less that you just have to be prepared to be so efficient.
49:17 · Yes.
49:17 · There's numbers on their burn compared to OpenAI. I've written about it, but they are so much more efficient in the in the way It's not It's not even Not even close.
49:27 · Yeah.
49:28 · But it's so clear, right? Like how to output max for the world. They have been prepared. And you could call that luck, but Yeah.
49:35 · Luck favors a prepared mind.
49:36 · This is one of the things that I was going over some of your old lectures, and you were like, you know, data, people think it's a moat, and like actually it's culture. And actually it's team. Actually. And and I it's There's different levels of moats, and this is the ultimate one that determines everything else. Which you you can then compound.
Culture, Hardship, and Anthropic’s P0
49:54 · You're saying culture is the ultimate one.
49:56 · Yes.
49:57 · Yeah.
49:57 · But the thing about culture is it's very fragile. So moats I I I don't think there's very few moats I found that are actually moats. There is is a nice concept. But in reality, you have to replenish your culture. You know, Ben Ben Horowitz was a the speaker in CS183 on Tuesday. And I I asked him this question about the culture bottleneck in teams cuz you know, there there are several AI teams.
50:21 · His book like Hard Thing About Hard Things.
50:23 · Hard Thing About Hard Things. But more more concretely, there are so many AI labs today that have all the cash they need. They have all the compute they need and they're still not able to ship anything Soda. And then you start seeing people leave and so on. And my diagnosis it's is it's the culture.
50:37 · And so I asked him, Ben you know, there were he's been one of the most aggressive investors in AI labs. He goes back to this thing which resonates in my mind a lot. It When I used to work at A16Z, I would book a conference room and right outside the conference room which is closest to the toilet cuz it it was the fastest way for me to go use the bathroom between Zoom meetings.
50:56 · Oh my god, I'll put my sink by toilet optimization.
50:59 · Okay, never mind. It was not healthy [laughter] in hindsight, but maybe this is TMI. But anyway, outside that conference room on the wall was this quote that was printed that said, "Culture is not a set of beliefs. It's a set of actions."
51:12 · And it's by Bushido is a Japanese philosopher. And if you stop taking the actions that demonstrate the mission alignment to what you've said to your team and to your the world matters to you, then your culture starts to fray. So it's not actually a moat, I would say. It's a very very brittle fragile thing that requires daily tending to like a garden.
51:38 · But if you figure out the system to keep that garden tended, which I think ultimately comes down to knowing yourself cuz you most naturally if you're authentic and so on, you'll naturally make trade-offs that seem effortless to you, but that reinforce your culture.
51:52 · And then that becomes this very hard thing for other people to catch up to. And at Anthropic, from day one, you know, there was this mission-like missionary-like zeal and belief that hey, these capabilities will scale. These systems are stochastic, not deterministic. There will be error bars. And until we crack interpretability, there's risk.
52:14 · Yeah.
52:14 · And at some point, people will go stop using Claude just for coding. They'll use it in some mission-critical context where there's it'll throw off a bug. And then people are going to come blame them. And they want to be on the right side of history where they said, yes, this is a powerful technology. We think it's going to change the world.
52:31 · And we want to be very measured and scientific about the fact that hey guys, these are stats models, statistical models. That's how statistics works. Like ultimately, when you're training neural nets, it is just a statistical system.
52:44 · And I think that that belief that safety is important and that it might seem toy-like in the early days and sometimes there you could say Anthropic is they totally over exaggerated the risk, you know, like 2 years ago when they said let's not launch Claude 1 or whatever. Well, okay. Maybe in hindsight, but hindsight is 20/20. And at the time, they didn't know how that model would be used.
53:08 · And to them, it felt existential if somebody came and said, you weren't responsible. It this wrote a bug. The liability associated with that is massive. So, how do you prevent against that? Well, day in day out, you say safety, safety, safety, safety. And when you start deviating from that, you have the team hold you accountable. You have the world hold you accountable.
53:26 · And I think that becomes a moat over time. At some point, that moat will get challenged and so on, and then it will become fragile. I hope it endures because that's the beauty of having founders run the show.
53:38 · Cuz they can make really hard trade-offs to do mission alignment. The hardest part is in the earliest days when you don't have a group of people who are going through difficulty, stress, crisis together, then your your culture doesn't get defined sharply enough. And that's what I'm worried about right now. Is there so much money going to these labs, there's no hardship. There's no 21 knows.
54:03 · There's no 21 knows. And that in hindsight was a feature not a bug for Anthropic. The number of people who said no, the number of people who said, "Sorry, we're already investors in OpenAI."
54:14 · That is competitive defense. It forces you to really understand like what is the hill you want to die on at the expense of everything else. What's the P0?
54:23 · And their P0 from day one was coding. The reason mechanism system there was if we crack coding, then we will crack AGI. You know, our mission is AGI. We want to get there safely. If we focus on coding, it's such a generally powerful capability. That it can accelerate all kinds of work on a computer. And if we can accelerate all kinds of work on a computer, we can get to AGI.
54:46 · You know, as a result, they've had to say no to so much other stuff. Here, superconductivity is the mission. Coding is not the mission. So, we use Claude. We'll use Claude. We don't care about that. The mission defines everything. And I I think teams who can raise too much money too fast too early, who don't have to define what the P0 is, because that's the only thing when you have scarce resources you got to Yeah.
55:08 · you got to invest in.
55:10 · Those cultures end up being the most fragile and brittle. And they almost don't even make it to take off.
55:14 · So, let's apply this to Periodic since we're here. Sure. What is the constraint or the hardship that they were forcing themselves to go through?
Periodic Labs, Physics, and Silicon Valley Mercenaries
55:21 · Dude, here? Here? Are you crazy? No.
55:23 · [laughter] Well, yeah, okay. So, on a technical level, it's physics. It's literally reality.
55:28 · I mean.
55:29 · But like is there is there is there another one that's like the company building Yeah. When I I Liam was a co-creator of ChatGPT. And Doge was skip level from Demis at DeepMind.
55:41 · Had created, you know, Gnomics, one of the one of the most important tools to come out of DeepMind. You know, at the time I was a visiting scientist at the Stanford physics department, and we had started benchmarking frontier models on physics and science capabilities, and they were not very good. They were good at like doing things like summarization of papers. But if you said, "Hey, could you like analyze the scientific data coming out of a condensed matter physics lab?" I was I was in the condensed matter physics group at at Stanford.
56:11 · They were terrible. So, it was not popular 12 months ago. You know, PR I could have And I won't go into details, but you know, there were people who said you know, as recently as a few months ago, who said they wanted to join the company. And they, for whatever reason, you know, took a job elsewhere. They kind of reneged on their commitments.
56:33 · They took a job elsewhere that offered more money. Then we had a technical breakthrough. Created a soda system and like Okay.
56:41 · I was excited.
56:42 · Yeah, yeah, we'll cover it. We'll we'll be doing a separate part on on periodic.
56:45 · And then they wanted to come back. And I said, "No. No way. You If you come here, you had your shot. You you had your shot.
56:52 · Cuz it's actually about culture.
56:53 · Of course.
56:54 · principles, yeah.
56:55 · You know, and look, I believe in second chances and so on, but time will need to heal. Some of those wounds were they will leave deep deep for them will leave deep scars, but because I started my company at 24, 25, I had I went through the whole cycle of betrayal and drama, and so you realize you know, Silicon Valley is both a very missionary place, it's also a very mercenary place.
57:17 · Um sometimes people lose their minds with with a when big money gets involved, which is in the grand scheme of things quite small money.
57:26 · Like you know, we I I guess you're taking Life-changing to me, maybe less to you, but you know, like a lot of people have not been taught how to deal with money. And yeah, we didn't come up from like that privilege of a background, right?
Rishi Valley, Singapore, and Money as a Measure
57:38 · a street dog, man.
57:39 · I look, I grew up in Rishi Valley. We We didn't have like This was enforced brutalism. Jiddu Krishnamurti started the school was like you will sleep on a hard slab of stone. Like the mattress was this thin, you know? And then you go up in Singapore. When I got to Singapore, I used to sleep I was a part of the scholarship program, but which which was amazing. I'm very grateful to the Singaporean government.
58:02 · But I was at St. Andrew's JC and our dorm, which was by Boon Keng, you know, MRT, Which is not a prestigious neighborhood.
58:12 · [laughter] It was a It was a transition dorm because you're building this beautiful like residential campus on site at SAJC in Potong Pasir, but the we were the last, I think the second last batch to be in the transition site, which was some old like I think I think it's like an immigrant labor Yes. That's where we keep the people who work on the the factories and stuff.
58:35 · Right. [laughter] So I lived in a my 11th and 12th grade, I slept in a bedroom the size of this like literally from from there to here, Yeah.
58:47 · right?
58:48 · They were like bunk beds. And so one bunk bed here, one bunk bed there, one on top, one on top, one more here, and then here was where our like we kept our toiletries and clothes and stuff. And when one guy would climb onto his bed there, this one would shake.
59:04 · Oh my god.
59:05 · And one of my roommates who was from and it was amazing. I loved every minute of it. You know, my my roommates were a guy who was a top-ranked Dota player from BRC from China, didn't speak a lick of English. Uh loved him. Amazing guy.
59:21 · I mean, all the Singapore scholars are fantastic. And honestly, we should treat you guys better cuz of what you're going to do, but [laughter] it's cool to know.
59:28 · No, I mean what I'm saying is I don't need much to be happy in life. You know, when you've lived through that, money is a way I think sometimes we measure ourselves, but you know, when it's when when it stops becoming, you know, it works good hearts law when it stops becoming just a byproduct and more of a measure, it stops having meaning.
59:50 · You use it to do more meaningful things. It's resources to pursue your mission. Uh I've kept you longer than I am supposed to, but we should continue this in the next part two.
Closing Thoughts
1:00:00 · know, I really enjoyed this. Yeah, yeah, I mean, you're you're so inspirational and yeah, there's more I want to dig into about how you've like set everything up, every single one of your investments, how amp is going, but we don't we're running out of time for that, but thank you so much for joining us.
1:00:13 · to see you, man. Let's get chicken rice sometime.
1:00:15 · Yes. I'm actually tomorrow. I'll send you I'll send you the details. I'm hosting a birthday party.
1:00:21 · I don't get an invite?
1:00:22 · to be a Singaporean birthday party, yes.
1:00:23 · Yeah, you get an invite right now.
1:00:25 · Okay, cool.
1:00:25 · [laughter] All right, thank you.
1:00:26 · All righty, thanks, man.
1:00:29 · [music]