I'm Kevin Roose, a tech columnist at the New York Times.
And I'm Casey Newton from Platformer. And this is Hard Fork.
This week, SpaceX acquires XAI. What does it mean for the AI race? Then Google lets Project Genie out of the bottle. Now, how do you rub it the right way? And finally, Moltbook founder Matt Schlicht is here to discuss building a social network for bots. Are there pokes on Moltbook, Evan? I'm not sure.
Casey, let's spend the first segment of today's show talking about our exes.
Well, all of my exes live in Texas. Is that who you mean?
No, I'm talking about SpaceX and XAI. There has been a lot of drama in the AI industry this week. We're going to talk a little bit later about some involving OpenAI and Nvidia. But now I want to start with the news this week that SpaceX is acquiring XAI. The deal was announced on Monday. These are both private companies, so the details are not public, but it's been reported by Bloomberg that this was an all-stock deal and that it may value the combined company at $1.25 trillion. This is of course in anticipation of SpaceX going public. That's expected to happen later this year. And if this is sounding familiar, it's because Elon Musk loves to do this. He loves to kind of bundle his companies up into mergers. Tesla in 2016 acquired another Elon Musk company, SolarCity. And just last year, XAI itself acquired X, the social media company formerly known as Twitter, in an all-stock deal. I'll just read a little segment from the memo. This was from Elon Musk. He said, quote, 'SpaceX has acquired XAI to form the most ambitious vertically integrated innovation engine on and off Earth with AI rockets, space-based internet, direct-to-mobile device communications, and the world's foremost real-time information and free speech platform. This marks not just the next chapter, but the next book in SpaceX and XAI's mission, scaling to make a sentient sun to understand the universe and extend the light of consciousness to the stars.' Casey, your reaction.
Ketamine is so powerful that it must be used carefully.
Talk to your doctor before using ketamine.
Now, this is a very grandiose statement from a man who is prone to grandiose statements. I think there is a less grandiose way of describing what we have just seen here, Kevin, which is that a very valuable and profitable company in SpaceX has acquired a cash furnace named XAI. Yes. So XAI, we know, is losing lots of money. They are spending lots of money building out data centers, training models. They are not anywhere close to profitable. SpaceX does generate revenue somewhere in the neighborhood of $15 billion last year according to reports. So basically some people are saying this is less of a merger and more of a bailout. This is a company that has a real business, that has real revenue coming in, in SpaceX acquiring something that is, as you said, a cash furnace.
Yeah. And it is kind of the second bailout of this kind that we have seen from Elon. The first one, of course, was when XAI acquired X for $33 billion. A pretty high price, I would argue, for what X was at the time. And now that combined company has now itself been bailed out by SpaceX. So, this has been amazing for all of the investors in those first two companies. And later this year, we may see them get another great exit when this whole thing goes public.
Yeah. And if you look at the reports, some of the SpaceX investors are a little bit nervous about this. They were like excited for the IPO. Now, some of them are saying, 'Well, you know, we trust Elon, but like we're not sure about bundling all this other stuff in with it, too.' But the vision, at least according to Elon Musk, is that this will all sort of be part of a unified, sort of vertically integrated strategy where you'll have, you know, X, the social media network, that'll be powered by Grok, the XAI chatbot, that will also run on data centers that are being put into space. And we should talk about that too because it appears that Elon Musk, like other companies we talked about, Google and its Project Suncatcher initiative a few months ago on the show, but he has gotten space data center plans. He wants to put the data centers in space.
Yeah, that's what he says. I imagine he will try to do it, but I think an important thing to say up top whenever you discuss Elon Musk is this man just loves to say things, right? Like there is no figure in public life, certainly not in the business world, who has made more promises over the years that have either not come true or have come true many, many years later. So, as we approach the discussion of Elon Musk's space dreams, I want to just note that upfront.
Yes, that's a good point. We don't know, you know, how quickly any of this will arrive. There are still a lot of good cautionary notes to strike on that.
You know, Casey, I almost forgot our disclosures. We're going to talk about AI. So, I should say that I work at The New York Times, which is suing OpenAI, Microsoft, and Perplexity for alleged copyright violations.
And my boyfriend works at Anthropic.
So, last week, SpaceX requested approval for putting 1 million solar-powered data center satellites into orbit from the FCC. And that is a ton. For context, the European Space Agency estimates that there are around 15,000 satellites currently in orbit. So this would be many times the number of currently existing satellites. And I think there are a few interesting angles here. One of them is this is obviously something that other tech companies are interested in too, but those companies do not have rocket companies attached to them, right? Google, if it wants to put TPUs into space, does not have a way to get there other than paying Elon Musk and SpaceX to send them up on one of their rockets. So, this could become kind of a land grab in space, if you will, where like Elon Musk may not entertain offers from other people to put their data centers into space because he wants to put his own there, too.
I certainly think it is a savvy story for Elon to be selling at a moment where he is preparing to take SpaceX public. He has just saddled it with this cash-burning company and he's going to have to convince investors that that will be worth it in the end. So, what does he say? Essentially, look, this is going to be the best AI company at infrastructure. AI demands will eventually exceed what we can provide for on Earth. The only way we will be able to fulfill demand is by putting data centers in space. And the combined SpaceX and XAI and maybe Tesla version of that will just be the winner of that race. So again, I think getting there is going to be an extremely shaky proposition, but I think if you're trying to make your S-1 look good before you go public, it's a smart thing to say.
Yeah. And I think it's not just about making the S-1 look good, because I think there's an argument that this actually makes the S-1 look worse, if you have this AI company attached to this, you know, money-making rocket company. But I think they're taking a calculated bet here that there is some sort of fixed amount of investor excitement about what's happening in AI and sort of whoever can get to an IPO first will get a big portion of that. So this is why you're seeing OpenAI and Anthropic kind of racing toward an IPO. I think this is why you're seeing SpaceX go out this year because there's this sense that like there's sort of one chance to be first in what we, you know, we'll see over the next few years, which is a bunch of these companies going public and they want to be the first there.
Yeah. And that makes sense to me. Although in a world where, you know, these companies continue to grow very quickly, I can imagine investors wanting to buy three or even four AI stocks. So, I'm not saying there isn't an advantage in getting there first, but I wonder how durable that advantage is going to be, you know, over the next 12 or 18 months.
Yeah, there are also a lot of hurdles and limitations to overcome before this idea of a space-based data center is even plausible. Right now it would be way too expensive and hard to do this. We don't even really know if it's physically possible yet. But there have been some experiments. For example, StarCloud, which is one of these other space-based data center companies, actually did train a very small large language model on a space data center in December. So, they have one of their sort of pilot satellites up there. They trained something called nanoGPT, which is like a very small version of a GPT-like chatbot, and they say that it worked. So, there's still lots of, you know, outstanding questions about whether this is possible. I would still put this very much in the category of experimental bet rather than sort of plan A. Their plan A, I think, for all these companies is to build these things on Earth. But if they can't do that for some reason, they want to go to space.
Yeah. And again, like just to inject another note of skepticism here, like think about how many times and how long ago Elon began promising that full self-driving would come to Tesla cars, right? And think about where we are in that journey now, which is like, well, it kind of drives itself in like these certain circumstances but not always, and there have certainly been an interesting number of fatal crashes, right? So that has sort of been that journey. And let me just say I do not expect the journey towards space-based data centers to go any more smoothly than the journey to full self-driving.
That's a good point, that's a good point. I think we should expect this to go more slowly than maybe the most optimistic people do. But this is at least part of the stated rationale that Musk gave for merging these two companies together, is that ultimately the future of AI is pointing toward needing to put at least some of the infrastructure into space and having what he calls the sentient sun, which is basically like capturing the solar energy and turning that into compute, is part of the reason for this rationale. Will investors buy that? I'm not sure. But I think that's where he wants to go.
You know, you also have a sentient son. He's lovely.
Here's what I'll say. If we're going to do this, I would like to actually see Elon go into space and build some of these, you know, data centers. I want to see him in like a spacesuit with a wrench doing his part. You know, they say a good boss, there's nothing that they would ask someone else to do that they wouldn't do themselves. So, I would love to see him up there.
I want to see Katy Perry up there with him. I want to see her fixing the high-bandwidth memory.
She has space experience.
Actually, when you look at like all of humanity, there are very few people who have been to space more than Katy Perry. So, I think it's a good idea.
All right, Casey, what effect, if any, do you think that being part of SpaceX will have on X, the social media network?
Well, my fear is that it is going to make them even less accountable than they have been so far. Right. So, as this acquisition is happening, X, I think arguably, is in the midst of a crisis. It's a crisis that we've talked about on the show. The Grok chatbot was used to create millions of sexualized images of women and children, and it is now under investigation by countries around the world. Just this week, French police raided X's offices in Paris as part of what has apparently been a very long investigation into potential crimes, including distributing CSAM, child sexual abuse material. And the UK's media regulator, Ofcom, has also opened an investigation. So, you know, I'm sure that those things are going to proceed. But my fear is that once X — it's already been swallowed up once in XAI — once that gets further swallowed up into SpaceX, now you start to get into some really weird issues where let's say that a country has hired SpaceX to put a bunch of satellites into space for strategic purposes. Now it's maybe going to be a little bit more reticent to go after them for these CSAM violations because they think, well, that's actually a really important relationship for us. We need to make sure that we have priority access so that we can get our satellites into space, and all of a sudden there's just way less accountability for X. So like these are very real concerns, right? Like look at what Ukraine has had to go through with Starlink in order to maintain its own access to internet connectivity, which is essential to its war with Russia. So these are not theoretical concerns, and you know this is one reason why if I were a United States regulator I might look actually very skeptically at this acquisition.
Yeah, I think that's right. I think there are certain ways in which like being part of a public company could force X to, I don't know, get its act together a little bit. Maybe there will be some additional disclosure requirements. We'll finally be able to figure out like how many people are using this thing, how much money is it making or losing? You know, what does the business of X look like? So, in that sense, I think it could be good for people trying to make sense of what the former Twitter has become. But I think you're right, it absolutely does give them tremendous leverage. They're going to have, you know, Starlink, they're going to have SpaceX, they're going to have XAI, they're going to have X, the social network. So, it's just like it is — if you thought Elon Musk was already too powerful and that was a thing that worried you, I think this is going to be a really negative development for you.
Absolutely. Now, there is one lightly comic note that I can inject into this otherwise, you know, dark aspect of the story, Kevin, which is that as part of the French investigation into X, Linda Yaccarino is being called back to appear.
And I always love it when they bring back characters from past seasons. Haven't heard that name in years.
That's right. The former CEO of X. She's now going to have to answer to the French government.
Okay. Well, I suspect that will go poorly. I think the other question I have here is how this affects XAI and their sort of strategy for building terrestrial data centers here on Earth. I've talked to some people at some of the big AI companies who are increasingly worried about XAI. I think they traditionally haven't been thought of as kind of a top-tier AI company. I still don't think Grok is sort of considered one of the best chatbots on the market, but they now have access to a lot of money through Elon Musk and then after this IPO through the public investors, and they are actually using that money. They just recently brought a big Colossus 2 data center online with a bunch of the newest Blackwell Nvidia chips, like 550,000 of them, which is more than anyone else has. And so there are people starting to look at how quickly they are building this data center and how many chips they're getting and thinking, well, if these guys can like make a good model, they could catch up quite quickly. So, you know, you can't just get to the frontier by outspending everyone else on chips, but it's an ingredient, and I think they have the money and they will have even more money after this IPO to do that.
This is the core of the story. Like this is how I understand all of the events that we are talking about in this segment, that this is Elon using the core advantage he has in the AI race, which is access to the most money and the ability to combine many different companies together. Right now, as a result of this merger, the fact that XAI is burning billions of dollars doesn't matter because it can now suckle off the profits of SpaceX, and that is going to let Elon buy time, right? And so the whole bet is we may not be at the frontier right now. We may be a second-tier lab, but we have a lot of money and we have a lot of ability to do these financial engineering schemes. And so we are just going to bet that if we have enough time, we will be able to catch up. We will be able to either come up with a breakthrough of our own or just copy off somebody else's homework, and that is going to let us win. That in combination with this massive infrastructure that we're going to build, that is going to let us win.
Do you remember when Elon Musk signed an open letter calling for a six-month slowdown, a six-month pause in the training of large language models? Do you remember the long-ago year of like 2023 when that happened?
Another move at the time that did serve Elon's interest because it again allowed him to play for time, right? Like what Elon Musk needs is time. He needs time to build. He needs time to catch up. He needs time to link all of his companies together so that he has access to the capital that he needs. That is how he is playing this game.
Yes. What he really meant was everyone else slow down. I'm trying to build rockets and data centers over here.
Well, there was one more big story in the AI world this week, which is the drama that is unfolding involving OpenAI, Nvidia, and Oracle. So Casey, can I try to sort of give the summary view of what is going on here and then you tell me what I'm messing up?
Okay. So here is what has been going on. There was reporting in The Wall Street Journal late last week around this deal between Nvidia and OpenAI maybe being a little shakier than people had thought. So as a reminder, last September, Nvidia announced a plan to invest up to $100 billion in OpenAI. This was part of their sort of dealmaking sprint that happened last fall. And according to the journal, there were some in Nvidia who were growing a little more skeptical of this deal and were expressing doubts. According to the journal's sources, Jensen Huang, the CEO of Nvidia, has privately emphasized to industry associates in recent months that the original $100 billion agreement was non-binding and not finalized. He has also, according to the journal, privately criticized what he has described as a lack of discipline in OpenAI's business approach and expressed concerns about the competition it faces from the likes of Google and Anthropic. So that's sort of turn of the screw number one is, you know, the boys are fighting and Nvidia is getting maybe cold feet about this deal. Then a report comes out from Reuters alleging that OpenAI is not happy with some of Nvidia's latest chips, specifically their performance on inference, and they sort of noted in this report that OpenAI has also been doing deals with a bunch of other companies — AMD, Cerebras, Groq, which is another chip company — which Nvidia has now sort of licensed or pseudo-acquired. So these two reports seem like they're kind of related to each other, with some people from Nvidia's side saying, 'Hey, we have some questions about the strength of your business,' and OpenAI or at least people sort of leaking out that they may be unhappy with the latest chips out of Nvidia. Am I right so far?
You're doing a great job so far. This is a complicated one. We're navigating it beautifully.
Okay. So then the CEOs go into damage control mode, right? So Sam Altman says he goes on X and he says, 'We love working with Nvidia and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time. I don't get where all this insanity is coming from.' And then Jensen Huang of Nvidia says basically the same thing. He calls the reports of a rift between the companies nonsense and he said, 'I really love working with Sam.' So into all this steps Oracle, which is also involved in all of these data center deals. They're the ones building the data centers that the Nvidia chips go in that are used to train the OpenAI models. So there's sort of the third piece of this tropple here. And they put out this post on X that says, quote, 'The Nvidia-OpenAI deal has zero impact on our financial relationship with OpenAI. We remain highly confident in OpenAI's ability to raise funds and meet its commitment.' Basically, some of Oracle's investors have been worried that if Nvidia doesn't give OpenAI this $100 billion, then OpenAI won't have the money to pay for all the data centers that Oracle is putting up. Did I miss anything important?
You did miss one important thing, but I missed it, too. Until this week I talked to Berber Jin, the great reporter at The Wall Street Journal, who helped to walk me through some of the twists and turns that have been reported, but I think that some folks aren't bringing together entirely. Here is the crux of the problem between OpenAI and Nvidia as I understand it. In September, part of the deal was that Nvidia was going to lease its chips to OpenAI. This is basically unheard of for Nvidia to do. Nvidia wants to sell you its chips. Chips are depreciating assets. It wants to get them off of its balance sheet. If it leases the chips to OpenAI, those chips remain on the Nvidia balance sheet and they're just kind of a drag on profitability. And of course, there is the risk that maybe at some point OpenAI defaults and it can't pay for all the chips that it has leased. And in that case, Nvidia would have much rather have just sold them to somebody else because God knows it would have been able to find a buyer. So, this was a very unusual deal that had been made in September. And over the next few months, Kevin, we have seen investor confidence in the AI buildout begin to decline. You can see it in the public stock prices. Oracle stock price way down, I believe, at various points by about half. Investors are that concerned. CoreWeave, another one of these companies that, you know, lets AI labs get access to chip capacity.
A neocloud, if you will. That's a cloud from the Matrix. They have also seen their stock price dramatically decline. And according to Berber, this is what the executives at Nvidia start to get concerned about, and they think, we don't want to be on the hook to OpenAI for all of these leased chips. And so that is the part of the deal that we are going to step away from. We are not going to do this leasing thing with them. We are just going to go back to pure sales. Now it is true that Nvidia is still happy to take an equity stake in OpenAI. They believe in OpenAI. OpenAI is a good customer. They do think that there's a good chance it's going to be successful in the long run. But you may remember back last year, Kevin, do you remember when Sam Altman would say things like, 'Our plans are literally so crazy, we're gonna have to invent new financial instruments'? We were in the room when he said this at the OpenAI developer day. This leasing deal that I just described, I believe, was one of those crazy financial instruments. And Nvidia said, 'This too crazy.'
Huh. So, this is a lot of like financial engineering and a lot of sort of investor talk about these companies, but...
And I love talking about it because it's our core expertise.
Yes. No, but I think like to take a step back for a second, I think what's happening here is there are just various levels of comfort among these big companies with the idea of spending some huge portion of their resources on building out these data centers for AI. And I think some people have said this is sort of the problem with these circular deals, right? If you have all of these companies doing deals with each other and financing everything in new ways, then like if this deal between OpenAI and Nvidia were to sort of collapse or were to sort of become much smaller, then that affects Oracle, it affects all of the other partners in that ecosystem. And the worry is like you pull out one piece and like the whole thing comes down.
Yeah, absolutely. And I don't think we are there yet, but when I spoke with Berber, he said that one thing that this does seem to put at risk is OpenAI's plans to build its own first-party data centers, right? Like raising the money to do that. It's, I mean, it's staggeringly expensive. Like the chips alone to fill one of these data centers costs in the tens of billions of dollars. Just the chips. You bought a graphics card in the '90s, it was what, $400? Now imagine you spent $35 billion on chips.
That seems likely to make OpenAI make deals with other cloud providers, right?
But all of this just speaks to, you know, once again Sam Altman having a very ambitious vision and then maybe struggling a little bit to get from A to B. So, do you think this beef between OpenAI and Nvidia that both CEOs are denying is a beef exists?
So, I don't know how individual OpenAI executives feel about the fact that this leasing deal is going away. I imagine they would be unhappy. It seems like that deal was quite favorable to them and the fact that it doesn't exist anymore is going to create some problems for them. At the same time, they do not want to poke Nvidia in the eye and say, 'You suck.' They want to take the $20 billion that Nvidia just pledged to their next fundraising round because they desperately need that money. So, this is all going to get smoothed over because ultimately like there are still plenty of like win-win situations for these two players. But I do think you can expect to see as long as investor confidence in these infrastructure plans remains shaky, Nvidia is not going to be making up any crazy new financial instruments, right? They're in the chip-selling business. They presumably want to stay in the chip-selling business.
All right. So, that's the big drama in the AI world this week. Casey, do these stories seem related to you in any way? The XAI-SpaceX merger, and the OpenAI-Nvidia-Oracle drama?
Yes, they are related. They are about the lengths that you have to go to in order to win the AI race and what are the different advantages that the companies have, right? So Elon has the advantage of being able to do a lot of financial engineering and use his very rich profitable company to subsidize his money-losing company. OpenAI does not have that advantage. What they have is the world's greatest fundraiser and maybe one of the world's greatest storytellers in Sam Altman, and he can get out there and he can, you know, persuade Nvidia briefly to invent a new financial instrument to try to finance his dreams. That happened to fall apart this week, but I'm sure he's going to continue to try to press that advantage as they try to win.
Yeah, it's going to get so weird. It's going to get so weird. The numbers are already getting so big that I'm just like I've sort of become desensitized to them. I'm like, 'Oh, like 15 billion.' Like it's a banana, Michael. How much could it cost? In 2024, I went to Anthropic to interview somebody and on their laptop there was a sticker that said, 'This is the last normal year.' And I actually think that sticker might have been right.
Yes. All right. Today, we're talking about something called Project Genie. This is a new tool that was released by Google last week that lets users create interactive worlds that they can navigate around like a video game.
So technically, Casey, they're not even calling it a tool. They say it's an experimental research prototype. So please don your lab coat before using Project Genie.
Yes, that is Google speak for this is very expensive and we don't want you to use it a lot. So this is something that has come out of Google's Genie 3 model, which is the underlying model that powers Project Genie. They showed that off back in August, but then they sort of spent the next few months building a product around that so that now you can actually go in and use this yourself. Right now access to Project Genie is still only for people who pay for the Google AI Ultra plan, which is their super high-end $250-a-month plan, and it's only in the US and only for over 18. So this is very limited because I assume it's very compute-intensive and expensive to run, but they did give us both access. We were able to play around a little bit.
Yeah. And one reason why I think that Project Genie is really interesting to talk about, Casey, is that it is just a different flavor of AI than we normally talk about around here. You know, so often we talk about these pure large language model, these chatbot interactions. Genie is something different. It's based around a world model, and that is an idea that is getting a lot of attention in the AI community these days.
Yeah, it's been a big topic of discussion in the last couple of months. I would say from people who are skeptical that pure language models can reach all the way to AGI, right? There are people like Yann LeCun, formerly of Meta, Fei-Fei Li, who's another big name in the AI industry, who have now started these sort of world model companies, and they are people who are saying, well, wait a minute, being an artificial general intelligence is not just about talking or even solving problems or writing code. You also have to have a sense of physics and the physical world if you want to do something like robotics. A lot of people are betting that these world model systems will help with that because you'll be able to sort of give a robot an understanding of the physical world around them. And so while it may look like something that is primarily for video games, I think the folks at Google believe that this is sort of part of their overall push to make these systems as smart and capable as possible.
Yeah. And if you have been on social media over the past few days, you may have seen some clips of Project Genie going around. I found several of them that I really like. Casey, I wonder if you saw any of these. Somebody put together a pretty good replica of a 1990s Blockbuster video store that you could walk around. I saw somebody who had recreated the crucifixion using a prompt. And my very favorite, somebody had made what basically looked like an Inception video game where you had sort of, you know, people flying sideways through the doors in the same way that they did during the Christopher Nolan movie, except you could actually, you know, control the character using your keyboard. So, some very fun things people are making.
Yes, these examples got a lot of attention, including from investors. And following this announcement of Project Genie last week, several big video game companies saw their stock prices fall. Take-Two Interactive dropped more than 7% and is down more than 10% so far this week. Roblox dropped more than 10% and is still down. Unity, another big gaming company, also saw its stock fall more than 20%. So people are saying, 'Wait a minute. If you can just generate video games like this with a text prompt, why would you pay for a video game like Fortnite or Grand Theft Auto or something if you can sort of make your own version of that?' Now, I don't know that that's like realistic. I think these investors may be just a little panicky. But I think the implications of this have been sort of most heavily felt in the video game industry since Project Genie, at least what it looks like today, has the closest resemblance to a video game.
Yeah, what I would say is that while I agree, Casey, we're a long way from a
tax prompt to create a full video game. If you use Project Genie, it's not clear how it's making a video game company more valuable, right? So, let's actually show people what it looks like. So the actual Project Genie interface basically looks a lot like a chatbot except it has two input boxes, one for environment and one for character. So you basically give it some ideas about what you want the landscape in your sort of video game to look like and then you describe your character whether it's a person or an animal or an object. You can pick first person or third person. So do you actually see the character or are you just sort of seeing the world through their eyes? And then you hit a little button that says create sketch. And Nano Banana Pro, which is one of the Google image generation models, goes out and makes a just two-dimensional sketch of what this world looks like. You can add some notes or modify it. And then once you're ready, you hit the button to generate your world. And it actually goes off for a few minutes and renders this thing that you can then walk around in. You can move around or move your view. You can jump using the space bar. You can kind of navigate this. And all this lasts for 60 seconds and then it goes away.
Yeah, you have. It's funny because when you're actually creating the world, they basically tell you like don't go anywhere because as soon as they start showing you the demo, a counter starts ticking down from 60 seconds, which like I've never seen a tech product like this before. And it just tells me that the cost of whatever this is is absolutely staggering.
They are incinerating TPUs over there. They can't even wait for you to get back from the bathroom. They can either show it to you right now or not at all. And that's the deal.
Yeah. I heard a rumor. I'm not sure if it's true or not, but that each of these generations requires at least four TPUs per user. Oh my god. Like they are, it's very compute intensive. Obviously, you know, this is an experiment and we should say it's like still even with all that compute, it is low frame rate. Like it's not a sort of very responsive sort of studio quality game.
Yeah. Maybe we'll get into some of my criticisms after we see some examples. But I would also say that over on X, a user there named Andrew Curran had posted side-by-side videos of what looked like the same prompt for Genie 2 back in late 2024 and Genie 3 eight months later and the difference was pretty amazing because the video was a lot shorter. It was a lot less interactive. The resolution was lower. So within a year this got radically better and whenever you see that sort of exponential improvement I think you want to start paying attention.
Totally. So Casey, what have you been making with Project Genie? Okay. Well, so there's one that I am proud of that I would like to show you. And I like to call it podcast microphone escaping the studio. Can we pull that one up? Okay. So, I gave a very limited prompt on this. I basically said for the world like create a Hard Fork podcast studio and for the microphone I said create a microphone with googly eyes.
And maybe I said something about you know he needs to escape the studio. And so I'll just show you what happened next. I will say this is actually a nicer studio than our actual studio. Absolutely. It's a hundred times nicer than our actual studio. And you will notice that I seem to have some struggle kind of manipulating my character. I was still sort of getting the hang of the keyboard controls. Also, I think you'll see as this clip goes on, it just kind of starts to feel more and more like a horror movie because it turns out there is no escape from the podcast studio. But I'll just show you what I mean. Here we go.
Okay. So, the microphone is trying to leave the studio. It's looking around. Great Hard Fork sign on the wall. And now it's going to jump off the table. Oh, it jumped off the table.
