Odyssey CEO Oliver Cameron is 90% sure we live in a simulation | E2358

This Week In Startups is made possible by: Paper: https://getpaper.ai/twist G2i: https://G2i.co/twist Northwest Registered Agent: https://www.northwestregisteredagent.com/twistdomain Plaud: https://Plaud.ai/twist Today’s show: Odyssey co-founder and CEO Oliver Cameron builds simulated worlds for a living, so we asked him the obvious question: are we all living in one, right now? He says he’s 90% certain that, yes, we are. To help make the case (and show off their world model technology), Oliver’s team dropped a simulated fruit fly brain into an AI world, that’s being rendered and created in real time. So far, on X, 6.2 million people have watched the virtual fly navigate an imaginary world. On TWiST, we’re digging into how world models learn about physics, gravity, and causality just by watching videos, what make Odyssey-3 an ideal trainig ground for AI agents, and why a world model only renders the parts of the world that are currently being experienced. It’s a peek inside the technology that may, one day, allow us to enslave people in The Matrix and turn them into batteries. Or train self-driving cars more effectively! One of those. Plus: SpaceX throttles Grok Bot over unsustainably high costs, but Jason says it’s worth every penny. And will OpenAI’s suggested name change from data centers to “compute campuses,” and Meta’s efforts to improve El Paso, help turn around American attitudes? Or is blasting them into space the only remaining viable option? Guest Oliver Cameron on X: https://x.com/olivercameron Odyssey: https://odyssey.systems/ Relevant Links The fruit fly brain flying inside Odyssey-3 (6.2M views): https://x.com/odysseyml/status/2111589148146749616 Meet Odyssey-3, Odyssey's foundation world model: https://odyssey.systems/meet-odyssey-3 FlyWire: https://flywire.ai/ Nick Bostrom, "Are You Living in a Computer Simulation?”: https://simulation-argument.com/simulation.pdf Do we live in a simulation? Chances are about 50-50 (Scientific American): https://www.scientificamerican.com/article/do-we-live-in-a-simulation-chances-are-about-50-50/ "Beyond Perception: Proposing Our Reality as an ASI Alignment Simulation" (Ali Eslami, IPI Letters): https://ipipublishing.org/index.php/ipil/article/view/350 The double-slit experiment explained (Royal Institution, Jim Al-Khalili): https://www.youtube.com/watch?v=A9tKncAdlHQ “A Glitch in the Matrix” trailer: https://www.youtube.com/watch?v=au8eT79WUJ0 Elon Musk throttles Grok Bot limits over "unsustainably high" costs (TechCrunch): https://techcrunch.com/2026/11/02/elon-musk-throttles-grok-bot-limits-due-to-unsustainably-high-costs/ Elon's post: "The cost of serving Grok Bot was unsustainably high relative to revenue”: https://x.com/elonmusk/status/2116992258520064397 Jason's post: "< $40k a year on tokens… getting $2m+ in value": https://x.com/Jason/status/2117126320962351277 Grok Bot: https://x.ai/bot Plaud Web: https://web.plaud.ai Founder University Japan: https://japan.founder.university/ Most Americans say AI is bad for society and oppose more data centers (ABC/WaPo/Ipsos): https://abc7ny.com/story/most-americans-say-ai-is-bad-society-oppose-more-data-centers-abcpostipsos-poll/19964027/ Chris Lehane's memo: AI will be the issue in 2028 (TNW): https://thenextweb.com/news/chris-lehane-memo-openai-2028-ai-election Meta touts El Paso water and grid pledges as council weighs incentives (KFOX): https://kfoxtv.com/news/local/meta-touts-10b-el-paso-data-center-water-and-grid-pledges-as-council-weighs-incentives UT Austin's Texas Advance Commitment (free tuition under $100K): https://admissions.utexas.edu/cost-aid/financial-aid/texas-advance-commitment/ Andrej Karpathy's HackerTube post: https://x.com/karpathy/status/2116957135661945283 Jason's "Unsponsored" Chrome extension for Amazon: https://x.com/Jason/status/2116380861167714594 Google's first data center satellite is online and writing poetry (TechCrunch): https://techcrunch.com/2026/10/30/googles-first-data-center-satellite-is-online-and-writing-poetry/ What is low Earth orbit? (Universe Today): https://www.universetoday.com/articles/what-is-low-earth-orbit Timestamps: 0:00 Grok Bot users are at their limit 3:36 How Jason and LAUNCH get $2M+ in value out of AI 5:46 Jason’s Grok-made jacket lookbook 10:35 Paper is the legal backbone for thousands of startup founders and fund managers - including Jason. Use https://getpaper.ai/twist to get your cap table free for a year - or two years if you're leaving Carta. 14:12 Everything changed in AI over the last 90 days 20:41 G2i has been recruiting software engineers for over a decade. Go to https://g2i.co/twist to find out how G2i vets engineers, and can help you hire them in days, not weeks. 22:06 Plaud: If your work depends on conversations - interviews, meetings, calls - you need a Plaud NotePin S. You can check it out at https://plaud.ai/twist and use code TWIST for 10% off! 22:37 Wiring our Plaud into our Grok Bot 23:24 Odyssey’s fruit fly video zooms to 6.2M views 27:32 All About Simulation Theory 34:34 Got a new business idea? Northwest Registered Agent helps you bring it to life. Get a free domain, email, phone number, and more - with no purchase required! Learn more at https://www.northwestregisteredagent.com/twistdomain 38:33 What is the double slit experiment? 40:47 How to teach a robot about physics 50:01 OpenAI’s plan to make Americans love data centers 56:53 Meta’s El Paso data center pitch 1:02:26 AI companies should overfund local schools 1:06:27 Karpathy’s HackerTube project 1:13:02 Google’s first space data center goes online 1:17:51 How much room is there in low Earth orbit? Subscribe to our newsletters on Substack: TWiST: https://twistartups.substack.com/ TWiAI: https://www.thisweekinai.ai/ TWiVC: https://thisweekinvc.substack.com/ Subscribe to This Week in Startups on Apple: https://rb.gy/v19fcp Follow Lon: X: https://x.com/lons Follow Jason: X: https://twitter.com/Jason LinkedIn: https://www.linkedin.com/in/jasoncalacanis Thank you to our partners: Paper is the legal backbone for thousands of startup founders and fund managers - including Jason. Use https://getpaper.ai/twist to get your cap table free for a year - or two years if you're leaving Carta. G2i has been recruiting software engineers for over a decade. Go to G2i.co/twist to find out how G2i vets engineers, and can help you hire them in days, not weeks. Got a new business idea? Northwest Registered Agent helps you bring it to life. Get a free domain, email, phone number, and more - with no purchase required! Learn more at https://www.northwestregisteredagent.com/twistdomain Plaud: If your work depends on conversations — interviews, meetings, calls — you need a Plaud NotePin S. You can check it out at https://Plaud.ai/twist and use code TWIST for 10% off! Check out all our partner offers: https://partners.launch.co/ Great TWIST interviews: Will Guidara, Eoghan McCabe, Steve Huffman, Brian Chesky, Bob Moesta, Aaron Levie, Sophia Amoruso, Reid Hoffman, Frank Slootman, Billy McFarland Follow TWiST: Twitter: https://twitter.com/TWiStartups YouTube: https://www.youtube.com/thisweekin Instagram: https://www.instagram.com/thisweekinstartups TikTok: https://www.tiktok.com/@thisweekinstartups Substack: https://twistartups.substack.com

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Jason Calacanis?

  All right, it's twist for Monday, November 2nd, 2026. There's snow in Tahoe. I'm going to Japan and Davos in January. Things are about to heat up.

• 0:15Speaker 01

  Cool down.

• 0:15Jason Calacanis?

  And the most exciting thing is Grokbot. Grokbot is the most exciting thing right now. What's going on with Grokbot this weekend? Because all I did all weekend when I wasn't with the family and going to a Halloween party was just hang out with my Grokbot. I'm so addicted.

• 0:31Speaker 01

  Yeah, it's a very compelling piece of software. So Grokbot users spent the weekend complaining that they were burning through their weekly usage limits in just a few days, Jason. Elon Musk responded on X, "The cost of serving Grokbot was unsustainably high relative to revenue. He did not give any specifics on cost or whether the limits were cut or by how much. This comes two weeks after SpaceX's earnings call, where Elon called Grokbot the fastest growing product he'd ever seen. SpaceX President Gwynne Shotwell said it hit 800,000 paid weekly users, more than doubling in a month. Grokbot starts with a $30 a month premium plan. The $100 and $200 plans get higher limits."

• 1:12Speaker 01

  And you posted last night that Launch currently spends under $40K a year on AI tools, including Grokbot, and gets, in your opinion, $2 million plus in value. So is this just a people need to suck it up and spend a little bit more to keep in touch with their Grokbot?

• 1:29Jason Calacanis?

  I think this product is so addictive, obviously, and you get so much value out of it that the usage limit you're going to quickly get to. And it doesn't kind of keep you informed of your limit. You have to go kind of hunt and peck to find that. I think what they need to do is put a bar on the top of the apps. And every time you're using it, it should just show that bar and give you just a little bit of an idea of what you're consuming. Because I have Ultra.

• 2:07Speaker 01

  Ultra. I'm on premium still.

• 2:09Jason Calacanis?

  Which is 200 a month?

• 2:11Speaker 01

  No, well, I, I signed up for premium, which is only 30 a month, and that's what I was using. But then the team, you know, we started adding other members of our team. So then it bumped me up. So I get the 200 a month plan just for myself. And then every other team member gets their $30.

• 2:28Jason Calacanis?

  Oh, by doing that as the team manager, they give it to you for free?

• 2:32Speaker 01

  I mean, I don't know. It told me that my premium usage, I was getting an increase because I'm the team manager or whatever. And so I don't, I don't run out as much or as quickly. I've only used 9% with only four days to go this week, and I have not adjusted my Grokbot usage, so.

• 2:50Jason Calacanis?

  So that is one issue is, and Perplexity does a great job of this. I've stopped using Perplexity as much because all the stuff I was building with Perplexity Computer, I'm now using a bot interface to do.

• 3:03Speaker 01

  Right.

• 3:04Jason Calacanis?

  And the bot interface is more pleasing and more efficient. And now that I have 100 different bots, I'm starting to put bots together. So I have a bot.

• 3:14Speaker 01

  This is crucial.

• 3:15Jason Calacanis?

  That is good for making videos. I have a bot that's good for doing research. I have a LinkedIn bot. I've got an HTML webpage creating bot. And for, and over this weekend, I, for the first time, had it, I gave it my credit card, had it register a domain name for me, and had it publish to the web for me some content. We were going to do, we're going to buy jackets, right? So I just want to give people an example here. We're going to buy jackets, right? And you can pull up my, a tweet from this weekend.

• 3:45Speaker 01

  Uh, right. So that, yeah, this was, this was Jason's response that we sort of quoted before. He's saying, "I'm spending about 40K a year on tokens and subscriptions from Anthropic, Lovable, Grokbot, Eleven Labs, long tail of other products, and getting about $2 million plus in value for Launch and this weekend startups. We've automated at least seven full-time jobs at this point while becoming 100% faster at the average task that we do. If you haven't used the last 90 days worth of products for at least 30 hours a week, you have no idea how profound the software is. The last 90 days, things have changed more than in the last two years. Many of the things I spent hours on last year now take minutes, and the results are mind blowing." You know, I really agree.

• 4:24Speaker 01

  I think a lot of people got fixed in time in like 2023, 2024, like this is what chatbots are, and this is how AI works. And they're still relying on those tropes of like, oh, it hallucinates and you can't trust it, and it doesn't have any context, and it forgets things. And it's like, Grokbot's a new world, man. You gotta, you gotta plug in.

• 4:46Jason Calacanis?

  Correct.

• 4:46Speaker 01

  Because like I, teaming them up is where it gets, I think, really powerful as well. Like I have my Substack analytics bot that is looking every day at what's happening on Substack and posting in Slack. But now I have that bot talk to all the other bots that helped me research those Substacks. So now they learn what works and what doesn't, and refine their recommendations based on the analytics and the readership we're getting in Substack. And like, that's how they, it can make smart decisions about what to write about without me, because it's already like, the last time we did a post like this with this kind of title that we could do again, it did really well. The open rates were 6% higher, like, whoa, man. That's like a full-time analyst looking at our stats that's just working for me now.

• 5:31Speaker 01

  That's pretty incredible.

• 5:32Jason Calacanis?

  Yeah. So what I did was, we were going to make a jacket, right? We wanted to make a jacket. And so I had.

• 5:40Speaker 01

  Like a cool jacket.

• 5:43Jason Calacanis?

  Yeah, well, I wanted to make like a cool jacket, right? Everybody wants to have a cool jacket. So what I did was, I decided, you know, people were like debating like what we're going to get for a jacket for Founder University in Japan and for Launch Accelerator. And, you know, I kept explaining myself to a human being. And the human being did human being things, which is make mistakes, ignore certain key factors, be fallible, just like I am, just like you are, we all are. And the Grokbot made this for me, posted it to my own URL. And I did this in about under 15 minutes of instructions, maybe 10, of just talking

• 6:28Jason Calacanis?

  to it, using Whisperflow to guide it. And then I just kept republishing it, republishing it, republishing it. I got so far here, this is something you would have paid a designer. If you scroll up here, you'll see.

• 6:41Jason Calacanis?

  And if you just scroll up, I'll tell you when to stop. So here's like, and by the way, this was no prompting. It just made me a table of contents. I say, give it, like, give me the best jackets, the most hip jackets, et cetera. And then I said, here's where I want to place the logos. And I gave it the two logos I wanted to place. So I like when you make schwag to have it not have a lot of outside branding, right? I like it to be subtle. So here, I said, where to place logos, scroll up. I said, you know, on the inside of a denim jacket, if you put these logos, that's one of our companies, and then Founder University, or in a leather jacket, scroll up. And it just made this. So now me explaining to somebody what I wanted, a human I'm talking about, would have just been back and forth, back and forth, hours, meetings.

• 7:23Jason Calacanis?

  And then I said, you know, I want to put something on the collar or where the tag is, keep scrolling up. And it just made me the perfect mockups of this. Now, keep going.

• 7:33Speaker 01

  It looks beautiful.

• 7:34Jason Calacanis?

  Not only did it do that.

• 7:36Speaker 01

  Designed, you know, like it looks like a human spent time and care in.

• 7:40Jason Calacanis?

  Like a professional designer.

• 7:41Speaker 01

  Exactly.

• 7:41Jason Calacanis?

  Yeah. And so I said, oh, I want to put Do the Work, Kaizen, Ikigai on the inside to kind of just reinforce with our founders, like some of the philosophies, maybe make it a little inspiring scroll up.

• 7:53Jason Calacanis?

  Here we go, more examples, keep scrolling up.

• 7:55Speaker 01

  We learned, interesting cultural thing that we learned, because originally when we were designing Founder University Japan stuff, we wanted to have Japanese characters and writing, because in America, we all think that's cool. But in Japan, they think it's really nerdy to have Japanese writing on your clothes, but they think it's cool to have English. So that's why I think we're now leaning towards writing the words in English.

• 8:17Jason Calacanis?

  Correct. And then there's the Ikigai logo, or the, you know, the Venn diagram, keep scrolling up.

• 8:24Jason Calacanis?

  And then I think it gave us like, oh, and then I said, go find out what other people have done with these same jackets and find me how they've done the logo branding.

• 8:33Speaker 01

  Right.

• 8:34Jason Calacanis?

  So it found like a work shirt and, you know, famous chef, Eric, there with it on, and different patch ideas I didn't have, keep scrolling up.

• 8:44Jason Calacanis?

  And keep going. These are leather jackets, like a biker jacket. Here's denim jackets and how people put logos on them and do other branding, keep going. And it sourced all this and linked to it. And then I love these work jackets, which is what chefs wear or like the French wear when they're working in their garden or whatever.

• 9:01Speaker 01

  Right.

• 9:01Jason Calacanis?

  And man, it found me all this information, keep going. This would have been two weeks of work for a project manager working with a designer, right? This is like 100 or 200 hours of work. And here we go, it just did, it keeps scrolling up.

• 9:16Speaker 01

  The research alone. I mean, not even designing this and writing all the content.

• 9:20Jason Calacanis?

  And then if you look at this paragraph above, Lon, check this out. Tell people what this is. Keep going, up a little bit. Keep going, yeah, right there. Perfect. What this means for our jackets.

• 9:31Speaker 01

  Right. Oh, wow. One placement zone per jacket. Patches look sharp when they are, sit together at the same size, like the chest grid on the Vansen Race jacket, or the stacked upper lip sleeve on a patched Perfecto. Scatter them and it turns into a costume. That's a great insight. That's true, actually.

• 9:50Jason Calacanis?

  Literally learned the best practices, right?

• 9:52Speaker 01

  Yeah.

• 9:52Jason Calacanis?

  And gave examples and everything. Now, just scroll down quickly here till you get to the work jacket. So here's all 25 types of jackets it found.

• 10:02Jason Calacanis?

  Here's where this workwear fits.

• 10:04Speaker 01

  Not the varsity jacket is nerdy unless you are a varsity.

• 10:07Jason Calacanis?

  True nerdy. This is the chore jacket. So this is what I'm gonna make for everybody. I decided.

• 10:11Speaker 01

  I like that.

• 10:12Jason Calacanis?

  Which is classy. You wear it, or you can wear it to dinner. You can wear it on a hike. It literally is very versatile, because you wear stuff underneath it and you have this, like, oversized work denim thick kind of cotton jacket. Beautiful.

• 10:24Speaker 01

  It's very Texas, like a ranch jacket or a work jacket that you can also wear to dinner. That's like, it's a very, it's a very central Texas look.

• 10:31Jason Calacanis?

  And very Parisian and very Japanese. Thank you. When you're building a startup, your legal paperwork is spread out all over the place. Entity formation here, fundraising over there, cap tables, contracts, it's endless. But Paper finally has a solution that's gonna save you so much time and so many headaches. Paper's introducing smart entities. Here's how it works. You tell Paper what you need to do and their pre-built agent takes care of all of it. That's it. This is not some chatbot you need to provide with a ton of context. Nope. Paper is an AI-native platform that's going to manage all of your startup's legal work. That means no more switching between disconnected tools, keeping track of hundreds of documents, and way less busy work for you to stay on top of.

• 11:17Jason Calacanis?

  That's why Paper is already the legal backbone for thousands of startups and funds. Right now, founders using my link, get paper.ai/twist, get their cap table free for one year. What an amazing deal. Or two years if you're switching from Carta. If you're building a venture-backed company, it's time to build your smart entity on Paper. You know, I would have needed to put somebody on this or hire somebody, so I would have spent 100 hours on this. Instead, I spent 15 minutes and I got a better product. So saving $100 an hour, and, you know, developers get paid $75 an hour. If you take their salary, 75 times 2,000 hours a year is 150 grand.

• 11:58Jason Calacanis?

  You know, a clerical worker, you know, an Athena assistant might be 10 an hour. A clerical worker in the United States might be 20, 25. At a school, it might be $30,000, $30 an hour, whatever it is. College-educated kid, 40 bucks an hour, something like that. You could actually put a price on these things. So I started putting a price on it and I realized, most of the companies on $20, $30 a month subscriptions, some of us are on 200. And we're probably only spending 40,000, but I looked at the amount of software we're creating. We would need five or six developers and two full-time designers and two project managers. It's nine full-time positions that would have averaged 100K each or more if it was developers, but I'll just put it at 100. That's like a million dollars right there.

• 12:43Speaker 01

  Yeah.

• 12:43Jason Calacanis?

  Then you look at what happened to producing podcasts, making clips. It's a similar thing. We would need three more people right now to do what we're doing? Two, maybe? Three?

• 12:51Speaker 01

  At least two, if not three. I mean, even just like, if you're thinking about the content alone, the amount of writing, the amount of research, like, not even getting into the editing and the post-production and everything, like,

• 13:01Jason Calacanis?

  Right.

• 13:02Speaker 01

  Yeah, we'd need to be a two- or three-person team larger.

• 13:05Jason Calacanis?

  I'm gonna say three. All right, so now we're at 12, 13, 14 people in the company, you know, and there are things we're doing now that we just couldn't have done. There's software we're making that is, you know, growing the business and you put all that together. I think we're getting about $2 million in value and 40,000 in spend. So now you kind of do some math there. You're talking about 50 times the value.

• 13:29Speaker 01

  Yeah.

• 13:30Jason Calacanis?

  So for every dollar we spend, we're getting $50 out. So I want to just let that sink in. A dollar invested in AI today, if you're doing it properly, is $50 of output. Now, can you monetize that? Is it profitable? Or are you just spinning your wheels? Well, that's up to you to define the goal, right, Lon? And what we've done now is we've defined the goal. We're hiring like, for 10 positions here at the company. And you say, wait a second, why aren't you just automating them? Well, things are growing so fast, there are things that are uniquely human, like being on air. Video editing, still uniquely human, having taste, running a community.

• 14:08Speaker 01

  Right.

• 14:09Jason Calacanis?

  Running events. You still, you know, that's in the real world. You still need somebody to go to the venue and greet people and all that stuff. So what we're hiring for are uniquely human people, but I have to say, this technology, since the bots came out, and in the last 90 days, I believe the last 90 days are the equivalent of the first two years of using this software. That much has changed, because we were trying to do this stuff with Open Claude. Then I tried to do it with Perplexly Computer. And I tried to do it with Claude Co-Work and projects. We tried to do it originally with ChatGPT. I mean, we've been down this road, you know, Slackbot, everything.

• 14:41Speaker 01

  It's the, I mean, people say it's the agents, and I think that that sometimes gets construed as agents are inherently more powerful or whatever. It's that the agents have context and access. That's, that's what changed. Like, Open Claude, the promise was, you're gonna have this agent that can act independently and is smart and can do things. And it's true. I don't mean to say it's not true, but it was very difficult, especially if it was in a virtual server somewhere, for it to get into Slack, get into Notion, get into email, get into your calendar, get into your messages.

• 15:10Jason Calacanis?

  So why did that happen? Well, one, all the services out there built connectivity, right? The second thing that happened in all this is that they put a virtual computer inside of Muse and inside of Grokbot. I think Claude might have it. Claude hasn't released their agent technology yet.

• 15:28Speaker 01

  Browser that you can hook it into your Google Chrome and then it can look at everything in your browser. It's simple.

• 15:33Jason Calacanis?

  No, but they're adding the virtual machine that the agent can use. Once you do that, and I've gotten comfortable with it, inside of Grok, I don't know if you've seen this, there's a vault. Have you started using the vault yet?

• 15:43Speaker 01

  Absolutely. 100%.

• 15:44Jason Calacanis?

  Okay, so there's a vault. So, you know, booking flights is also another one of these things that is a back and forth, back and forth. You're giving it to a human, etc. It's just hard, scheduling, another one of these. So what I'm doing now is, I'm having my Athena assistants use Grokbot. I got them, I have two Athena assistants. They're gonna use the same instance. And then I'm having them do every task they do. They have to build a Grokbot for it.

• 16:14Speaker 01

  Right.

• 16:15Jason Calacanis?

  So now I sucked in all the Notion pages of my travel preferences, my, you know, miles, numbers, whatever, and I just started giving it goals. Now they're using that too. Boom. This stuff is all just getting done and it actually finishes the task. And that was the problem previously. It didn't finish the task. People didn't know how to give it goals. It wasn't connected. There wasn't a computer.

• 16:35Speaker 01

  It would timeout and get confused. And then you'd be like, what happened to that? It'd be like, oh, you're right. I didn't run that. I'm sorry.

• 16:42Jason Calacanis?

  It was brittle, to use a Chamath term that he likes to use. There was a brittleness to it. And now it's not brittle. It's the opposite. And so I, my advice for everybody with the Grokbot is, you're gonna need to do the assessment like I did. If Grokbot was $1,000 a month, and, you know, I think we've got four or five people on it. Don't clip this and send it Elon, but if it was $1,000 a month, of the four people I have using it, I probably would have three of the four keep using it for $1,000 a month. Because even then, if my 40,000 a year bill went to 200,000, it would be 10 to 1. So every dollar I'm spending, a 10. So I think what's gonna happen is, this is kind of like Uber and Lyft when they were discounting rides

• 17:28Jason Calacanis?

  to get people to use the software. They're gonna start boiling the frog. They're gonna start raising the prices and take the losses out. Elon's just the type of person who will tell you that's what he's doing. You know, Claude, and I haven't heard Dario just come out and be like, by the way, we're losing money on this. It's not sustainable. It's growing too fast. We need to change the price a little bit. We gotta change the deal. Elon's just upfront about it. Anybody who's complaining, look at what you're getting out of it. Please go do some simple math. Are you getting 50, 30, 20 for every dollar you spend? Just increase your spend. Just go to the ultra plan. Go to the $100 plan. And you really want this to, I do think some people felt rug pulled

• 18:09Jason Calacanis?

  because their browsers, their bots stopped working. They might have been dependent on that.

• 18:14Speaker 01

  Right.

• 18:15Jason Calacanis?

  This is why you need to give everybody the status bar that shows usage.

• 18:20Speaker 01

  I agree.

• 18:20Jason Calacanis?

  As you go. Countdown clock.

• 18:22Speaker 01

  I figured it out because I got so frustrated. We were doing, I had Grokbot doing like a long form research project. And so I was running out of compute even before all of these changes, just because it was combing YouTube and doing so much research for me. So I discovered you go to cursor.com/dashboard/spending. That's where you could actually keep a running eye on how much usage your Grokbot has left before restarting the next week. Uh, I usually mostly use it up by the end of the week, but I'm going to keep an eye on it now. But anyway, that's where to go. But I agree, having it in Grokbot itself, so you could just monitor it more closely would be very helpful. I think people don't realize that if you do give it those long range assignments, it can suck up compute pretty quickly.

• 19:05Jason Calacanis?

  I am on a Ultra Plan and it, uh, yeah, I, I don't know how I got my Ultra Plan. It might be because I'm part of the all in organization and I have that icon next to my name. And, uh, so I, somebody's paying for my Ultra. I don't know what Ultra costs. I think that might be two or three or four.

• 19:26Speaker 01

  I think that's the two.

• 19:27Jason Calacanis?

  Two hundred, okay. I have Ultra, I'm at 51% weekly usage and it resets on November 3rd. So they gotta just let everybody know what they're spending and don't make you search for it. This.

• 19:40Speaker 01

  Yeah, cursor.com/dashboard/spending, folks. That's where to go.

• 19:44Jason Calacanis?

  And somewhere in there is like the bar chart I just shared that you could show. This needs to be on the top level. It should literally be part of the interface, just showing it to you. And that every single one of your jobs should give you a little piece of feedback at the end, when it finishes its job, 0.7% of your weekly total, this many tokens. Just educate us in real time. And you could turn it off, I guess, but I think the reason people didn't have this on is they didn't want to give people compute anxiety. But if we're in this dance, right? With, uh, the providers of this technology and we want to use it, I think we just have to be transparent about it, like, which, you know, for a business, you kind of want to let somebody run out and then have to upgrade, like video games or whatever.

• 20:28Jason Calacanis?

  I think in this case, you gotta switch it since it's a business case and just say, you really, based on your work, you're at the $20 level. We highly recommend you go to this $100 level. It's going to be this much more per year and you're worth it.

• 20:41Speaker 01

  You know, it's even harder than writing great code. Consistently recruiting people who can write good code. And if you're working on a project and you need to hire great engineers, but you don't want to go through a painful, time-consuming hiring process, G2I can help. G2I has a network of more than 8,000 pre-vetted software engineers, the crème de la crème, all with more than five years of coding experience. These are the elite developers you need. And these aren't people who hack some automated coding test. The team at G2I puts engineers through live technical interviews to ensure these are folks who really know how to get their hands dirty building stuff. And the best part for founders, they make the entire process fast and easy.

• 21:23Speaker 01

  You get a dedicated Slack channel where you communicate directly with the G2I team so that you can share exactly what you need and start getting matched with engineers today. So go right now to G2I.co/twist and get $1,500 off your first hire. That's G2I.co/twist. I think it, it stung a little bit, maybe extra because it, I, I think a lot of people don't think of it as an enterprise application. It absolutely is. We use Grokbot for our business every day, but it, it also had a consumer friendly vibe. And I think that's what maybe people were like.

• 21:56Jason Calacanis?

  Sure.

• 21:56Speaker 01

  Oh, I've just been enjoying my Grokbot and now all of a sudden I have to pay more. But for me, I would pay, I would happily pay more. It's, it's worth it. Speaking of things I would pay more for, but don't have to, Jason, you'll notice we're wearing our Plaud pins. I don't know if you have a Plaud pin on you.

• 22:12Jason Calacanis?

  Well, I mean, I have it right here on the charging station for a reason.

• 22:15Speaker 01

  They're extremely useful.

• 22:16Jason Calacanis?

  Pop it on here.

• 22:17Speaker 01

  Record whatever you're talking about, whatever you're meeting about, whatever you need. It's going to store it for later and not just record a transcript, but then you can apply that transcript, use it, reformat it, however you want to just sort of keep at track of everything that was said in whatever format that you want it to take away in.

• 22:36Jason Calacanis?

  You know what we need to figure out by the next episode? Is there a connection between Grokbot and Plaud? All right, good. Do that for the next episode, Lon. Do a little extra work here. I put a little extra work on you. Like get that set up and I will show it next time and show us some real world examples of how you got this work. Like maybe it's giving you to do list ideas and it sucked it out of your transcripts.

• 22:56Speaker 01

  You know what I'm going to do? I'm going to double up. I'm going to use this for the show and then I'll write about it on the This Week in AI newsletter. So that's it.

• 23:02Jason Calacanis?

  Oh, yes. This Week in AI Newsletter. Great one.

• 23:05Speaker 01

  Great one. So if your work depends on conversations, interviews, meetings or calls, or you have a Grokbot and you want it to be able to listen in on your meetings, you could check out Plaud at plaud.ai/twist. That's P-L-A-U-D.

• 23:17Jason Calacanis?

  Don't wait.

• 23:18Speaker 01

  A-I slash twist. Use the code TWIST and they're going to give you an extra 10% off on your Plaud.

• 23:23Jason Calacanis?

  Do me a favor, don't wait. Go buy one now. Tell every boy Jake House, don't you? Okay. Uh, let's keep going.

• 23:28Speaker 01

  Uh, we have a, we have an interview guest. Do you want to bring in our guest?

• 23:32Jason Calacanis?

  We do. Wow. Let's, uh, let's, uh, let's talk to our guest, of course.

• 23:35Speaker 01

  Let's bring in Oliver. He, Oliver Cameron, he is the co-founder and CEO of Odyssey. Of course, they build world models, AI trained on video that learns how the physical world works. So things like robots, cars, drones, physical AI, they could practice inside simulated worlds instead of a real one. We also wanted to bring you on because a few weeks ago, uh, Odyssey posted a video of a simulated fruit fly brain flying around inside a world.

• 24:00Jason Calacanis?

  Oh, yes.

• 24:00Speaker 01

  It's AI was imagining in real time and we, we gotta dig in a little bit more into this.

• 24:05Jason Calacanis?

  Thank you for getting Oliver on the program. How you doing Oliver? I, I saw this. I was quite taken by it and I said to my team, bring me whoever is running this Odyssey ML. And because we're, we're playing the video and they summoned you, uh, they found you. This thing went viral. You got 6.2 million views on it. X.com is where like everybody is talking about AI. Like if you're on, I saw Bill Ackman said the other week, Lon, if you're on X, you feel like you're two weeks ahead of the news cycle because whatever's happening on X, kind of like two weeks later, the, the regular press gets to it. It is kind of true. And so we're looking at what looks like, I don't know, Spain, maybe a Portugal.

• 24:49Jason Calacanis?

  I don't know what town this is, but there's a tiny, um, little bug there flying around that world. Is that world a AI constructed world or is that a video of the real world from a drone?

• 25:04Speaker 02

  So this is entirely imagined by world model. And importantly, it's, uh, imagined in real time.

• 25:12Jason Calacanis?

  Imagined in real time. So it's building the simulation on the fly, like a video game might. Lon, when you saw this, did you initially think, did it cross the uncounted valley? Did you think this was a real video?

• 25:22Speaker 01

  I did.

• 25:23Jason Calacanis?

  From a drone?

• 25:23Speaker 01

  The, the first, my first impression was the simulated fruit fly was the AI part. And this was just real world footage. And then you read it and you're like, oh no, that's just like, it looks like a river in a city and, but no, it's just, uh, it's just all made up.

• 25:38Jason Calacanis?

  So the question, I guess, is why did you do this? Why did it go viral? And what does it say about simulation theory?

• 25:46Speaker 02

  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 world models are essential to enabling a billion robots and a trillion agents. For robots, world models are this general brain.

• 26:30Speaker 02

  And it allows robots to better predict the consequences of their actions, uh, and to thus better interact with the world. And then for agents, uh, 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?

• 27:14Speaker 02

  And as you can see here, the fly flew around as a fly does. And, uh, I, I think the reason it went viral, you know, is that it's this perfect, um, perfect moment of saying, are we living in a simulation? Right? We, we have increasingly capable artificial intelligence. Uh, 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.

• 27:56Jason Calacanis?

  And for people who don't know, Lon, there is a concept called simulation theory.

• 28:00Speaker 01

  Yes.

• 28:01Jason Calacanis?

  Uh, you as my, uh, very astute follower of science fiction can probably do a, an exceptional job explaining what simulation theory is and why this, uh, you know, is a milestone in some ways, a waypoint.

• 28:19Speaker 01

  Sure.

• 28:19Jason Calacanis?

  Uh, in simulation theory, in, from your perspective.

• 28:21Speaker 01

  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, uh, 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. Uh, 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.

• 29:06Speaker 01

  What, if we had much more sophisticated computers, much faster GPUs, much more, uh, 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.

• 29:24Jason Calacanis?

  Yes.

• 29:25Speaker 01

  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 want to learn.

• 29:37Jason Calacanis?

  What percentage? We, we have P do them 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?

• 29:55Speaker 02

  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.

• 30:14Jason Calacanis?

  Sure.

• 30:15Speaker 02

  But, yeah, on its way.

• 30:16Jason Calacanis?

  Got it. 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? You know, 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?

• 30:40Speaker 02

  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 a faithful to that idea. They sort of short circuit a few of these and have versions that I'm not quite convinced by.

• 31:21Speaker 02

  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. And so-

• 31:39Jason Calacanis?

  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 you're, 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-

• 32:09Speaker 01

  Core, core reality, the matter you're in.

• 32:13Jason Calacanis?

  Core reality. So where are you at, Oliver? Be honest here, you're not going to lose investment or anything. People need to know where people building world models sit. Where do you sit? Be honest.

• 32:23Speaker 02

  I'm 90% certain that this is a simulation, a very advanced simulation of some sort.

• 32:29Jason Calacanis?

  This is why I'm the world's greatest moderator, Lon. I can convince people to actually answer the question. I am 50/50. I'm 50/50 at this point. Where are you, Lon?

• 32:38Speaker 01

  I used to be a lot more skeptical. I sort of thought this was like, people read too much sci-fi and then they get like wacky ideas. Because it's always like, we'll be like, well, what's the point? You know, like, that was my way in. It's like, why would an alien species do this to us? Like, why would you run this all day? What would be the point? But now that we are humans building them and putting animal brains in them, it's like, oh, the point would be to see if you could do it, like that, unlock that for me. So now I'm a lot more open to it because it'd be like, yeah, if you had the technology, someone would be like, oh, let me put 6 billion human simulated brains in here and see what happens. Like, maybe we're a see what happens. I can't rule it out.

• 33:20Jason Calacanis?

  I'm gonna, I'm gonna show you a paper. I have a friend, Ali. We used to play cards in LA when I was there. And, you know, when you play poker, and this paper, we'll put the link in the, in the show notes, or you can search for it. "Beyond Perception: Proposing Our Reality as an ASI Alignment Simulation."

• 33:38Speaker 01

  Sure, okay.

• 33:40Jason Calacanis?

  So it's proposing that our reality, Lon, is an ASI simulation, a super intelligent simulation. And the concept is like, you know, we, our universe is a simulation designed to train and test superintelligence. In other words, somebody created this in order, and it's getting kind of meta here, to see what happens in a simulation when the participants in the simulation make superintelligence and can they reach alignment? Or will they blow everything up? Right?

• 34:18Jason Calacanis?

  So this is like, you know, and as part of the evidence in this study, the proposal would be the fidelity of the world. And Oliver, we'll see if this dovetails with your work. So, you've been listening to the show for long enough. You've been inspired. You're starting your own company. AI maybe pushed you. Superintelligence maybe pushed you to the edge. And now you're gonna jump in. You need to have a partner you can trust to help transform your idea into a real business. And I recommend our friends over at Northwest Registered Agent. Northwest has just launched free identity services. They're gonna give you a free domain name, plus open source website hosting, a business

• 35:00Jason Calacanis?

  email, a phone number, and all the other important assets that make a brand new startup feel like an established professional company, all with no purchase required. And this is free done right. Northwest doesn't hand over your data to any third-party vendors and they will never sell it. So whether you've got an idea you can't shake, or you're already knee-deep in building something amazing, start with Northwest for free. Find out more at northwestregisteredagent.com/twistdomain. Would you, as a creator of a world, a world creator, would you produce the entire world? Let's say your world is Italy and what we just saw was, you know, Venice.

• 35:40Jason Calacanis?

  Would you produce a simulation of the entire tens of millions of people in Italy in every single nook and cranny, under the bridge, 10 feet down in the water, and a fish swimming? And you would have that entirety simulated? Or, when you create a world model, is the fly's path high fidelity and everything else is dormant?

• 36:05Speaker 02

  The latter.

• 36:05Jason Calacanis?

  And why?

• 36:06Speaker 02

  Simply that if you want the fly's path to be the best possible performance, that's where you focus your GPU time. And so, everything outside of that, unessential to be computed. And so, I think we, you know, some of our work is around how do we scale up agent learning to massive scale? Right now, we can support something like 20 agents inside a world model session simultaneously.

• 36:38Jason Calacanis?

  Got it.

• 36:38Speaker 02

  And so, we can generate 20 viewpoints, unique viewpoints, simultaneously. Now, how do you scale that to a billion? How do you scale that to 8 billion to-

• 36:48Jason Calacanis?

  Yeah, if you were actually simulating 8 billion people's reality, half of them, or a third of them, would be asleep at any point in time and dreaming, or not dreaming.

• 36:58Speaker 01

  But I mean, if we really are in a simulation, how do you know there really are? Like, you only know there's 8 billion people on Earth because you've read it, because you've been told, like, we don't, maybe all of us who are being simulated are in Texas and nothing outside of Texas exists and, you know, like-

• 37:13Jason Calacanis?

  Yes, when I go to Japan, it pops up Japan as a concept and a construct.

• 37:17Speaker 01

  It creates it because now one of the people with actual perspectives are there. There could be just 10 of us in here.

• 37:22Jason Calacanis?

  And that's a limitation of the tech. Now, if we were to look, Lon, at what we experience as humans in evidence, we try to understand things like dark matter. We try to understand the universe. We try to get off this planet. It's pretty hard to get off the planet. It's pretty hard to get to the core of the planet. It's pretty hard to get to the Mariana Trench. But we have these little interludes where we can get to the moon, or we can send a probe out to the end of the, you know, our solar system, right? We can do our solar system, but we don't get much further behind that. And then whatever else we get is part of the Big Bang or not the Big Bang, this crazy story we tell each other that we're kind of guessing on, but is incredibly low fidelity.

• 38:06Jason Calacanis?

  That is evidence that you're in a simulation and that what Oliver just described to us, like, why would I do the compute of more than 20 people in the simulation? I don't have that kind of compute money. Nobody does. If we took all the compute on the planet, Oliver, we took all the computer on the planet, we gave it to you for a simulation, how many people would you have in the simulation? A million? 2 million?

• 38:27Speaker 01

  Yeah.

• 38:29Speaker 02

  Like, have you done that calculation?

• 38:31Jason Calacanis?

  Yeah.

• 38:31Speaker 02

  But we haven't, no. But I do think so. Our logo is the double slit experiment, or it's inspired by the double slit experiment.

• 38:39Jason Calacanis?

  Explain.

• 38:40Speaker 02

  Yeah, so the double slit experiment is this famous experiment that basically says, when you pass light through two slits and you observe it going through those two slits, do you get a pattern that is conducive to light being a wave or to a particle? And that is something that we very much believe is indicative of us understanding how the universe works. When you see something, it observes this pattern. When you don't observe something with an instrument, it has this different pattern. And that might be indicative of some sort of,

• 39:27Speaker 02

  I don't know, performance characteristic of the universe.

• 39:30Jason Calacanis?

  In the double slit experiment, you have like a light going through two slits in a piece of wood, a piece of paper, whatever it is. And then you have it projected. And so here's your light source. It goes through a metal sheet and then it creates light and then the light hits a screen. Now, when you're observing it as a human, you see a different pattern than when you're not observing. In other words, the observer changes reality. Am I correct in describing that as a layperson there, Oliver?

• 40:02Speaker 02

  Yeah, to an extent. I don't think we really understand yet what observation really means. Like, you know, to us, yes, we're observing it via eyes. Is it that simply particles are interacting? Is that observation? And there's really limited understanding of this.

• 40:18Jason Calacanis?

  And then, popular science fiction, Lon, you have this moment in time when Kryptonians come to Earth because they have supernatural perception. When they take their mask off and they get hit by the sun, they hear everything on the planet. They see everything on the planet and the General Zod drops to his knees because he's disoriented and Superman's been on the planet for 30 years.

• 40:40Speaker 01

  Earth's yellow sun, yeah.

• 40:42Jason Calacanis?

  Yeah, but Superman's learned how to tune things out and fine tune it. So what is the business here, Oliver? Now that we've gone down a science fiction rabbit hole for 15, 20 minutes, which makes it super interesting. You know, obviously, you're not building a simulation to have, you know, philosophical discussions like this and pop culture discussions. You're doing it for a business. What is the business here and how close are you to that business?

• 41:05Speaker 02

  So the business is world models for robots and world models for agents. And before starting Odyssey, I spent eight years building driverless cars. And it turns out, we were building that technology all wrong, I believe. And the reason we were doing it all wrong is because what intelligence should learn from is a very general set of concepts before it tries to accomplish a task. Now, what we did in driverless cars is just show the same task over and over again. We showed those models. Here's how you drive a car. Here's how you drive a car. Here's how you drive a car. And so then, we're a little bit surprised when it can't handle crazy edge cases

• 41:50Speaker 02

  or it gets stuck and does something dumb. Well, the reason it does that is because all it's seen is the world through the lens of a car. And so really, what a world model proposes to be is a much more general foundation for robots, big and small, like cars, to build upon. And so the paradigm now with world models is you take this very general foundation model, a world model, and then you show it just a few hours of experience. So for example, we as humans, we learn for 17 years. Then we go learn how to drive and we're exposed to about 20, 30 hours of experience of driving a car. And so the equivalent of that for world models is that you show it 20 to 30 hours of humans driving cars. And then what we see is that the model's able

• 42:35Speaker 02

  to take those very general concepts and the 20 to 30 hours of experience and then become a really good driver. And this same concept applies to robots that you can show it 20, 30 hours of manipulating some task that you want the robot to do. And then it can perform that task with lots of robustness to the world. And so this, this whole, that's our business, to build that foundation model, to then work with.

• 43:01Jason Calacanis?

  And the customers are robotics companies, factory, companies that own factories or own depots.

• 43:07Speaker 02

  We sell to physical AI companies. And so that's robot labs, humanoid labs, self-driving car companies, drone companies. They take our base model, build upon it, and then they have more reliable.

• 43:18Jason Calacanis?

  Have they started actually paying for it? And they don't see that as their core business? They think it's better to just buy yours because it's, you're working on that problem and it's like they got enough problems on their own to get their hands working on these things?

• 43:30Speaker 02

  Yeah, yeah, we generate revenue today. And I think something changed when language models really landed, which was, I think everyone went from in deep tech, we have to build everything ourselves to be more open to doing post-training as the value add. And so the robot labs we work with, they primarily post-train Odyssey 3, our latest foundation world model. And then they deploy that model into their robot. Then they deploy that in the field. And so that's definitely been a shift in the, in the space.

• 44:02Jason Calacanis?

  I mean, I'm curious because, so it's all visual input is what I'm hearing. That's how you're training. That's how you're building the world model, just showing footage. But so much of how we engage and learn about the world is, is, is more tactile. Like, I, I know which of these objects is going to be heavier before I pick it up, just because I have years of experience gauging object size and weight. And, you know, like if I'm driving down the street and there's like a paper bag blowing, I'm like, I can hit that and it won't matter. It's not worth veering out of the way. If it's a rock, I go like, ah, I gotta get out of the way. How purely from visuals can robots and cars learn everything they need to know about our world and how matter moves through it?

• 44:43Speaker 02

  So really, if you think of a world model like this, you've got this really attentive learner. You place it on a couch in front of a TV screen and you show that really attentive learner many, many thousands of years of video. And it's just sat there. It never gets distracted. It's got this notebook and it's writing notes on every single thing it sees. It's keeping notes on dynamics, on motion, on humans and everything you could possibly think about. And then at the very end of that, this notebook is all filled out with all of the knowledge that it's learned through that process. And that, that notebook is now massive, right? It describes so much about the world. And that really is how these models learn. They're just simply taking copious notes of how physics evolves over time, how motion evolves,

• 45:29Speaker 02

  how humans work together, how we don't work together. And it just learns lots of concepts from, from simply doing that. And so when it comes to world models, one of the hills that we believe is most important to climb is physics accuracy.

• 45:42Jason Calacanis?

  Yeah.

• 45:43Speaker 02

  If we climb physics accuracy, that means the model is learning fluid dynamics. It's learning gravity. It's learning all these concepts. And so what we see now is they, they are just from a very simple learning objective, which is predict during training time how the world's going to evolve next. It learns very general concepts.

• 46:03Jason Calacanis?

  Amazing. All right, listen, continued success on it. Where are you based and are you hiring at the moment? I know you raised some money, yeah?

• 46:09Speaker 02

  Yes, we've raised $370 million from folks like Amazon, AMD, Nvidia GB. We're based in the Bay Area, London and Zurich. And definitely hiring.

• 46:17Jason Calacanis?

  Love it. And the domain name, so people can go apply to work on this vision and get in on maybe some of these RSUs, get some equity in this company over the next four years because you're going to crush it.

• 46:27Speaker 02

  Thank you very much. Odyssey dot systems.

• 46:29Jason Calacanis?

  All right. Thank you, Oliver. If you find yourself in Texas or Tokyo, just say my name three times and we'll go get some barbecue or I'll take you for some soupless ramen. Yeah. Great job, Oliver. Thank you, sir. Appreciate your time.

• 46:43Speaker 01

  Yeah, his system is going to have to generate Austin for him to come here. That's got to fill it.

• 46:48Jason Calacanis?

  Absolutely, yes. We're going to have to do some fidelity of Terry Black's beef rib. The beef rib is right now dormant on a CPU. It's just in a cluster somewhere, waiting to be popped up. Man, we haven't been to Terry Black's in a while. Man, I gotta go get that beef rib. We need to do like a meeting. You gotta get back to the standing meeting at a barbecue joint. What are we doing here?

• 47:08Speaker 01

  Yeah. The last time I was there was the birth of Steelcore Capital. You and Mark Jeffrey.

• 47:12Jason Calacanis?

  Oh, yes. Yeah, man, we hit that beef rib. I mean, it's like, the beef rib is so expensive, but it's so good and it's so fatty and delicious. I can only eat like a third of one now with my GLP.

• 47:24Speaker 01

  Yeah.

• 47:24Jason Calacanis?

  Yourow.co/twist. Get yourself a GLP and lose 40 pounds. I was at a Halloween party, man. People were like, Jake, how you look great.

• 47:33Speaker 01

  Yeah.

• 47:34Jason Calacanis?

  You look great. And I was like, thank you.

• 47:35Speaker 01

  Thing about row.co, it's a two-letter domain name, Jason. How did, how did they, that must have been, that must have cost a pretty penny.

• 47:42Jason Calacanis?

  Pretty, pretty penny, pretty penny.

• 47:46Speaker 01

  Worth it, right? I mean, if you're, if you're.

• 47:48Jason Calacanis?

  I need to sell these two domains, mahalo.com and inside.com. The auction on my own auction software didn't really work too well. Didn't hit the reserves. So I'm going to put, I think I'm just going to put it on GoDaddy. You ever hear of this GoDaddy?

• 47:58Speaker 01

  I've heard of GoDaddy, yeah.

• 48:00Jason Calacanis?

  GoDaddy's amazing. I do all my register and domain names there. They also have all kinds of services around it. Like, if you want to set up a website, you want to set up mail, you want to redirect it like Disgraciad. I just go in there, disgraciad.com.

• 48:14Speaker 01

  That's right.

• 48:14Jason Calacanis?

  And I just redirect it to somebody disgraceful as a bet on All In. And, you know, it's just, it's just sitting there in my GoDaddy and I just redirect it, beep beep boop, and it's done. Two secondes. All right, listen, thank you to our friends at GoDaddy. GoDaddy.

• 48:33Speaker 01

  That's where to go.

• 48:34Jason Calacanis?

  It's the best register. And you know, you want something safe, the lock that's. They'll lock that ish down like you would not believe. I got that locked down like Fort Knox. I never lose anything. If my credit card goes bad, they call me, they email me, hey, your credit card is, because you know, if you're there for 20 years, every four years your card goes dead.

• 48:52Speaker 01

  Yep.

• 48:53Jason Calacanis?

  So then you could put in a backup card, you know, that's a good feature too. You put a backup card in, like if it doesn't hit this one, you hit this one. You never want to lose a domain.

• 49:01Speaker 01

  Yeah, I don't want to lose.

• 49:02Jason Calacanis?

  I just got an alert from the.

• 49:03Speaker 01

  Very important.

• 49:04Jason Calacanis?

  Yeah, I just got L O A A, Loa, Loa. When I was in my Hawaiian domain names.

• 49:10Speaker 01

  What does that mean in Hawaiian, Loa?

• 49:12Jason Calacanis?

  Um, Loa, L O A A, uh, is to get, obtain, receive, or find. So to find something.

• 49:20Speaker 01

  To find.

• 49:21Jason Calacanis?

  L O A apostrophe A, Loa.

• 49:23Speaker 01

  Loa.

• 49:24Jason Calacanis?

  Uh, like I received the letter, you know, I found the letter is, uh, ooh Loa. Ooh Loa, I received, I found. Uh, Loa without the extra A has a different meaning. It means far, like a long mountain, a far distance, that kind of thing. So I was like, when I was doing mahalo, uh, which was like, thank you and heartfelt, thank you, uh, appreciation. I, I got Loa first to find, right? Like a Google.

• 49:50Speaker 01

  Right. It was a search company. So that makes total sense. Yeah.

• 49:53Jason Calacanis?

  Yeah, man, I, talk about being ahead of your time. Can you imagine? Everything we did is like the perplexity search. Ah.

• 50:01Speaker 01

  Data centers, Jason, they remain consistently unpopular and opposition is strongest when people are asked about building them in their own communities. I pulled a few, uh, little bits of data. The, uh, ABC Washington Post Ipsos poll from October 2026 says 66% of Americans oppose data centers being built anywhere in the U.S. 30% support them, uh, and support peaks among 53% among MAGA Republicans specifically, uh, 80% opposition to data centers among people 30 years old and under. Now Axios reports that OpenAI is planning a PR push for after Tuesday's midterm elections. Uh, the idea on the table would be AI speed limits, pushing

• 50:43Speaker 01

  the U.S. to help write global AI standards. Selling AI is something that empowers people day to day. And here's the best one, renaming data centers as quote, AI compute campuses. Uh, OpenAI's, uh, chief global affairs officer, Chris Lehane, told Axios that the midterms will be a bit of a movie trailer for 2028, when he expects AI to be the signature issue of the presidential election. Jason, my question's for you. Do you agree is AI the thing we'll be debating for the next two years into 2028? And two, is it too late for an AI data center rebrand? Can we call them compute campuses? Do you think that's gonna, that's gonna fly?

• 51:25Jason Calacanis?

  It is going to certainly be the topic of the rest of our lifetimes. So not just this midterm tomorrow. Check out the all in live stream tomorrow night. That'll either be, uh, a bunch of Republicans crying in their coffee and people on the left celebrating the two most unpopular, the most unpopular political party versus the most unpopular president of our lifetime. It's gonna be pretty wild to watch.

• 51:53Speaker 01

  I will say that we've, the expectations now, if Republicans sneak away with even a halfway decent showing, they're gonna, they, they win. Like they can win without winning at this point because the perception is that it's gonna be this Democrat blowout.

• 52:08Jason Calacanis?

  And I don't think it actually matters when it comes to the data centers and AI and superintelligence, who wins this midterm. And I don't even think it matters who wins in 2028, I'll be honest, because the person, the people who win are going to be the ones who are the most pro-human and anti-superintelligence. Why? Because right now, people correctly, um, realize there's going to be significant job loss to this technology. Now, what we all jump to as technologists, and we've had this conversation endless times in the last three years, is new jobs will be created, new opportunities, we're going to retire a certain work, we're

• 52:50Jason Calacanis?

  going to have more work. Obviously, that's all true as well. But I think, you know, you will selectively pick this argument based on where you sit in the capital stack. We sit in the top half of the capital stack. I sit in the top half, the half of the capital stack, the top 10% of the top capital stack, because I'm investing in the companies, you know, I hit a unicorn, you know, a, a bunch of money comes dropping from the ceiling on my head and everybody says, wow, he's just a luck box. He's got no skill. He's just gotten lucky 14 times. Um, which, you know, like maybe that's possible that somebody could get lucky 14 times. I kind of think after 12, you gotta give the person some credit for at least one or two of them. Um, and then if you're on the bottom half, you're probably

• 53:33Jason Calacanis?

  looking at this technology and you're seeing that you don't own the stocks, therefore you didn't get in on SpaceX or Anthropic or OpenAI or whatever the next company is. So you are not connected to the equity wealth creation. Then you probably know somebody who drove for Uber, Lyft, or DoorDash, who had their job replaced by RoboTaxi, Waymo, ZipLine, you know, DoorDash's delivery vehicles, automated delivery vehicles. So that's what's gonna happen over the next three, four, or five years. And then you're gonna have both sides of the equation. The libs on the left who can't put their graduate degrees to work

• 54:12Jason Calacanis?

  and are gonna be like, well, I got a $250,000 graduate degree undergrad loan and, um, all the work I did can be done by a bot. And I actually don't have judgment. I certainly don't have motivation to do things. I don't want to start a company. I want to be overpaid, pay down these loans and, uh, yeah, be a unique snowflake in the world. And then on the right, it's gonna be like, well, I had this physical job that now Optimus is doing, uh, Figure is doing, some robots doing. And the robots could also be self-driving, obviously. And so that's gonna start hitting both sides. And unless we have some way to make the bottom

• 54:53Jason Calacanis?

  half feel like they're part of this, which, you know, Trump accounts, Invest America on a margins is starting to, uh, we're gonna just need to really do a full-on effort. This is why I told my friends who were involved in Trump, and I said it on the All In podcast, where I talk about politics more than I would prefer, just do three things. Make a ton of homes. Make trade schools free, a fixed education, and then have a baseline for, and use this technology to lower the cost of healthcare dramatically and give people a baseline of technological, technologically delivered healthcare, and you could turn this around. If people

• 55:33Jason Calacanis?

  didn't have abject fear about those three issues, we would be a lot further along. And I know that to be true because I live in the great state of Texas where the home issue's been taken off the table. People feel great about housing here. Uh, if they're complaining, it's like little nitpicky things. And they feel actually great, Lon, about college because UT is free for anybody in the state who makes, who have a household income. I think it's gonna be a hundred thousand. If your state income, it used to be 60 or 70, but now it's a hundred thousand. If you're under a hundred thousand, you can go to UT or any of the, the, the state schools for free. And if you have to pay for it, is it $10,000 a year

• 56:15Jason Calacanis?

  or something absurdly cheap that anybody could afford? Um, so you, you, you put those two things together. Now, healthcare, we don't have figured out, but man, if you go to urgent care here, it is dirt cheap, no wait. I've been to it. Um, and it's pretty, quite reasonable. So, uh, that's what I think's gonna happen.

• 56:33Speaker 01

  I think the companies think that they can just do the same kind of like, what we used to call like cause washing or greenwashing, you know, and the companies would be like, they'd take out a bunch of ads touting their environmental record and you wouldn't necessarily see any of the results, but you'd be like, well, Chevron says they're cleaning up, uh, the lake, so I guess I should trust them. I think that's what's happening now. Pull up that, pull that up, Jacob. If you live in Texas, you see these ads all over the place for Meta, like, we're not just building data centers in El Paso. We're pumping money into El Paso. We're making El Paso a great place to live. And like, I, I, I don't live in El Paso. Maybe it's true, but it doesn't seem like, it, it feels like this is marketing.

• 57:15Speaker 01

  And I don't feel like there's a groundswell of people in El Paso, Texas, who really feel like, thank God Meta is here and building these data centers because El Paso is a much better place than it used to be. And I just feel like, I don't know if we've crossed that chasm where the, the companies are actually spending the money in a way that is felt by everyday people as opposed to building websites where they talk about how they're gonna do that.

• 57:42Jason Calacanis?

  I mean, if you're building a data center in one of these towns like El Paso, it's going to have a major impact eventually. Um, it may not have the impact today. The website might be a bit, um, you know, manipulative.

• 57:56Speaker 01

  They're running commercials for this too, where it's like guys in hard hats and they're like, everybody used to have to move out of El Paso to go get a good job. But now that the data center's here, we all work for Meta. And it'll be like a guy with a bulldozer and, you know, like they're, they're really, they're all in on the idea that like Meta is gonna save the town of El Paso with this data center. And like, I haven't talked to a lot of El Pasoans. Maybe they do feel that way, but I'm skeptical.

• 58:21Jason Calacanis?

  Okay. So there is do no harm. Like, don't make my electric bill go up.

• 58:25Speaker 01

  Right.

• 58:25Jason Calacanis?

  That is table stakes and it's done.

• 58:28Speaker 01

  Yes.

• 58:28Jason Calacanis?

  Uh, don't be selfish and take something from the community. That used to be what we did 10 years ago in the industry. You would bid one city off another to get tax breaks. And remember, Amazon did it. And I was on CNBC talking about it and they were gonna like have their second headquarters somewhere and they wound up splitting it between two cities.

• 58:46Speaker 01

  It would be this like long ongoing, like maybe we'll put it here and then they would see what the city council would offer them.

• 58:52Jason Calacanis?

  Right, exactly. Right. And that's ended because that felt like, well, and correctly so, people were like, well, wait a second, what are we getting out of this? Now, if you said 10,000 jobs, great, okay. But when you talk about a data center, you're talking about very few jobs after the thing has been finished.

• 59:08Speaker 01

  Right, they're building it and then once it's there, there's a bunch of services.

• 59:11Jason Calacanis?

  Onto the next. Unless they're gonna continue to build. So we went from, we're gonna extract something from the local community to, we're not gonna do any harm, we're gonna pay our way for the electricity, to now, we're going to subsidize this town, we're gonna train people to be electricians and HVAC folks for free, which Met is doing, so give them a lot of credit for that, we're gonna pay for the local stadium, we're gonna lower your electric cost, we're gonna lower your water cost. And I think they need to be more explicit about, when I look at this website, investing in El Paso, 300 operational jobs is nothing. I mean, it's not nothing, it's something. I don't know how many people live in El Paso, but it's, or how many people are working in El Paso right now, but I don't think it's that much.

• 59:54Jason Calacanis?

  4,000 skilled workers on site at peak construction, again, they're in and they're out, they're probably causing as much of a ruckus at the local bars as they are contributing, if you watch Landman.

• 1:00:04Speaker 01

  650 to 700K people in El Paso, so a decent sized town, not that much smaller than Austin, really.

• 1:00:11Jason Calacanis?

  How much bigger? How big is it?

• 1:00:13Speaker 01

  Austin's about a million, this is about 650,000 in El Paso.

• 1:00:17Jason Calacanis?

  Okay, so 6,000 would be 1%, 600 would be 10 basis points, 300 would be 50 bips. So, I mean, it's, it's, I mean, it's some sustainable jobs.

• 1:00:29Speaker 01

  It's not zero, it's not zero, yeah.

• 1:00:31Jason Calacanis?

  Right. So, they need to figure out, like, how much is actually going to the libraries, the roads, etc.

• 1:00:40Speaker 01

  That's what I'm saying.

• 1:00:41Jason Calacanis?

  That is not for them. And they, they, they, they don't exactly quantify that. I think what they should say is, we're gonna be giving $100 million over 10 years to the community to spend as they see fit.

• 1:00:55Speaker 01

  Right.

• 1:00:56Jason Calacanis?

  So, and, and just put the, where is the tax number here? Like, how much are they gonna pay in local taxes? They should just say, we're gonna pay $100 million a year in local taxes, $1 billion over X number of years. And then people could say, okay, yeah, and, and give it to the community, because local taxes, they say tens of millions, but...

• 1:01:15Speaker 01

  Why isn't there, why aren't they getting like the Meta El Paso arena? You know, like, it's just like, do, do something visible that everybody in town is like, well, Meta built that, it's great, I love it, we go to the Meta Dog Park every weekend, you know, like, I feel like it's always...

• 1:01:30Jason Calacanis?

  Well, they do do that kinda stuff. I, I would argue the other side of that, which is, don't do the, like, hand wringing signaling stuff, because I feel like that's a little bit performative. I think just giving a certain amount of cash in taxes at a minimum, we're gonna do a minimum, instead of it being 8, it says 80 million in taxes, and who knows how they get to, we're just gonna give 25 million a year in taxes as our, you know, minimum tax payment for the next 10 years. So as long as we're here, we're gonna pay 25 minimum in taxes, and we're, if you don't have a way to give us that money, we're just gonna gift it to you. And just, they know the economic value of this thing, so I think they should just give a, pick a number, give it to the city council, and then let the city council have arguments

• 1:02:14Jason Calacanis?

  of how to spend it. Because they might say, you know what, the football stadium's the priority. No, we think training's the priority. No, we think this bridge is the priority.

• 1:02:22Speaker 01

  Yeah.

• 1:02:22Jason Calacanis?

  Whatever it happens to be.

• 1:02:23Speaker 01

  Paying teachers to move to El Paso or something.

• 1:02:25Jason Calacanis?

  I, you know, I, one of the things that happens, you know, with these local communities, there is an interesting phenomenon in rich communities. And I've lived in, uh, how many times have I lived in an affluent community? Twice. Brentwood and then, and then in the peninsula.

• 1:02:42Speaker 01

  Yeah.

• 1:02:42Jason Calacanis?

  Yeah, in the peninsula. What happens is, you have a lot of expensive homes, you get charged a percentage of the home value, and the local schools are run by that district, they become overfunded, and people who are in there, who are affluent, send their kids to private schools, so the people who do send their kids to the local school have, I'm not kidding, a 5 to 10 to 1 teacher to kid ratio, it's better than the private school ratio. They get overfunded, in other words. That would be what I would do if I was Meta, I would overfund the school system and say, your school system is gonna have a 5 to 1 kid to teacher ratio because of what we're doing. That would be, that would resonate with people.

• 1:03:23Speaker 01

  Something visible that people can feel that's visceral, that's not just a commercial or like guys in hard hats, but like, something that people are like, okay, it's super valuable, I'm glad Meta's here. That's, that's what I think.

• 1:03:34Jason Calacanis?

  I gotta say, on the self-driving front, listen, and I, I'm invested in like, every self-driving company, from Uber to Tesla to Noro, not Noro, uh, you know, all the publicly traded ones, Ryride, Pony, Rivian, Lucid, I have like little pieces of everything, because that's a basket I made of, you know, I think this is gonna change the world.

• 1:03:52Speaker 01

  Sure.

• 1:03:53Jason Calacanis?

  I think, and so I'm working against my own book here, I'm not opposed to throttling the pace of automation and self-driving, because I do think what we're seeing here in Texas is that these companies can go a bit faster in Texas than other locations. We're gonna lose all those jobs real quick, man. The pace these things are going at, Waymo is expanding, Uber's expanding, Zooks is expanding, Tesla obviously is expanding very quick. I don't, I'm, I mean, are you taking Uber or are you taking robotaxi, Waymo? What are you taking on when you...

• 1:04:26Speaker 01

  I, if, if Waymo is an option, every time, every time I would do a Waymo. I, I prefer the Waymo experience to these straight...

• 1:04:34Jason Calacanis?

  Because it's quiet and it's, you're alone?

• 1:04:35Speaker 01

  No, like, nobody, nobody bothers me, you don't get that chatty, what out of every four to six Uber drivers, they're bored, they've been on the road all day, they wanna have a little chit chat, I don't wanna do a little chit chat. You don't have to worry about that with Waymo, you don't have to tip with Waymo, and honestly, I find it, I, I find it relaxing, I, I like being alone on the road.

• 1:04:55Jason Calacanis?

  It is zen.

• 1:04:56Speaker 01

  It's very zen.

• 1:04:56Jason Calacanis?

  It's a moment of zen.

• 1:04:57Speaker 01

  I could get things done, I, I prefer it significantly, I would take a Waymo every time.

• 1:05:01Jason Calacanis?

  There should be two, there should be two additions to the ride-sharing Lyft, you know, Grab and Uber and everything. Number one, when you set your preference, you can set preferences for, I think Uber Black, but I don't think they have it for Uber X.

• 1:05:14Speaker 01

  Right, they don't.

• 1:05:15Jason Calacanis?

  They should, they gotta make setting preferences for Uber X a thing, temperature, music, no music, then they should put small talk.

• 1:05:22Speaker 01

  Yeah, I just, I know some people love that.

• 1:05:25Jason Calacanis?

  No small talk, love small talk.

• 1:05:26Speaker 01

  Yeah.

• 1:05:27Jason Calacanis?

  Don't care.

• 1:05:27Speaker 01

  And big U is...

• 1:05:28Jason Calacanis?

  And then, when you're reported at the end, they should have cleanliness of vehicle as an option, it was very, it was pristine, it was tidy, it was messy, then they should have smells. It was neutral, or it was, it, it, it had some smells that were offensive.

• 1:05:50Speaker 01

  I think I've only over one time had a real smelly Uber situation, that's really not one of them.

• 1:05:54Jason Calacanis?

  I have two versions of this, one, you know, the perfume things, they put everywhere, I'm very sensitive to, so I start sneezing, so I just have to ask, and you know what, the black cars have that more than anybody, they put like three of them on every...

• 1:06:06Speaker 01

  Right, they want it to smell nice, yeah.

• 1:06:08Jason Calacanis?

  And they put that in the, um, the blower, so you put air conditioning on it, it's blowing it, so now you're basically got an, uh, you know, a, a wind blowing machine with peppermint or lemon scent blowing in my face, I literally, so sensitive to it, my ear, my, my eyes can tear and I start sneezing. They should have scents. I prefer no scents, etc.

• 1:06:30Speaker 01

  Yeah.

• 1:06:30Jason Calacanis?

  And so they should put that at the front of it, because it's very easy to actually get rid of the scent, you just put those things away.

• 1:06:35Speaker 01

  So Anthropics, Andre Karpathy got fed up with YouTube's recommendations, over the weekend, he built his own YouTube, it is called HackerTube, he had Claude go through 43 million hacker news comments going all the way back to 2007, he pulled out the YouTube videos people kept recommending year after year, so the result is a customized library of 6,454 videos across 150 topics, no login, no algorithm chasing your clicks. Karpathy said it took about two hours of Claude code with Opus 5.5 to build it, and he calls it a significantly better viewing experience. Have you, have you tried out HackerTube, Jason?

• 1:07:15Jason Calacanis?

  I, I pulled it up, it's, it's awesome, um, it's just great to, um, make new interfaces for data sets. YouTube should be thrilled, because if you click on these, it's gonna still have the ads embedded in it. Um, what you've done is you've removed the cruft, and I think this is going to be a very interesting moment in time. Anybody could have built this, you know, in 60 days with two developers and a designer, but nobody would take the time to do it. Um, I did something similar this same past week. I've talked before about my frustration on Amazon in just trying to get to a, uh, a product to buy.

• 1:07:55Speaker 01

  The real product, below all the ads and sponsored whatever, whatever, yeah.

• 1:08:00Jason Calacanis?

  So, I, in, while we were on the pod last week, we were talking about...

• 1:08:05Speaker 01

  Yeah.

• 1:08:05Jason Calacanis?

  Shopping or whatever, and I just wrote like two sentences, I gave it a simple prompt, and then I forgot I did it, and then I got back to my desktop and there was an alert in Claude, and, uh, take a look at this tweet I sent you, um, you pull it up there. I made my own, unsponsored, and if you click on the thing on the left, the image on the left...

• 1:08:28Jason Calacanis?

  It made all this. So it, and I didn't give it any instructions other than I just want to get rid of these and it with a, with a Chrome extension. It said, uh, okay. Um, I'm going to build five different things you could turn off. Sponsored products, brand banners and videos, sponsored carousels, display ads and Amazon brand promos. And it by default click four of them off. And then it said, removed listings are hidden, dimmed or outlined. And then it keeps a log of it and it keeps track of how many you do. I didn't ask it to do any of that kind of stuff, but it did it anyway. Made this beautiful branding name and hit the right arrow key there. It, uh, here's just an example of the red bar around it. Like, you know, I'm going down through, look, I just put in best coffee machine.

• 1:09:13Jason Calacanis?

  And you know, it's like, I'm going to get slapped with this like vegan French coff- pressed coffee makers. I don't want this in my feed. I, I, I just want to know the best ones.

• 1:09:23Speaker 01

  Right.

• 1:09:23Jason Calacanis?

  And start removing it. This is going to increase my usage of Amazon. Now, is this illegal or not for the terms of service against it? I, I would never publish this and give it to other people because I think that would be against the terms of service. But what you do on your local machine is up to you and how you view stuff. There's an exception for this typically. So if you own the New York Times and you want to cut it up and make a montage, or you want to take a webpage and edit it, you can do whatever you want on your local machine. It's when you publish it or give it to another person and you subvert their ability to make money, that this becomes an issue. It's not an issue when you do it locally, but I would never have hired anybody to write this, but I could certainly tell it in two seconds. Uh, and that's, this custom software and custom view of the world is coming soon, folks.

• 1:10:06Speaker 01

  You have to treat, retrain your brain. As a non-developer, I think you have to re, because for me, I don't think about like new products, like because I never had the ability to build them. So I have to like, I have to retrain myself to like, if something's not working, maybe you can fix, maybe you and Claude can fix it. Like just tell Claude what you want. And like, I don't even have those ideas, but now I have to start because everybody can be a hacker and a developer now. You just have a, a friend who can just do it for you.

• 1:10:39Jason Calacanis?

  I did another thing. I went through my entire, uh, D, I had, have DMs open on X. I said, uh, and I never go through them. I said, tell me who, uh, you know, the most followed people are who have DM'd me that I missed.

• 1:10:53Speaker 01

  Oh, yeah.

• 1:10:53Jason Calacanis?

  And then anybody who's a founder, tell them you can apply to our programs here, here and here. So it brings me up to the list. And then like a very famous basketball player had DM'd me in August and said, hey, I'm in Austin. I want to get a cup of coffee. I didn't see it.

• 1:11:04Speaker 01

  Oh, man.

• 1:11:05Jason Calacanis?

  So I just DM'd them and said like, uh, Knicks in five and, uh, here's my phone number. So you have it. And yeah, let me know next time you're in town. Sorry, I missed this. Uh, and then it sent 500 people from the last couple of months. A DM that said, hey, this is Jason's bot. I always have it do that. Here, uh, you can apply to founding university, launch accelerator, or here if you want to engage with our firm. And then, you know, half the people were like, no, no, I want to talk directly to you about this. Can you? All of them started talking to my bot.

• 1:11:32Speaker 01

  Sure.

• 1:11:32Jason Calacanis?

  They're like, tell Jason, I have this incredible thing. Tell Jason, I want him to come speak at my event. And it, it didn't exactly get the response right, but I was like, better they get something than, than nothing.

• 1:11:43Speaker 01

  Yeah, exactly.

• 1:11:44Jason Calacanis?

  So here we are in LinkedIn. I can't do that unfortunately. LinkedIn is not letting automation of, so this is a message to LinkedIn. Let me use my bot to DM people, but charge me five cents, 25 cents every time I do it to throttle it. So not everybody does it and limit me to 10 times a day or a hundred times a day. I don't care what you charge me for the in-mails, but since you are letting people pay you to do intros or to DM people who are on the system, why not let my bot do it? Charge me double when my bot does it. Or make a bot yourself.

• 1:12:22Speaker 01

  Yeah.

• 1:12:22Jason Calacanis?

  Yeah. Just have agent mail and just be like, hey, this is Jason's agent. Cause people are already doing this with, you know, uh, other tools that are hacker-like tools, just incorporate it into the platform. Let throttle the number of people. And if people complain about it, just give the power to those people. On the bottom of it, just say, uh, never want to hear from this again. Or, uh, and then you never hear from that person again.

• 1:12:45Speaker 01

  I, I feel like all these platforms that are being so bot and agent hostile, like it's a losing, like you, it's, it's over. Like people are going to have agents, they're going to have bots, they're going to want to access your reddits or your hacker newses or your LinkedIn's. Like there's, there's nothing you can do.

• 1:13:00Jason Calacanis?

  Not a big deal.

• 1:13:01Speaker 01

  You just, there's, I don't think there's any stopping. You can become irrelevant or you can let people's agents in. Those are your two choices.

• 1:13:08Jason Calacanis?

  All right. So tell me, uh, what's going on with space? I know, uh, Google and Planet Labs are doing an experiment together. And then Nvidia and SpaceX have their own plans, but they're all going up on SpaceX rockets.

• 1:13:21Speaker 01

  Yeah.

• 1:13:21Jason Calacanis?

  Which tries to be neutral about who can use the rockets.

• 1:13:23Speaker 01

  We're, we're, yeah, we're talking about Google had the big breakthrough. Their first data center in space is now up and running. It's a prototype satellite from Project Sun Catcher built by Planet Labs, as you said, carrying four of Google's TPU AI chips. It launched on a SpaceX Falcon nine earlier this month. This is the, uh, the first test that Google did to make sure things were working in space. They asked Gemma four, that's the model running on board to write a haiku about being in space. And here it is. Blue marble below, data streams in silent grace, starlight feeds my core. So it's working. You got the syllables right.

• 1:14:03Jason Calacanis?

  All right. So it's like a hello world. The new hello world is make a poem.

• 1:14:07Speaker 01

  Write a, write a haiku for us. Yeah. And so there are real limits. Only one of the four chips can run at a time. And it only runs for about 20 minutes before it runs out of power. They've got two more satellites with more advanced chips planned for 2027. Google is aiming for space data centers that make economic sense in the 2030s. So we were just talking about how much Americans hate data centers. They don't want data centers around. Is this a viable option? Could we just blast them into low orbit instead?

• 1:14:35Jason Calacanis?

  Absolutely viable. Um, after three or four iterations, the same way internet via satellite was viable, but we spent 10 years in the wilderness with slow internet. You know, if you were on a, on a ranch somewhere, you could have a very expensive internet connection via satellite. It was originally giant satellites and medium sized and smaller, huge satellites, whatever they were. And, you know, sometimes I'd rent an Airbnb, you know, somewhere on the edge where they didn't have broadband. And they would give it to you, but you know, you'd watch one Netflix movie, it would timeout and say you hit your limit.

• 1:15:10Speaker 01

  Yeah.

• 1:15:11Jason Calacanis?

  So they had it like limited by a certain amount of bandwidth they charged. It wasn't unlimited. So, um, that all changed. And now we sit here with Starlink and we're like, this is the greatest thing since sliced bread. Um, and it is, and, and there'll be three competitors. Uh, Amazon's got Leo, I guess. Uh, their low earth orbit satellites. It's going to be a same, similar timeline here. We might spend a decade, you know, making proof of concepts and things that can lightly do this stuff. But it really is all based on Starship. Starship has a massive amount of payload. I don't know if it's 10 or 20 times the previous Falcons, but it's significantly more. I actually have been inside that when they were first making them, I was at Starbase with Elon

• 1:15:55Jason Calacanis?

  and we were standing inside of it, the first prototype that was like hand welded together. I mean, and you look up, it's like you're looking into a small, you know, uh, office tower, you know? I don't, I don't know how to explain it, but you're like looking into a payload that could be 200 people in seats. You know, like looking into a Boeing.

• 1:16:17Speaker 01

  Starship can lift a hundred to 150,000 kilograms to low earth orbit. That's about four to 6.5 times more than Falcon nine's capacity.

• 1:16:26Jason Calacanis?

  Okay. So if it's five times the Falcon launches and they think they can do multiple a day, then this kind of project becomes viable. I believe these satellites are going to be bigger, more complex, more batteries than the, uh, Starlink ones, uh, which I think are, are, are smaller. But who's this? I mean, these could have Starlink and compute in them. So like, why would you put the Starlinks out there and not include some computer in them? They obviously have some computers in them. So this will be viable. And then what we're going to see is people on Planet are saying, please stop putting them in space. Can I get them in my backyard? Cause they want the revenue from it. The real estate cost is free out there for life.

• 1:17:04Speaker 01

  Yeah.

• 1:17:04Jason Calacanis?

  Free for life. And I think the energy cost out there is going to be also with solar, free for life. So you take those two costs out and that's the end state. Now getting to the end state is going to cost tens of billions of dollars, you know, or maybe billions of dollars, tens of billions of dollars. But if you're building trillions of dollars worth of compute, if it costs $10 billion to perfect this technology, well worth it, right? And maybe it takes four years to pay off, three years to pay off the cost of getting to space. But there's no real estate. There's no permits. There's no issue. And people don't understand the vastness of space, Lon. If you look at planet Earth, like if you, if you just think about the surface area of the Earth,

• 1:17:49Jason Calacanis?

  it's huge, right? And, and most of the surface area of Earth is water. And then there's a bunch of it that's not used, like the Arctic.

• 1:17:57Speaker 01

  Desert, tundra, exactly.

• 1:17:59Jason Calacanis?

  Ice caps, like none of it's used. So then, and rainforests, there's still a lot of those, thankfully. Now imagine you go out into space.

• 1:18:11Jason Calacanis?

  And it's two times the surface area because, and I don't know if low Earth orbit, this would be a good question for Claude, if you can type it real fast. What is the surface area of the circumference of where the satellites, the Starlink satellites are compared to Earth? Like, it would probably be twice as much surface area, or a third more surface area. It's some larger number of surface area, right, Lon? Because you're expanding out into space.

• 1:18:40Speaker 01

  I think it's, it's, there's a three dimensionality aspect, too. Like, the surface of the Earth is sort of flat. Not like the Earth is round, but like, the surface is 2D. Whereas you could, like, in low Earth orbit, one thing could be up here, one thing could be down. You know, like, I don't, I don't know if it's an exact. But Earth's surface area is about 197 million square miles. So if you think of the inner boundary of low Earth orbit as a spherical shell, it's about 5% larger than the Earth's surface. But if you think of low Earth's orbit's outer boundary as a different spherical shell, it's 73% larger than the Earth's surface. So that's what I was saying, like, 3D space, low Earth orbit is a huge

• 1:19:25Speaker 01

  distance ring around the Earth. So where you measure on that sphere makes a huge amount of difference.

• 1:19:32Jason Calacanis?

  Yeah. So the question is, how many miles deep is, uh, low Earth orbit?

• 1:19:38Speaker 01

  Yeah, it starts at about 160 kilometers above the Earth, and then it goes to 2,000 kilometers above the Earth.

• 1:19:45Jason Calacanis?

  Wow. So it's gotta be hundreds.

• 1:19:50Speaker 01

  I'm gonna look it up in miles.

• 1:19:52Jason Calacanis?

  It's probably 1,000 miles.

• 1:19:53Speaker 01

  It's 1,143 miles.

• 1:19:57Jason Calacanis?

  Okay, so you have 1,000 miles.

• 1:19:59Speaker 01

  Of space going.

• 1:20:00Jason Calacanis?

  So let's just say it was the same as the Earth. Let's just take out the calculation, that low Earth orbit. Let's just say it's the same as the Earth. It's not 73% bigger, whatever. You have 1,000 times the surface area of the Earth, to put these things. 1,000 times. And we're not using the surface area. The current surface, data centers on the current surface area of Earth are what? 0.001%?

• 1:20:26Speaker 01

  Yeah, nothing.

• 1:20:26Jason Calacanis?

  Data centers probably take out like 0.00001% of Earth. This is why the whole debate over them is silly.

• 1:20:35Speaker 01

  Yeah, a negligible fraction of a percent is what AI says.

• 1:20:40Jason Calacanis?

  Yes. So even if we were at 1% of it, now you've got 1,000 times that ability, and you don't have anybody to complain, "Oh, you know, you're taking my energy, my water. It's, there's noise pollution. It's just, it's just up there." So it's the obvious end state. And additionally, you would think that at some point, if we master this and we can get things back and forth, you might start manufacturing stuff out there. Now, this sounds crazy, like, why would we put a factory up there and make cars and then ship the cars back down? That seems crazy, but if the cars are staying up there, we're eventually gonna have manufacturing in space.

• 1:21:16Speaker 01

  Exactly. It's like to, to, to fuel the factories and the mining and all the other things that are happening in space. You wouldn't need to send stuff back to, to the Earth.

• 1:21:25Jason Calacanis?

  All right, listen, this has been another amazing episode of This Week in Startups. Let's do a little promo here. Yeah, we have this new tactical and practical newsletter. If you email [email protected], [email protected], you email Lon and say, "I would like a free year of this newsletter." You can pay for it if you want to give us the 100 bucks, because we're trying to make it sustainable. But if you ask Lon by his email to give you a free one, and you make it to this point in the program, you just email Lon, [email protected], and it'll give you a free year.

• 1:21:53Speaker 01

  Please put that in the title of the email, so it's easy.

• 1:21:56Jason Calacanis?

  Subject line, "Free year."

• 1:21:57Speaker 01

  Yeah, "Free year of the newsletter." I'll, I'll find it, no problem.

• 1:22:00Jason Calacanis?

  All I ask is that if he gives, if he gives you the free year, you give us feedback. You write a comment or two. We're getting likes on these, but I would like more comments on them. And then you're gonna also be able to subscribe at x.com/twistartups.

• 1:22:12Speaker 01

  You can right now. It's happening.

• 1:22:13Jason Calacanis?

  Oh, is it back at $10 a month, $100 a year?

• 1:22:16Speaker 01

  I will check if it's been properly repriced, but they are being posted.

• 1:22:19Jason Calacanis?

  Have we started posting the articles there yet?

• 1:22:21Speaker 01

  Yeah, yeah. Every new one goes up on both X and Substack now.

• 1:22:25Jason Calacanis?

  Make them free on X for now.

• 1:22:27Speaker 01

  Oh, I can do that, too.

• 1:22:28Jason Calacanis?

  Yeah.

• 1:22:29Speaker 01

  I can do that, too.

• 1:22:29Jason Calacanis?

  Just make them free for now and see if we get engagement on them. And then what we can do is you can always change them to pay later, right? We can flip them, but for now, we'll do that.

• 1:22:40Speaker 01

  I could go back and retroactively flip them. I've been putting, as you can see, I've been putting them up for subscribers, but.

• 1:22:46Jason Calacanis?

  Have we gotten our first x.com subscriber yet?

• 1:22:48Speaker 01

  I don't, I don't believe we have. But I wanted to start filling in the archive.

• 1:22:53Jason Calacanis?

  But this will be great. We want to, like, literally get people the ability to subscribe there, subscribe here, and then I also want them to be able to subscribe on YouTube. So I want to set up YouTube subscriptions. So you're gonna be able to pay 100 bucks a year to be a part of the community, get the newsletter, get the show, because they get the show without ads first, before anybody, right?

• 1:23:12Speaker 01

  Well, because it's done, it's done earlier. So you are getting the, the show as soon as it's finished being cut, no ads. It goes out to the Substack folks first. And then, you know, we, the, the other people have to wait for us to edit and put in the ads and finish the show. And so it's a few hours.

• 1:23:28Jason Calacanis?

  I am obsessed with getting, uh, our 1,000 true fans. So if you're one of the 1,000 true fans, give us a hundy. I don't need the 100, but I want to put all that money towards a full-time writer to put these tips together and make it sustainable. And we created a job in the world just based on all of us, you know, paying 100 bucks a year. So please go buy a subscription to This Week in Startups Tactical Practical Newsletter. It's on Substack, twistartups.substack.com. You can search for it on there, and you'll just see this beautiful design we made with Claude, back and forth. And we will see you next time. Bye-bye.

• 1:24:06Speaker 01

  Bye, everybody.