The Next 3 Years of AI: Lessons from Elon Musk’s First Investor — Silicon Valley Girl Podcast

Marina: What will the next three years look like?

Steve Jurvetson: I have this gut feeling that it'll be something architecturally variant.

Marina: This is Steve Jurvetson, an early investor in SpaceX when almost nobody believed in private space. He backed Tesla before electric cars went mainstream. For over 30 years, he's been betting on the future, and history keeps proving him right.

Marina: When someone comes to you with just an idea, what will be the best thing that they can do?

Steve Jurvetson: Single person with an idea? I might try to find a co-founder. It's rarely a individual. Jobs and Wozniak, Batman and Robin.

Marina: You're working a lot with Elon. Top three principles that everyone should learn from him?

Steve Jurvetson: I do try to observe leaders in action. Even with a focused effort, it's not always obvious, but a few things. One is this insane ability to

Marina: Thank you so much. This is going to be very exciting. Steve, I am so excited to have you on this stage. What a fun time, right? SpaceX IPO, and you were there super early. What did you see that most investors didn't see back then?

Steve Jurvetson: So, the simple answer to the question is there were almost no investors considering space. It wasn't a category on any on any site. So, the slightly varying the question is why in the world would we invest in a sector that is just not a sector for venture. Same same could be said for automotive with Tesla, um energy with nuclear fusion. There were, you know, a handful of investments with very few. So, the short version is obviously an incredible entrepreneur, um someone we've worked with before. I've known him for oh gosh, 29 years now. And uh invested in all of his companies of the century and his cousins, too. Um

[laughter]

So, uh all in, if you will. Um the uniqueness of the opportunity. So, what we've come to appreciate in a sort of fuzzy way then, but now more in a more crystallized manner is the way in which a sort of software-centric system engineering approach to to sleepy industry that hasn't had any change for decades can actually unlock incredible value and opportunity. You can see it in aerospace, you can see it in automotive now. It was sort of a long bet when we first invested, but now we can sort of see in in retrospect how that's going to play out in almost every industry over time. How they become information businesses.

Marina: Obviously, you're so good at predicting future and part of this podcast I really want to understand how you think about the future. You have this amazing graph, 130 years of compute and it basically grows exponentially. What does it mean for all of us? What will the next 3 years look like because of what's happening to compute?

Steve Jurvetson: Now, I'm just curious, how many people have seen this version or this abstraction of Moore's Law that was originally by Ray Kurzweil in like '99 book Age of Spiritual Machines? Looks to me like 25% of the room. Okay, I always ask cuz I'm curious how much it has entered the zeitgeist cuz I think it's the most important thing ever graphed and I give credit to Kurzweil for even seeing this pattern back when no one knew they were fitting to a curve. So, just for those who don't know, this curve is like five different technology substrates from mechanical devices to relay-based computers to, you know, discrete transistors and integrated circuits. And only in the most recent era with what Gordon Moore called Moore's Law be almost a refraction of a much longer-term trend transcending you know, all kinds of dramas of companies that came and went. It's almost cosmological. Like, why has humanity's capacity to compute compounded for 130 years? And for sense of scale, like that's an exponential scale, right? Logarithmic scale, so straight line is exponential. This graph shows a 10,000 billion billion X improvement in computation that a dollar can buy. This is what customers care about. No one buys transistors when they're buying ICs. They don't say, "How many How many transistors does that one have? I'll buy the one that has more." No, they buy compute capacity or memory and both have been on rails. And so, to your question, the first and foremost thing would be to just predict that it's going to keep going for 3 more years. Like, why would it suddenly just stop and hit a red brick wall the way Intel's been saying it would? And when companies say that, like Intel usually a sign they're losing their business to someone new like Nvidia 15 years ago. So in the next 3 years I think you'll see the analog chips continue to carry the mantle of Moore's law. Some of the more esoteric um and customized AI silicon that does discrete discrete matrix multiplying and add really efficiently. And this is what's going to carry the juggernaut that we all just take for granted that it keeps going. In fact, I'd say without this sort of exponential change in technology, you wouldn't have startups, you wouldn't have this disruptive innovation opportunity like we talked about in SpaceX or in a bunch of companies because if business is predictable, if there isn't disruptive technological change, the big get bigger. You know, they have all kinds of ways to prevent new entrants from competing with them. It's usually somebody reinvents an industry and usually it's based on something computationally based. I think AI and everything we're talking about today's conference is the epitome of this. It's like the most intense crucible of compute-centric innovation, economic growth, and and and sort of innervation of the economy. The the translation of formerly industrial crappy growth margin businesses into information age information-centric businesses that over next 3 years means it ripples through I think energy, agriculture, construction, three industries that are enormous, growing as a percentage GDP, and the least digitized industries on the planet. Not to mention healthcare right soon behind that.

Marina: Are those the industries where you think we're going to see the most change? And what will cause the change? Is it going to be a more advanced LLMs? Or do you think there's something else? I don't People are building world models. People are deep into robotics. What will be the technological driver for the most changes in the next 3 years?

Steve Jurvetson: That's a great question because it's very difficult to answer with any certainty. I I have this gut feeling that it'll be something architecturally variant. It might subsume the models that we know now. You can almost think of like a mixture of experts that's subsuming other architectures or the diffusion model we heard about earlier today that ultimately translates to a transformer, but it's a different way of thinking about the transformer, massively parallel form of a diffusion model. Um, in the back of my mind, I we have not So, what I'm about to share I we've not invested in this. So, I'm I've met with some companies and I've been intrigued and something my gut says they're going to they're probably going to make a breakthrough and this is the whole new generation of neural labs focused on reinforcement learning cuz we're almost going back to the founding premise of DeepMind, um, which then they kind of, you know, just put to the side for a while when the whole LLN thing took off. And so, um, if you can imagine what would be the You could phrase this in a agentic language and say, "What is the, you know, multi-decade long agentic process?" Yeah, minutes or hours, not driven by some outsider, you know, pulling puppet strings, but something that says almost like the drive evolutionarily for creatures or for humanity for whatever we consider the mission statement of our lives or humanity in general. What would be that thing? Is it, um, you know, to understand the universe the way Grok and xAI says it? Does that become a driver for an artist like that? Is it something like a novelty-seeking algorithm that says, "I'm going to continue to learn about the world and use novelty as my filter for Oh, I just discovered something new. I How do I know if I'm making progress?" What is, if you will, the selection pressure in an evolutionary algorithm? What is success? It's not just reproductive fitness in the biological sense, it's it's something grander and I I know that some of these groups are working on what is Is there a single reinforcement learning algorithm with continuous learning let loose in the wild with all the data sets of the internet that could bootstrap intelligence in that sense in the way that we think we're seeing in the large language models today, but it's it's largely we ascribe I think consciousness to other beings, we ascribe meaning to other things, and we see patterns where they're not. And so, I think a lot of it is a bit of a um it's a fun interaction, but it's not quite the same thing, right? We just We know there's nothing there inside. There's no light on inside, if you will.

Marina: So, are you just describing I think is it superintelligence when it's learning by itself, setting goals to itself? Are we going to see some version of that in the next 3 years?

Steve Jurvetson: I know Jack Clark, uh co-founder of Anthropic, gives it a 30% chance it happens next year, um which I think is kind of

Marina: Superintelligence or just

Steve Jurvetson: Yeah. Absolutely. So, I thought, well, that's kind of fun. Um there's at least one person putting a stake in the ground. I don't know. I don't have

Marina: they have a lot of strong opinions.

Steve Jurvetson: Well, they do, but they also think that they're on the path. And And there's a big debate as to whether this recursive self-improvement thing that they wrote about today and that Jack's been talking about for a few weeks now. I spoke with him about it last month or 2 months ago. Is there going to be some leap that we don't currently see for how these systems take on purpose and meaning in in the what I was referring to just a moment ago. Because right now, everything that they do is directed by a human. There's like, yes, the self-improving AI loop that they're witnessing already, these huge improvements, are coming from a number of steps that are still directed by humans. There's, you know, automated verification, improvement loops in the process of training itself, you know, adjusting hyperparameters from one training run to the next. A bunch of ways you could imagine hyper word experimentation being mediated by the AIs, but what is the goal? The goal setting is still by the human. And there So, there it may only be a thin veneer of activity that it's not yet doing, but it's in some ways the most important, right?

Marina: Yeah.

Steve Jurvetson: Um And And they'll admit they're not sure how does that just happen, right? What What makes that transition? And I don't know if it'll need to recapitulate some of the um functional specialization of our own brain. Like, we evolved to where we are today with a history of reactive limbic systems and what have you, emotional centers that then the cortex and more and more cortex layered on top of it, that whole construct may, as we heard in earlier speech, be the bootstrapped consciousness as a perception of what we perceive. Do we need to have the same things in our robotic {slash} AI systems, right? There There may be So it's a philosophical argument. The main answer to your question would be I do not know. Um and I don't really even have odds on it. I give it the fuzzy future kind of it Yeah, that might happen. But only because that's more convenient as an intellectual shortcut to actually thinking about it as a as a serious hard problem. Um is to put off that 3 years feels far enough in the future that it's hard to predict almost anything.

Marina: So we're seeing all the demos of robots and current technology, I think, is stronger than the deployment itself. We're still adopting, we're still adjusting. What's this gap? How big is it? From From what technology is actually capable of versus how we're using it?

Steve Jurvetson: All right. Yes, that's a very good point. Um and there'll be inherently very differential domains of acceptance. So here's a great e

Reed Hoffman: example, very simple to understand is if it involves the world of atoms, it takes time. So even though it is obvious today that fully autonomous vehicles are the inevitable future, that every car will be autonomous, every train, every airplane, everything that moves on Earth will be fully autonomous in the future. How could it not? It's It's insane to think now we'd argue that it's not, even though we've been saying this for decades. Um the the pace of switchover is going to be It's going to feel glacial in certain parts of the world, right? The people will keep cars for an average of like 11 to 12 years. So you just have the physical swap-out cycle for the car cycles. You have you know, the change in mobility doesn't happen overnight. Okay, that's an obvious one. Physical robotics might be the same. How long does it take to make a billion robots? That takes some time, even with recursive uh manufacturing techniques. Um And so the place where I think it just sweeps like wildfire can be in areas strangely uh that we sometimes held as uniquely human are the creative arts, uh the you know, the movie making, the images, what have you, uh which we've already seen. So, it's somewhat shocking that that came first. Um and then the white-collar jobs I was mentioning because a white-collar job um capability Take call centers, right? It's like 1% of US GDP. That like that just happens like that, right? I mean, you just do not need uh to uh wait uh for decades for that to switch over almost entirely. And interestingly, people um will increasingly prefer these to human interactions when they're better, show more emotional understanding, um more reading of the situation, and that's seen in everything from physician bedside manner um to uh to chatbots and or or customer service agents that the AIs do a better job with emotional connection than humans.

Marina: It's crazy how in some industries it's happening super fast, especially when it comes to software engineering. Some of my friends were editing 70% of AI-written code a year ago, now it's down to 30%. I wonder what it's going to be in a year. I am constantly looking at our AI stack. What can we make faster, smoother, more useful for the team? So, I want to tell you about a workspace we genuinely keep coming back to, and it's also sponsoring this episode. That's Miro. Miro is the AI innovation workspace where AI lives on the canvas, not in a separate chat. It sees the whole canvas, every note, every source, every decision the team already put there. We've picked a few specific use cases where it actually helps us. One of them is guest research. The team drops everything onto the board, and it re-transcribes articles, podcast clips, a dozen sources on one person. Then we run flows with a custom sidekick that reads all of it and pulls the angle, the best quotes, the questions worth asking. That used to take a full day. Now the guest dossier is ready in 30 minutes, which means my team and I actually have time to prepare properly for the next interview instead of just surviving the deadline. Especially now when my schedule is so crazy, I do four to five podcasts every single week. And it works for a lot more than guest prep. The same setup turns a messy retro into an action plan or planning session into a prioritized backlog. You can now build something similar yourself in Miro. The link is in the description.

Marina: So you worked with some of the most amazing entrepreneurs. You're working a lot with Elon. I know a lot of people in the audience are builders. Is there anything, like maybe top three principles that everyone should learn from him?

Reed Hoffman: It's funny. People have been sending me these books that just I guess they directed an ad or write about AI, you know, about Elon how he thinks, the secrets of Elon. I've actually been accumulating them on my bedside, but I'm not sure if a human's written any of them. A lot of people ask, you know, Elon's mom, you know, "Hey Maye, how did How did you How did you parent Elon? How did you get him to be the way he is?" And that's a tough question. She hasn't been able to answer either. And so I'll take it with a bit of humility that uh even as a close observer, by the way, I do try to observe leaders in actions. I worked with Steve Jobs briefly briefly and it's like I put all kinds of energy in trying to understand how that guy works. Um but even with focused effort, it's not always obvious. People are complex. So but a but a few things. One is this insane ability to focus, which may seem ironic given how many companies he's simultaneously running, setting new records for that in a way that you know, when Steve Jobs was CEO of two companies, that seemed strange and now it's all the rage. But one thing that allows you to do is use the fact that you've got obvious competing needs for your attention as a way to focus, prioritize, and not go to meetings. If a normal CEO of one company didn't go to their holiday party, you know, and it might be seen as weird and like, "Whoa." But that No one questions that of Elon. He's got other things to do. He's got other companies. So what that Whether it's an excuse or just works out this way, he says no to things so effectively that are distractions that are not critical mission-critical right now. I mean, for example, years ago I was trying to hook him up with Craig Venter to brainstorm ways we could, you know, terraform Mars more easily and do a sample return of life from Mars with gene sequencers and reinstantiating. Anyway, microbes on Earth. It was a fascinating topic to me. I was like, "Wow, this is so fascinating." But he's like, "No, it doesn't matter until we get Starship flying. None of this stuff on Mars matters. I got to get that thing working first before we think about what we do when we get there." There's, I think, maybe more importantly than what I just said. Even more importantly is this in in maniacal focus on the uh what I would generalize as the cycle time of innovation, which is how rapidly can we run experiments or iterate in our learning loop? What is the core learning loop? Whether it's the launch cadence, um whether it's um the data gathered from all the Teslas before fully self-driving vehicles came that could be used to train the models. How can we make sure that we have a leg up on anyone else on the rate at which we're learning from customer interaction, product features, and and technology in general. And uh as an example of like how powerful that is um when you do it right in the data flywheel you can get for make for AI. As just one example, the Tesla cars today in their cameras gather for their AI training set more data every 4 days than Waymo has in its entire history. And the brilliance was enabling every vehicle, whether or not the customer paid for full self-driving, to be a data collecting vehicle. Okay. So, focus, learning loops, and this um a whole series of well-honed skills on uh identifying talent that um I wish I could replicate. I just can't. It's like there's Sometimes there's a pattern recognition um and he'll share bits and pieces of this like you know, not not leaning on credentials or specific background or experience. In fact, it's often an albatross. But like having people really walk through major uh engineering crises or problem-solving things and and drilling down further and further and further to show if they did they really master the do they have mastery and understanding of what it took to make something successful. So, broadly defined, um being a magnet for talent, finding a way to pitch and and and refine a vision that people want to join you. So, like one of one of his brilliant things at Tesla, SpaceX, everywhere is not just saying, "Oh, yeah, we're making rockets or we're making cars." But, to really think of something much grander, right? Catalyzing the, you know, transition to sustainable energy or making humanity multi-planetary, uh understanding the universe when you know, now that XAI is merging into it. Uh these are the sort of lofty goals that that that motivate some of the best and the brightest to want to work with you and that is a is sort of compounding benefit that ripples out through the whole organization, right? Because great people want to work with other great people, right?

Marina: I'm talking [clears throat] to a lot of entrepreneurs and especially these days with things moving so fast, there's this new shiny thing every single week. How do you stay true to your mission when the rest of the world, 99% of the world, tells you of it's too early, like talking about space, we have so many problems here on Earth?

Reed Hoffman: That's an interesting question and I realize I have a bit of a sample selection bias in that I've tried as best I can and I've done VC now for 30 years to only work with the people who have a true, sincere, you know, messianic mission in mind that is driving them and they're not the arbitrary seeking opportunists to see the next bright shiny object or oh gosh, um you know, where where should I go to next? And one of the ways and I'll get to your question, but one of the ways that by the way that I filter for that in meetings is let's say we're getting really excited about a company, I'll often ask, you know, okay, what does your business look like in 50 years? And I get usually two reactions most often. One will be a chuckle, like, "What a ridiculous question, you know, like the arbitrary seeking opportunist is going to be like, 'I'll be in my third startup by then, like what what what how would I possibly know what my startup is in in years?'" Like they just laugh at the question. Uh and then we pass on those. And then the best is when the person's like so relieved like oh thank god. Now I can actually tell you what I've been wanting to say all day long which is this is what's driving me. It's this thing that's so many steps ahead of what you would probably want to invest in today. Like making, you know, colonizing Mars is an uninvestable proposition. Go back in the founding days. It's like when you start a business day one I'm going to colonize Mars like you know, next, right? For most investors, right? Like that's not a door opener. And so most entrepreneurs that have that true sincere vision have found a way to like subjugate and put off what their true dreams are and talk about something much more prosaic and and near term. So I think the answer would be as the entrepreneur it just happens naturally and try to find investors and partners and certainly employees who are with you for that long ride and um have a path to get there that is plausible. So you know, this is sort of the joint tension I think in the best startups that's hard to simultaneously satisfy which is an audacious, you know, 50 to 500 year vision. This is [snorts] what this company's going to do to the economy or the universe um coupled with oh and by the way over the next three years we're going to iterate with real customers, learn from that and have a can paint the path from where we are now to that future that is uh chaining sometimes they chain back from the past to the present like to get there what do I have to build now and then and then move forward along that path but it's not like go into a research lab pop out in 20 years and you know, solve all the world's problems.

Marina: Yeah, this is a really fascinating feature that I see with a lot of greatest entrepreneurs. It's like if they're reverse engineering from 50 years ahead. Is there anything surprising that still surprises you about those amazing entrepreneurs?

Reed Hoffman: Well, I suppose it's a bit surprising in a way each and every time it goes incredibly right. Um so weirdly, this may sound weird, I don't think I've ever thought about that question before um or been asked it before. And so, the perpetual surprise for me is like, wow. Like in the year 8, 9, 12, some new opportunity that opens up and unfolds from the in a sense the expanding option value of

Reed Hoffman: going into some new frontier of the unknown. So, what I mean by this is we try to invest in by the way, at at our firm Future Ventures, in things that are unlike anything we've seen before yet adjacent to where we've been. So, ideally, it's a company that's literally one of a kind based on things we are used to, whether it's AI, whether it's something synthetic biology, whatever it might be, but they're taking it in some new direction. So, the the window in as long as you're have an agile mind and you're looking at it, wow. Like no one thought of that when we started. So, for example, when Tes- when we first invested in Tesla, there was no concept of whatsoever autonomous driving. Was not in the business plan, there was no talk of it, was not in anyone's mind. The way in which electric drivetrain uniquely enables that and control of fidelity was fascinating. Or in SpaceX, the space the Starlink, you know, opportunity, like, oh yes, of course, when you lower cost of launch that much, you can have mega constellations, but what would be the new thing that that would make sense that we weren't doing before? Not just we invested in Planet Labs for earth observation, yes, constellation of telescopes, but this whole notion of building a a network um you know, backbone for the internet in the sky was and then direct to cell phone. Like each one of these things has an unfolding then to then orbital data centers, right? Not on the dance card even 5 years ago. So, uh that continues to surprise me. In some ways, it it's not easy, but it it seems so much more powerful as a business vector than purposeful design, if you will. It's almost like exploring the the the option space that that or the light cone, if you will, of possibilities in an economy versus, you know, sort of planning out something 10 years in advance and having it go according to plan, if you will.

Marina: It is so fascinating how you were successful in so many different bets that you made in the past and they're so different from each other in different industries. What are you betting on now? What should we be looking out for?

Reed Hoffman: Plastic snow. Let me think that.

[laughter]

Reed Hoffman: Some people who know the old movie. Let's see. So we So taking that thesis that AI and information technology will innovate every economy at meaning at a nervous system to everything. We saw in automotive and aerospace just expanding on that thought a bit. We are looking for additional things in energy. We've invested in a variety of nuclear fusion and and and subcritical fusion that doesn't trigger NRC regulations. Basically avoiding the nuclear regulatory commission but figuring out energy, which by the way is the third bottleneck for AI. It's not just good people and a lot of compute, it's also energy. There are a bunch of things that you could imagine 500 years from now have been solved and we're trying to figure out the entrepreneur will open our eyes to how we get there. So free healthcare forever via a cell phone, all diagnostic information you could possibly need for your personal health should be a free service globally. Trying to figure out how to get there. Probably won't be in the US that it launches. You know, bypassing FDA, bypassing insurance and reimbursement. On food, we won't slaughter animals for meat. It's the the products are getting there but the the you can sort of see the future. It's so close, you can almost taste it, so to speak. Whether cellular ag, mycelium, or other techniques, mycelium being the fastest growing thing. But but we are going to eat meat-like things that are delicious, healthy, and not involve slaughter of animals. Uh construction growing as a percentage of GDP and like labor productivity has been flat for 30 years. So that it's such a hard industry to change that we've tried and failed a few times. But we're looking. Again, so what the best I can do to answer your question is I don't know what the answer is but I know there is categories that don't we want to look at. Recently we've been investing in epigenetic editing across a variety of things from crop health, pesticides, herbicides, um uh, uh, human health. It's it's fascinating. It's basically the software of biology instead of going to the firmware of our our genome. And uh, we've been investing in materials, critical minerals and metals, uh, everything from deep sea mining to copper refining, uh, because of this incredible need. It's sort of like the workhorse of all these chips is you need these materials to make the stuff. And there's couple of that, a reshoring or you know, bringing back to the US capacity to build that which we had uh, atrophied over many years. Analog AI I mentioned it, we have three different investments coming at it from different angles using AI to develop to design AI chips, uh, sorry, analog chips.

Marina: Analog chips.

Reed Hoffman: Yep. Uh, analog, um, in memory compute from Mythic where they can do 8-bit multiplying that in a single transistor. And then unconventional which is taking a very very strange and forward-looking, um, big bet on you know, in every case trying to get 100X and then another 100X on power reduction, power per per per calculation. Overall, we're about 40% life sciences, 60% IT. And and we in the life sciences decide just see we look for the weird things that are like on the edge, you know, harvesting organs for transplant, you know, growing humans without brains so that you can use their organs. Um, there's a company here actually in the audience doing the same thing. Um, a male birth control pill, uh, you know, improving IVF dramatically, um, blood things that fall through the cracks of a traditional pharma VC.

Marina: Some hearing agriculture, uh, biotech. I'm just thinking in my head how can I replicate your strategy with ETFs and all that?

Reed Hoffman: Well, it's hard to rep See, our strategy I can say it, it's very unusual. I can say it openly and then it's hard to replicate cuz when I say we invest in things that are unlike anything we've seen before, well, that's great. But how do you know what we've seen? So, you know, what we're actually doing is not as good.

Marina: areas. So, if they are dramatically changing a market, then it's going to be a reflective

Reed Hoffman: Especially if it's an old crappy business that hasn't seen a new entrant in years. So, like Boring Company for tunnel boring machines. It's like like the four largest companies were all started in the 1800s. That's who you're competing with.

Marina: We have a lot of entrepreneurs who have crazy ideas. Can you give them a 30-day plan to execute on that idea? What would be the best thing that they can do?

Reed Hoffman: What stage are they are you seeing them?

Marina: have an idea.

Reed Hoffman: Oh, single person with an idea?

Marina: Yeah.

Reed Hoffman: Mhm, 30-day plan. I might try to find a co-founder who agrees with you or whoever this person is. And the reason I say that is a lot of startups tend to have a dynamic duo at their founding. It's rarely a individual. Um and in you can you can imagine Jobs and Wozniak as a mental model for this or these superheroes, Batman and Robin, uh you know, Sergey and Larry Page. Even Larry Ellison had Bob Miner, who's less well known cuz he's an introvert, but you know, there was not like a singular cult of personality of a founder. And And part of the reason to have someone is I found this as an investor, having a colleague, Mariana, my co-founder, is I am so much better as an investor having someone to bounce ideas off versus like being the sole, you know, like an angel investor or something. And similarly for a startup, having a diversity of backgrounds, like an engineer and a marketing person, an extrovert and an introvert, whatever it might be that have mutual respect for each other, not only makes it better that you like some you got someone to bounce ideas off of in a rapid iteration loop when there's just two of you, but it also sets the culture for everyone that you'll hire. It's not like, oh, there's a singular person that everyone works for. It's more like there was a pair and they're very different and that that ripples through the culture of a firm and types of people that are hired and the and the cognitive diversity that follows. So, finding And the reason I say that is finding someone who agrees that your crazy idea is worth pursuing is better than finding zero people. In other words, uh I think the best outcome is if you're literally your premise to your question, crazy startup where no one else is doing it, it's one-of-a-kind, and most people tell you it's crazy. Well, it is possible that it's crazy, right? So, if 100% of people that you've ever met think it's crazy, take that as feedback. If it's, you know, nine out of 10, that's pretty good. If it's eight out of 10, that's pretty good, too. If it's like only two people think it's crazy, that's bad because it's clearly not bold enough. If it's an obvious idea, other people will do it, right? And ask yourself, is is this a business that couldn't have been started 3 years ago? If the answer is yes, that's good, right? If it's like, oh yeah, I know it's Anyone could have started this business if they just had this idea, probably a bad sign. And then somebody, that your co-founder, uh agrees with you and thinks, oh my god, this is me That just shows that as almost as a kind of test case, you can persuade persuade someone to give up their job and join you in this mission, then before you go out and fundraise, that that says a lot more than just this old person with an idea, like the inventor in a garage, you know, off all by themselves. There's so many cases like that that just never manifest as a business because they just never made that first step of being able to persuade anyone to join them in the mission.

Marina: It's great advice because a lot of people start with building an MVP or like even pitching investors right away. The co-founder sounds incredible. What from all the startups you founded, where did the best co-founders meet? Is that university or

Reed Hoffman: Yeah, good question. I don't know. I'm not sure. I haven't I haven't thought through that. Um

Marina: Cuz it's so hard, especially

Reed Hoffman: often they often come to us having already done that. And often, yes. So, for all the university ones, they're many of them are from That's probably your your question had embedded within it the most common answer, which is you know, we met in some interdisciplinary way at a university, which is fascinating, by the way. The word disciplinary, you know, or disciplines, academic disciplines, are a way of stovepiping information into a system's vernacular and domain expertise that often doesn't cross-pollinate. And universities one of those few places where you get these cross-spanners, you get these undergrads or other people who take courses outside their department, unlike the professors in their little stovepipes. And despite a lot of institutional efforts to share information, it's often the students that are the cross-pollination between academic disciplines. And that's at those boundaries or interstices between formally discrete disciplines that you find, I think, most breakthrough innovation, certainly in the sciences. As a quick aside, that's something that large language models do very well, translating between academic domains, seeing patterns in the, you know, almost the translation, if you will, between languages, between concepts. And that, I think, is allowing a fountainhead of possible idea discovery using AI to figure out new ways of cross-pollinating between academic disciplines that I think we're only beginning to tap into.

Marina: This makes total sense. Um I think I can be talking to you for hours because you are someone who's really good at predicting future and betting on it and seeing where we're going. I have one last question before we open it up uh for Q&A. When machines do everything, what's the meaning of life?

Reed Hoffman: Yeah. Uh yeah, and your question, I think, is an interesting one to contemplate. What do we do when m

Steve Jurvetson: achines do everything that we do better than we can. Every physical activity, everything that involves employment, um and it's going to come soon, right? Roughly 19% of global employment is in driving vehicles, um and that's obviously going away. It's just not as rapidly as we might imagine. Um I think we we want meaningful work. I think all humans have a fundamental desire for symbolic immortality, this belief that we've contributed something to the world that transcends our brief time on this world. And uh we see that, of course, in the drive to have children or in writing works or in philanthropy or creating companies, sometimes even named after their founders, like Hewlett-Packard or what have you. These are instantiations of that urge. And so, I think there's still a creative desire um and and I can translate the question to be like, what is the mission statement for humanity? It's a question that Yuri Milner and Elon Musk and others have asked and they come to similar conclusion, which is to understand the universe, to try to contribute to the wisdom, the accumulated knowledge that we have. You could think of human culture and our knowledge base that we pass on from generation to generation as the primary vector of our own evolutionary progress. It's not biological evolution, that's glacial in comparison. And any progress we feel humanity is making is not because we changed our biology, it's because we changed our accumulated basis of knowledge, the the way we comport ourselves, the rule of law, the understanding we have around what works and helps with human flourishing. So, I think we all want to contribute to that. Um it doesn't have to be paid employment though. So, I you can't imagine some sort of hyperspace jump cuz that's conceptually what it requires cuz there's no way to imagine how we get here from here to there. But somehow if we just jump there to world of abundance like Peter Diamandis envisions, um you know, everything physical costs a dollar a pound, there's nothing that requires human labor, we all are in the indentured rich like in the days of yore we had, you know, servants or serfs or slaves that did all, you know, menial work and we could just be, you know, philosopher kings or artists or pursue whatever we might want and and and some people some, not all, but some people really love that era. Well, the machines will be those slaves, right? Not because even in slavery humans will not be cost-effective. And I say that somewhat tongue-in-cheek, but it's like finally the scourge of human slavery might finally end when that's no longer even cost-effective compared to machines. Um what does that leave for the rest of us? And so, I think it's going to be a man's search for meaning that really is the core question. Um I think it's going to be a really fun if we could hyperspace there, but I will add the caveat that's not the path we're taking. Like there's nothing that indicates that we're just going to peacefully march from an economy of full employment to an economy of no employment and pass through the 30, 40, 50% unemployment points. Well, that's going to be turbulent. That's going to be tough. And I don't see any politicians taking long-term perspectives on any of that. So, I don't want to end on a downer. Let's go back to that hyperspace to abundance. Um I think I think we inherently find that in our curious exploration of the of the universe.

Marina: Yeah. I really like the rule of going back to your mission statement cuz a lot of us these days are questioning our jobs, what we're doing, is it going to exist in the same shape and form in 3 years? And again, going back to your mission statement, I think this is brilliant. Thank you so much, Steve, and let's open it up for Q&A. Quick pause here. If you're enjoying this podcast, you will absolutely love my Inner Circle newsletter. So, what I basically do is I take all the tips from these podcasts and apply them to my personal life, to my investment portfolio, and to my businesses, this media company and my language teaching business. Sometimes we get amazing results and I share our real tactics. Sometimes we don't and I share that, too. Think of it as an insider version of this podcast. The link is in the description. Join my free newsletter to stay ahead.

Marina: Thank you, Steve, for your fireside chat. Uh since you're a big investor in Elon Musk's companies, um I'm curious, have you invested in Neuralink?

Steve: Yeah.

Marina: Yeah, so honestly, in my opinion, everybody is excited about SpaceX, but I'm looking forward for an IPO of Neuralink. Do you think it's happening soon? And honestly, I think it's uh basically brain-machine interface is the future. Cuz essentially, currently, if we use the voice mode on ChatGPT or OpenAI, right, or we type, we're limited in our throughput of how many tokens we we uh we send to the LLMs. And if we have brain-machine interface from Neuralink, we're able to unlock even more creativity and faster throughput from our brain to machine.

Steve: Yeah, I can't comment on IPO timelines, um but um the enthusiasm there is interesting. It was originally sparked, as many things are, from um a science fiction novel, Ian Banks' Surface Detail, where there were neural lace. Fascinating book, I recommend it. Um and I think what you see

[sighs and gasps]

So, I have a I have a somewhat unique perspective not shared by Nerualink. So, I'll but I'll just share my perspective, which is I think um it is an amazing capability for expanding the sensory cortex, adding the pros- prostheses to the mind. In other words, restoring function once broken, expanding function, like let's say seeing more in more wavelengths or hearing better than we could hear, not just repairing hearing, fixing spinal cords, basically working from the periphery of these systems. As opposed to a much more difficult and yet to be solved task, which is uh upgrading core functionality, like just making someone smarter. So, I think the example you gave is is a very interesting one. Could you have a higher data rate communication? Absolutely. I think that is very doable. And the reason I have this belief, it's more of a um pattern recognition across decades of complex system development. Basically, the high-level statement would be any product produced from an iterative algorithm, which would be evolution, genetic programming, all neural networks, um you know, cellular automata, whatever it might be. Um if you iterate something billions of times and accumulate complexity from that algorithm, the thing you make is inherently inscrutable. It is an artifact of absolute like inscrutable complexity. Despite attempts at mechanistic interpretability in AI, I don't think that's going to bear fruit. I don't think control and uh and alignment is possible in a cutting-edge system that is pushing and back to AI for a moment, pushing the capabilities of of what we can build. Similarly, it'd be like asking about controlling, aligning, um mind-controlling a teenager. So, I swap teenager and AI whenever I think about this. So, when it comes The reason that's relevant is from the brain as a complex system itself. And reverse engineering its inner workings for uploading or for, you know, brain-to-brain or just like adding speech like the way Jeff Hawkins thinks he just like cut and paste this like a French speaking module into a human brain or or a neural net. I don't think it's going to be possible on a time frame of relevance meaning it'd be easier to build a new intelligence than it is to reverse engineer one you've made. So I do think Neuralink is fascinating but I don't personally um get faith that it's going to keep up with AI. Maybe that'll be the safest way to phrase it. Not that it can't be done but the time scales, you know, FDA cycles, human biology. Nothing happens on a time scale comparable to the rate the learning loops. Back to the Elon Musk saying focus on learning where do you learn more quickly? You can learn more quickly in the synthetic domain. I think humanity always wants to believe it's part of the future in that regard but they the Kurzweil's uploading I can just see why he wants that to be true within his lifetime and that's what he predicts will happen. But it doesn't mean it will.

Marina: Steve, I'm really curious. What do you think about Penrose's argument that the consciousness is go far beyond algorithmic processes to quantum level processes meaning that AI would never be able to to develop consciousness itself just by by its nature. So what do you think? Can AI develop consciousness or it's will be only imitated and that's it?

Steve: And you're referencing Penrose's

Marina: Yes.

Steve: Yeah, so Penrose is a brilliant guy in UK um generally but here he has this gut feeling that there's some quantum process in the brain that makes it unique and yet there's no real clear mechanism by which that would happen. There's some argument around some lithium isotopes that might be a coupling but but it's wishful thinking. We don't so to speak. Um but I can also generalize your question. Is there something a vitalistic naturalistic unique to our brain that is irreproducible in others and and there was a reference earlier to Neil Seth's work. I find the arguments completely uncompelling that there's something vitalistic or unique to the substrate. Just because it's the only example we know of of consciousness and consciousness for example is a tricky thing like do how do we know if the dog is conscious? What how do we how do we test for this, right? But we believe we see it in ourselves. I mean I don't know if you're conscious but I'm kind of just guessing you are, right? Um and I mean you seem awake and you're human therefore we generalize it conscious. Okay, so I've not seen a compelling argument just because we have an example of one doesn't mean it's the only possible example. It's you could make a similar argument that says does all life need to be carbon based, right? And there is something unique about carbon and it's be able to do single, double, and triple bonds and and all the weak bonds. It is kind of you could actually make I think a better argument that says carbon is special to life than you could to say neurons as we have them or essential to consciousness. Now totally different question is but I won't I won't digress is like is anything that we're doing in AI development going to lead to conscious? That's a different question because you could you could argue that's a dead end and won't get us to conscious but it doesn't mean it's not possible. It's much higher order proposition to say something is impossible than to say I don't know. And so my answer would be I don't know but I certainly wouldn't say it's impossible and I don't believe that we have any evidence of a quantum process going on in the brain and if and if we did why couldn't we replicate that with quantum computers? I mean that's a different question. And then if I brought in your question farther uh just animus or spirit or life is does it have to be a living thing to be conscious? And the analogy I would use is imagine you substitute the word memory for consciousness. I and I just picked memory reason just randomly. It's an overloaded term. Do we mean memory like I have memories in a human sense, human memories which are holographic and and they can have graceful degradation and they're not at all the way we do memories in a computer chip but when we talk about computers they have memories too and we don't debate is memory possible in a computer? Can it remember things? Well, in that at that level of abstraction, of course they can. And yet it doesn't have human memory. And that's fine. Um so consciousness, it may not have human consciousness, but maybe it has a different kind of consciousness, whatever that thing is. If we could be more precise about defining it. And I don't think you make the argument that everything we have in our brain is essential for consciousness.

Steve Jurvetson: In other words, there's a lot of there is a garbage collection for our metabolism, you know, that you know, things that happen when we sleep and cleaning up but you know, waste products and the way mitochondria work. You don't you don't have to have all of that in a computer to be intelligent or to have memories. What you don't need all that baggage for consciousness either. But that doesn't mean we know what the minimum set is. But it does I think we'll figure it out one day. So in other words, I'm more on the my gut tells me, oh sure, I think one day they will be conscious. I don't know if we're on a path to get us there. Maybe something more akin to evolution and and reinforcement learning algorithms would get us there more more obviously just because whenever you're recapitulate what we've already done with our biology, that makes me give hope that won't Why can't we do it in a different substrate?

Marina: Thank you so much.

Steve Jurvetson: Woo!

[cheering]