This Week in Startups
00:25:46 · Oliver Cameron
[00:25:46] S2: To speak a little bit about Odyssey and what we do. And again, thank you, for having me on. So we build world models, and we believe that world models are a true foundation model. What we do is we train them on vast visual observations of the world, and then they gain this really cool ability to predict the possible futures of the world. And so through this training process, what they do is they learn physics, they learn dynamics, cause and effect, all of these fundamentals of the universe. And so what is this useful for? Well, we believe that WARP models are essential to enabling a billion robots and a trillion agents. For robots, WARP models are this general brain, and it allows robots to better predict the consequences of their actions, and to thus better interact with the world. And then for agents, world models generate infinite environments for agents to learn within, and this increases their performance on tasks that they would learn inside those environments. And so what we did here is say, hey, there is this really cool, you know, fruit fly project. Someone's mapped the brain of an... You know, a species. That's incredible. Let's now put that brain inside an environment that's being generated in real time. And will the fly decide to fly? Will it just sit there? What will it do? And as you can see here, the fly flew around as a fly does. And I I think the reason it went viral, you know, is that it's this perfect perfect moment of saying, are we living in a simulation? Right? We we have increasingly capable artificial intelligence. This dream of superintelligence is no longer a dream. And I think it's on lots of people's minds now that with increasingly advanced technologies, whether all of this is in fact a simulation. And so I I think it just sparked that thought in people. It's like, wow. Okay. Now now we're starting to see real visual evidence that this might be a simulation.
00:27:56 · Jason Calacanis, Lon Harris
[00:27:56] S0: And for people who don't know, Lon, there is a concept called simulation theory. Yes. You as my very astute follower of science fiction can probably do an exceptional job explaining what simulation theory is and why this, you know, is a milestone in some ways, a waypoint [00:28:19] S1: Sure. [00:28:19] S0: In simulation theory from your perspective.
00:28:19 · Jason Calacanis, Lon Harris
[00:28:19] S0: In simulation theory from your perspective. [00:28:21] S1: I mean, thankfully, we have a great pop culture reference point that I think we can all sort of agree on, which would be the, 1999 classic, The Matrix, which sort of proposes, I think, the general idea. Get away from human beings or the batteries or whatever. But the idea that we could all be living in or our consciousness could be living in some sort of a virtual reality simulation of the world rather than the real world, and that we don't have any objective way to know because everything we sense about the world is through our senses. And so if we could suspend a fly brain in a computer simulation like this and it believes it's a fly, it's flying around, it's flying in a virtual world, and it has every reason to think that that's a real world. What if we had much more sophisticated computers, much faster GPUs, much more, you know, like, elaborate models? Could any one of our brains be in this kind of a computer simulation as well walking around a virtual world, doing a virtual podcast, believing that it's real Yes. But not really is is, the idea. There is a great documentary called Glitch in the Matrix that plays on this idea and that speaks with people who believe we are, in fact, in a simulation that I that I really recommend if you wanna learn. [00:29:37] S0: What percentage? We we have p doom a lot. Oliver, what percentage are you? Somebody building world models and putting essentially, this is a simulated fruit fly. Yeah? This is a... How do you simulate the fruit fly's existence then?
00:29:37 · Jason Calacanis, Oliver Cameron
[00:29:37] S0: What percentage? We we have p doom a lot. Oliver, what percentage are you? Somebody building world models and putting essentially, this is a simulated fruit fly. Yeah? This is a... How do you simulate the fruit fly's existence then? [00:29:55] S2: There was this open source project, that takes this mapped brain and then connects it to actions, that it can perform. So I wouldn't necessarily take from this that this is the brain of a fruit fly as a fruit fly operates in our physical world. It's a sort of, I don't know, baby version of that. Sure. But, yeah, on its way. [00:30:17] S0: Got it. And and at some point, are you planning to put... Is there a way to connect a fruit fly's brain or some a mouse's brain, a rabbit's brain? Obviously, there's Neuralink and other things. But in three, four, five, six years, what would you be simulating here in the world model? Could you actually connect a biological creature to a world model and just shoot it into their brains?
00:30:17 · Jason Calacanis, Oliver Cameron
[00:30:17] S0: Got it. And and at some point, are you planning to put... Is there a way to connect a fruit fly's brain or some a mouse's brain, a rabbit's brain? Obviously, there's Neuralink and other things. But in three, four, five, six years, what would you be simulating here in the world model? Could you actually connect a biological creature to a world model and just shoot it into their brains? [00:30:40] S2: I mean, our focus is primarily on these as environments for agents to learn in. And so, really, one thing that strikes me as really interesting to pursue is this idea that we humans exist inside this world model and all of these systems that we are raised in shape us, shape our intelligence, shape what we do and how we do it. And language models as they're currently trained, I don't feel are faithful to that idea. They sort of short circuit a few of these and have, versions that I'm not quite convinced by. And so what I think is really exciting is this idea that you could place agents, both mature or nascent agents, inside these environments that to the agent are utterly real and actually place it with other agents too and say, hey. Learn the world. Learn how the world works. Try to accomplish tasks. So
00:30:40 · Oliver Cameron, Jason Calacanis, Lon Harris
[00:30:40] S2: I mean, our focus is primarily on these as environments for agents to learn in. And so, really, one thing that strikes me as really interesting to pursue is this idea that we humans exist inside this world model and all of these systems that we are raised in shape us, shape our intelligence, shape what we do and how we do it. And language models as they're currently trained, I don't feel are faithful to that idea. They sort of short circuit a few of these and have, versions that I'm not quite convinced by. And so what I think is really exciting is this idea that you could place agents, both mature or nascent agents, inside these environments that to the agent are utterly real and actually place it with other agents too and say, hey. Learn the world. Learn how the world works. Try to accomplish tasks. So [00:31:39] S0: Percentage you're convinced that we're in a simulation right now. If you're being honest, like, and you've had a range of this from, you know, over time. And let's just assume 0% of us before seeing The Matrix or having read some of the philosophers who the books were based on. If everybody starts at zero, where are you at now in terms of this reality we're in? Is the base reality or we're in a simulation. Base reality would be we're the ones running the simulation versus [00:32:10] S1: Core core reality. The matter you