Sam Altman's AI Optimism, New Presidential AI Policy & Defense, Endless AI Video Games & More AI News
40:26 Share

OpenAI C E O sam altman claims that A I development .

is speeding up, while billionaire CS are actually warning that .

is slowing down. So is we an A I bom or an a Kevin?

I give you two big boobs bomb.

I only vegan understand that reference, and I want you to stop also A, I is at the center of a national security conversation.

And just as important, future of video game.

Okay, i'm saying doom。 Now we're doom, doom.

This is a ee for humans, everybody.

Okay, everybody, there are some very big stories as we can. The world of A I first inform us we are going to be talking a little bit how the election will influence the air world going forward. But before we get into that opening.

before we get into that, I don't know what your politics are, just if you're listening to this, please get out there and vote. Everybody can make a change. I was in line early this morning, Gavin. People wrapped around the block. I'm saying they were only there for the hot dog and soda combo that I had to be a member to get inside.

Ridiculous about the costco guys.

Still I talk. I was in line of a costa.

this one I bit. I have a little video said in london last week, at in a death day can take less of this.

I think in five years, IT looks like we have an unbelievably rapid rate of improvement in technology itself. You know, hope people are like, man, the A G I moment came, went, whatever they like, the th Epace o f p rogresses, like totally crazy, or discovering all this new stuff, both. But A I research also about all of the rest of science.

And that feels like if we could sit here now and look at IT IT would seem like they should be very crazy. And then the second part of the prediction is that society itself actually changes, surprising the little. Yeah, we got to the point where computers passed the turing test.

And you would imagine five years ago, that when that happened to be dogs walking up right in the streets speaking french, and the everything would be up ended, society would be wild. But in fact, IT happened. In here we are. We're still playing the new season of diablo and trying to grind our hard core characters. Like.

not a lot really changed. No, no. Yeah, not a lot. But here we are. So I think that this is an interesting point of like am sees clearly that there is massive uh, technological advancements that are gona happen so going forward.

And yet he says that was clip obviously, that there is a sense that like maybe we won't notice IT as much. And I think to his point, the world of A I just kind of keeps getting bigger and bigger and is a small little story, but in video has now become the most valuable company in the world. The company that is power the A I boom, which we all know, picks and shovels, power the I boom cells.

picks and shovels. Oh, the gold rush is beneficial to them. What were the odds that? right?

So this is where we're at right now in kind of the AI hype cycle. He says something .

really like a couple of things that are impacted for that, right? If you believe that we have in fact satisfied the touring test and I I am assuming you do def, I think so.

I mean, I think that in the traditional sense, we definitely have .

yeah I think if most people were given a black box with the speaker in IT and IT was connected to open a eyes advanced voice model, they would think they are having a conversation with probably a human being on the other side of that, right? okay. So that's satisfied.

I would have said, wow, that would be a watershed moment for society of computers. Are that capable? Some things are gonna change, but I also think some things have changed.

We are living IT and breathing IT and reporting on IT each and every day. My life has changed a little bit. My job has changed a little bit.

I'm looking out on the horizon and seeing a lot of things changing. So when he casually says he is going to happen, artificial general intelligence, these machines are going to be capable. They're going to be agents that can go out and problems solve and work with each other and collaborate.

And he just says, and then that will go by and not much will change. Do you think that'll be the case? Do you think A G I will come and we will wake up the next day and just go like back to the fruit loops.

I think what he's talking about here is how we as humans process change. And I think that's what he's getting up more than anything else. I think that we have an unbelievable capacity to process change.

I keep thinking about this thing, my phone and and what that looks like now and what I can do and how i'm pissed off that if I can get the right A I uh you know gmail announcement to me where I know two thousand and seven, two thousand eight, the fact that like, I can watch almost any show of any given time on my phone and it's a perfect connection to the internet is so crazy to think about. So I think what he's getting out here is that we will process this in a way that doesn't feel like it's like, oh, my god, is the crazy thing i've ever seen. But I do think he's saying very specifically that this agi sort of world is coming, but that we may not feel IT as much as he thinks in that same.

We're interesting. Well, i'm going aside with the other billionaire right now, Gavin, you are saying it's surprising. I know.

I mean, i'm just do IT for the sake of this podcast. Obviously, i'm kind of pot committed to the idea of A I being the next big everything. But let's this just for the fun of this little audible scenario that we're walking through. Let's say I side with market and Jason and ben horowitz, Gavin, let's say that these billionaire s who believe everything is slowing down and grinding to a halt, that we are extracting less value, that the rate of change is absolutely becoming just pure dog water. And my paraphrase your person, you're periphrases.

Let's let's you where they said, and I think it's important to get in to there is .

a very interesting aim, total kind of thing. It's happening right now um where you know two years ago there was one you know L M that was like way out ahead of everybody else, which was open the eyes and sitting here today and there's like six um that are like on par with that. And interestingly, at least for right now, they're all a some coding at the at sort of the same point.

They're to get the the ceiling um on captain ties. Now know there's lots of smart people industry working to break through the ceilings. But you know sitting here today, if you just looked at the data, if you just look at the charts of performance, uh, over time, what you would say is there's there's at least of local topping out capabilities is happening, right? You look at like the improvement from GPT uh, two point out to GPT three to three point five and then compare that from like three point five to four. You know, we've really slow down in terms of the amount of improvement .

I do want to say I enjoy listening to that podcasts, a part casses on youtube, the band mark.

Sure, whatever those are, the two guys out of the backers of sixteen, they always are interesting, right? But what I found anything about this particularly think Kevin, is that like what was IT mid last year, where mark and reason came out and had his big manifesto, the technical optimist manifesto, and really talked about the idea of like A I can lead us to this next big generation. IT just feels weird.

And IT feels like a weird pivot to me a little bit to hear them talk about this stuff in. I IT doesn't ring true to me from what i've seen things change. Um but I don't know what is your initial thought on that. And I think one of think we should talk about is like what's been going out of anthropic because I think that kind of might lead into some of this stuff one hundred percent.

I think that's causing a lot of this signal right now. Some people have said, obviously, scale as all you need that we've solved IT, and it's just a matter of bigger data sets, more compute. There were also others that claim that for us to really get there, we need another breakthrough.

Maybe we're in that in between, Gavin, and where we know that we can extract a bit more performance out of these things, going to take a lot more energy, a lot more power, a lot more compute. We're going to get a little bit out. But we still are a breakthrough away, whether that comes from A I or somebody in their dorm right now hacking away some .

new or your boyfriend, you know who i'm talking about, john car mac, john carr max out there in his little shack. He doesn't .

to put a label on our relationship. He's you be ready to commit a very one sided or c but you mentioned anthropic, which had a massive release this week, which is not signaling amazing things for the company. And we should walk through that because cloud high school three point five is out. The new version of IT, the benchMarks went up a little bit across the board and the Price went up a lot of bit.

And that is something we have to talk about because IT is the first time in the AI space where a company has said, hey, the models gotten Better, so we're going to charge more for IT. And I think that points out to some of this stuff that like benchmark might beginning to this the conversation, which is at some point you have to pay the paper, right? You have to figure out where the money is going to come from.

And I think anthropic is in a little bit of a risk of being A A second in a race that is running very fast in burning a ton of cash. And i'd be curious to know, kind of like where IT goes. There's been a lot of weird rumors in the A I space.

And this is completely unfounded that for some reason they didn't opus run for the next generation of of claude opus, which is the premier frontier model. And then I didn't work in some way. Now, again, there is no proof out there.

There are people saying this is not real. But the other ideas, except the new clauses on IT, which is, again, claudes on IT three point five parenthesis new if if you're going to look for what the model name is, is very good in its improved. And we talked about computer use on this show and how good IT is, but maybe they are hitting some sort of like upper peak of what this is. And I think that's where the next like three months to say nine months to a year of A I is gna look at is like can we push through that? Does opening I have a new front to your model that's .

going to be commonly in one sense, mark and ben are right when they say that we're asm coding. That everybody else is catching up. All these models are getting to the same point, the same capabilities. But IT also in some ways ignores the fact that oh, won a side a lot of what we've seen from OpenAI, a stuff that was delivered months, potentially even a year ago, they had quite a lead. So do you think their next foundation or something behind the current will be a massive eyes open or govern?

So here's a theory i've ve been having for a while and having sit in the pocket. Ia, but my theory is part, it's us that is the problem because when we go from GPT two, no, I mean, us is as people, people in people, humans, not just we are part of the solution. Kevin is so clear already that we are part of the solution.

I want me sure this is your part of the solution anyway. So going from two to, say, three GPT two to three, if you're not layer GPT two was like an early breakthrough. But when you read the sentences that would come out of IT IT was kind like broken and are almost a like a presolar some way in the GPT three was this really interesting kind of step.

And for us, IT was like watching a kid go from a pressure or to like a grade school, or you're like, that is a improvement. And then you go from three to three, five and four, and you go like, all my god, I went from a grade school or to a high school, or this is where we come in. Oh, one. I don't think I have pushed all one where I can go, right? Because I am not a physicist, I am not research .

theory on the physics, what exists within a cork.

I tried to use IT on my classic, a benchmark, the game n benchmark, which is fighting sixty four bees against each other, to see which one wins. And I gave me, basically he result. The gp t four, did I thought about IT for two seconds and spit IT out? So I think this is what i'm saying is it's really hard for us to now see the further advances.

And this is why agencia I the idea that A I will be out to go to stuff for you is interesting because instead of IT being that the AI is getting smarter, if the AI gets more capable, that feels like I could be a step in the right direction on Kevin. Now we have to talk about kind of the elephant in the room that he just happened here, which is we entered into a different world now, right in the world, at least in america, with the election that went down. And I think there's a lot to be said about what this new administration will look for in this world .

not too long ago, the tightens of these industries were begging for regulation, and many assumed that this was to stop competitors from popping up overnight from training their own models. Now some are saying, no, this is great because there aren't gonna be regulations.

In fact, there's been a signaling that the evite and administration who said that models above a certain size or as a certain capability will need to be regulated in report to the government and have guard rails put in that a that that's gonna rolled back. So let's talk, boom, mr. Dom Gavin, do we think that the of the political fallout from the events of this week are going to be a boom .

m or doom for A I? Well, I think my first and foremost most take is that you may have seen this all. But like sam altman congratulated the new president, a bunch of these tech companies, ata IT, exactly.

They're all kind of kissing the ring now, in part because I think they know that this is the next step in trying to figure out how to work with in this world. I just hope that this administration takes this very seriously. Now, you know, the argument of republican versus democrat is going to be always like free market versus governments in some ways, but at the same time, this now directly connects the national security. In fact, there is a fascinating story that came out recently that met as lama was basically forked by china to create their own A I. So the open source A I .

and I got ta get in china's gonna a big troubles on this one again, because that is against their terms of service. IT says specifically that with lama, you're not allowed to use IT for military purposes. You can't in china, violated that. And I, the slap on the wrist is gonna be felt around the world. I guarantee IT this.

We just got very kind of monkey. And IT is an important thing, everybody to understand what is the big deal, open source say, I means the idea that, like the code is out there, people can use IT that can fork and and and do whatever they want. With that, mark zr work company matter made lama is not as powerful as the most advanced anthropic and open eye models, but IT was open source means that anybody can go and build on IT. They can do anything with they want. In china, a government was obviously able to get a hold of IT and kind of build some of their national defense stuff into IT.

I hope the reaction to this isn't to jerk the wheel and say, well, clearly open source bad, that only a handful will have access to this technology, and theyll be the ones who decide how things are aligned in guard rail. To be sincere for half, as I can be, meta wala licensing terms prohibit the use of models for, quote, military warfare, nuclear industries or applications, a and sb ono's. So they've come out and said, hey, the people's liberation army, this is an unauthorized used again, that's a bit of a finger wag. That's about all they can do, I think in this O T.

they have. And to show really that this .

is the most they can do, they then turn around and said, OK, guess what? U. S. Government go head in islam, that's fine. And they have basically turned about as well and said, like, if they're gonna do IT over there, we gotta do IT over here. So those of the two schools of thought, right, one is you ve got a, you got to clamp down and say, no open source because it's dangerous. And the other is, say, literally, weapons free, go head and it's out, do whatever you want.

And Alice wang, the founder of scale A I just announced defense lama, the LLM purpose built for american national security. This is a product of collaboration between meta scale and defense experts and is now available for integration into U. S.

Defense system. So we've telling people for a while this story was gonna cross over into national politics and international politics. And guess what, we are here now, and you know, fingers cross. This story goes well. But we're not too many far steps .

away from spaces. And that people, at some point, I think people accuse us of being a little bit when we said, look, genes out of the bottle, you're not going to rain IT back in. You can regulate IT here, but right over there, someone's going to do whatever they want with IT.

And I think we are just being real about IT. I'm not trying to shatter our hands, patting ourselves on the back here, but this was the viewpoint. So I think any effort to constrain IT is only going to get in the way of efforts here nationally.

I think so. And Kevin, I know what people can't constrain is their ability to subscribe to this podcast. Because everybody out there, you know, these are the most important times of our lives.

You know, these things matter. And what matters most of all is that you love us when we love you, and we do love you. So if you love us, please hit that subscribe, but and leave us review on apple podcast.

We ve got a very nice shut out from the podcast review this week. Thank you to those people. We love making this podcast. We bringing IT to you. But we can only do that when you share and you share your own .

love of IT as well, can't regulate your love for us. That's exactly so like subscribe. And if you want to throw five dollars in the patron tip jar will take that as well.

Also, we have a newsletter. We'll shot that out later. It's very good.

K, let's move onto what is another kind of divisive story here. The card A I and echt both introduced oasis, which is the world's first real time AI world model. What IT is is a general A I game engine in which you can explore minecraft, or a generated version of minecraft and actually play IT.

So in this game, again, you are not playing a game that was written code. You are essentially going around a minecraft world playing minecraft. But each frame of IT is being generated in real time based on your movement.

So if you kind of go and you walk towards something and you hit, you want to, like dig into the ground there that is not be rendered or or coated, sort away. It's pretty crazy. I didn't think we are gonna get to this sort of level of genre games this fast, but IT is a big deal. Once you get in there, you realize a first of all, it's hlubis ating over the place, which is literally what this is doing, right? Trying to hallucinate the next idea, make up what things you're going to do next.

The point where if you move the camera to the left a little bit right again, and the google back, IT has completely forgotten whatever IT is you were looking at.

which is a lot of people who criticized this idea said, specifically, this is why generate vi games may have trouble. Is that, like, yes, you can generate stuff based other people playing. And we've talked about a dome version of this that existed not that long, goes well, but that if you can't remember stuff or the game does not remember what you did, and it's not really a game.

it's more of like a simulate world tims we need founder and CEO epic games replied to the announcement of oasis and said, quote, wow, every single thing about this sucks.

Which is a pretty big statement from the founder and CEO of one of the biggest video game companies in the world. Now have a lot of people. I tried to discern what he's getting at here, but this does really bring forth some of that divide in the world. The video games, right? Because.

sure, someone complained that, oh, the game with the lines and the bot, that's not tennis. This poem is blurry and it's slow and there's not even any good music. This game sucks. I am not a hundred .

percent against what tim is saying here because IT does suck slightly. I don't mean that in a way we're like i'm hating on this thing entirely. But the question is, is that is .

tim has said everything about this sucks.

not everything, not everything I will say not everything .

about this suck, but what doesn't suck about IT, it's not like they're shipping this on steam and saying sixty dollars. Go play A I minecraft they're saying, like, look, this is the first step and we criticized people in the past for like being met at A I imagery because horse wearing a hat look like a Brown glory blog.

You know you can even do hands Brown that yeah what this statement sort of feels like when you say every single thing about this sucks. You are also saying that technology socks as well. So IT is a it's a pretty broad and sweeping statement. But you know, is IT ready for prime time? No, I don't think anybody saying that is the thing .

that I think is really more interesting about this is what would IT look like for this kind of technology to be layer onto a game and you like. In fact, fortnight, particularly tim's engine, is built with the idea that I can be completely changed by skin IT can be completely change the u unreal 4IH creative engine and all that stuff you can build into IT what wont look like for a generation world to be kind of placed in an engine like that？

I guess if I say a prompted game sort of thing, and the dream of this is like, I want to play, I to play a wild west game, where is me versus jeff baths? And jeff BIOS owns s like some sort of polar bear company, and I own a penguin company, were each other using polar birds in pending, right? Like that's just the perfect version of what a random game look like.

I feel like we are quite away, away from missing that to a an, A I engine. And suddenly the A I can just spin up that idea and haven't be fun. Guide from squair, who invested in this game, responded to him to and said that come back to me in five years will talk again. So in his mind that take a five year window and maybe five years is the time where we get to the place where a game can be prompted by a person.

There's going to be a transition where traditional game is like unreal or unity, are going to pile A I features on top of them that will let you style, transfer your game and at all sorts of effects or whatever. I think about those grand of total videos where people are piping footage of old GTA three and four into runway and making your super amazing and as gorgeous as yeah, of course yeah.

Let's let's get that with Better consistency and it'll be a huge value at we're gonna see A I bult IT on to traditional engines in the near term and people will be OK ay with that. And then eventually we'll see, I think, a full AI engine. Where's mostly A I with a little bit of the existing engines that we know today, bolt IT on to that and that it'll slowly probably start to change switch over.

But the notion that everything about this sucks is an interesting bumper sticker and it's a cool tweet. And I got people interested. I actually don't believe that him thinks that that that about sucks so you can say that sucks because all they trained on someone else's game and it's approximating IT OK. I think the core techy might go like, oh, well, the technology is actually interesting what they're doing there, but this implementation of that sucks at the moment, but that's not nearly a gravy tweet.

no. And also it's interesting because that is that underlying there's a little bit that let I thing right, which is like something we've talked about the show a lot. IT does IT suck because it's an AI doing a lot of the stuff right now that people were doing and and does that mean that like that's what suck.

It's hard to know what he thinks sucks, right? Even we are not the suck. We are not the suck owners. We are not the ones who are talking about the suck in the scenario. I'm going to keep saying, sucks as much as I can keep out the sock police.

I wish you were the suck police. I think we .

now have .

to transition because IT changes each and every week. But it's time to take our caps off and wig them around and say, hey, eye for a holy eye. Would I .

for hollywood? I for humans and hollywood 哈利波。 I for hollywood.

I can feel a certain website on recommending our podcast, right?

We're punchy today. We're punchy.

It's all A I interesting very interesting way. Let's talk about A I disruption because we talked about IT briefly with the defense department. We've talked about IT now with games, but we have said that prompt to hollywood is going to happen within five years time.

We're about a year and change into that, and we are still on pace for that, I believe, Gavin. And every week, there's new announcements. We want to focus on the ones that are we think you're going to change hollywood and moving pictures industry, if you will.

One of those is by a chinese company. They've have a model called clean and IT is arguably one of the best video generation models out there. And now they have a new feature, which has been rolled out to a handful of influencers, a custom face models. So if you want to train yourself into their video model or UI, generate a person, and you can give IT multiple looks of that person now, you can get a consistent character and pilot them in your text to video generations.

Clean as a chinese company, as is minimus, which we discussed, how china and amErica kind of fighting back and forth. And clean does not have, like mini max, doesn't either the IP restrictions, so you can generate all sorts of interesting real life people. But I think the interesting thing right now is you mean this is rolled out a bunch influences that will go out wide to, I think, paid clean user.

Soon we have seen people use customers to make themselves and make their own videos in themselves. But you know what Kevin for from prompt to A I A A for hollywood, what this also allows people to do is put actors faces and have consistent actors. So say, are the rights to, uh, junk lad than damn john clad than dam is on my side? He wants to have an the iverson, I can make bloodsport seven polar bears versus penguins, and have john claude van damn lead an army of polar bears against jeff eos s penguins. And I can make that happen because now I have a facial, Laura, within emotion engine.

I guess, the saving Grace to stop someone from taking your lightness and making one of these video models is that IT does require several clipsed to train their engine, and they need to be of the principal person and different lighting conditions from different angles with no other faces around. But if there's enough media of you out fare in different lighting conditions with nobody else in the frame, conceivably someone can make IT.

And there in lies the rub. As you mentioned, this is a chinese company, which as of what we're able to see right now, isn't particularly respecting write holders. And let's talk about an american company now, Gavin, and we love runway is simple IT works. You can get some powerful results, and they just gave users control over the camera of their generation.

So runway genre has been one of the most interesting products and that they have been rolling out a lot of additional features to IT. The video they released was this really interesting video, where you can see one character with red hair, and they kind of goes in different directions to the right into a laugh and like kind of a buffer and below her. And again, this just adds to the feature set, right? And when we talk about prompt to hollywood, we're talking about giving the creator even more control.

And you know, camera control is something that like when you're making a film or you're making a show, is one of the most important things you can have, because then as a director or or a creator of this stuff, you can focus on a certain thing or you can move a camera and interaction you want to the idea of like an A I cinematographer is a pretty big deal. And I do want to point out, this is not A I deciding which way the camera goes. This is the A I tool allowing you the human cradle to decide which way you want to move the camera. And that is like an empowerment thing, right? And and I think that's .

why give us art is about making choices. And now when you have another choice to make with the tool, so is generated A I slop becoming art right before our very eyes.

Well, I can tell you, i'm gonna make that polar bears bing's movie like right after this, and I would have some really dramatic pushes and moves across the faces of jasus as I take him down. So we'll see how that goes. And finally, there is one more big, big hollywood stories. Cavan involves a Roberts zmcc s and tom hanses new movie here.

So here is a film that takes place in basically one location. There's almost this locked off room where you watch history take place and you watch, you know, tom hanks have a relationship, and you watch them go through the different decades. But because of that, they needed to rely on special effects to D.

H. And then actually age time hanks, that you can see him play all these versions of himself. And the lead kind of got buried here a little bit because I don't think the movie is doing particularly is by doing that.

Well, trunk fan over on ex has a good summary of everything that was being said, but makeup can take hours to put on C, G, I. Very expensive. Semco said that here couldn't have been made three years ago, that it's fifty million dollar budget, wouldn't have been able to afford the C.

G. I, if you were to go in by hand to do this. That are so they A, I, basically deep fake technology. And this is the beginning of exactly this, right? I think in in five years time, this will not be part of the press cycle in the news stories about how movies are made.

I just made a video couple ago where I have mentioned others of math. H hangs joint IT wasn't go nation, but IT was the polar express. Do you member the polar express? So in the polar express, tom hung's basically got more capped and played every single role.

In fact, I found a piece of video from that, which was in all the E, P, K. And all the press club, where IT showed tom hank's playing the kid and playing the conductor and playing all these people in that movie. If you go back and watch, IT is miserable in a lot of weird ways, right there.

There's a charming Christmas story in there, but the eyes of the kids look dead like IT looks like a bunch of kind of zombies. And this was like the big uncanny value argument was like we're never gona figure out how to make um these moke APP images look like real people. And then, you know loan behold, and the circus comes along.

Peter Jackson makes all of the rings. We're now at a place where, you know anos in the avengers movies, all of these things get Better and Better over time. And like this is just another exaggeration of that thing, like we're just in the technology cycle that I think is happening again. I move on to some of our favorite things we saw this week in A I C. What you did there.

Then the day you stop 半 烧。

Ethan molech, who is a great follow on ex or linked in, he is a professor and he does a lot of really interesting conversation about the I looks into a lot of A I papers, but also just does some funny I experiments. He had this thing where he invited the hair avatar into a zoom call, which is a big thing. Now you can use hand A I avatar s and bring them in tear zoom call to represent you, which I still think is a very developing idea.

But he brought in two agent advice, and they started to talk to each other. And can I might just be work playing as the tiny st clip of this because, like, IT gets really ridiculous about halfway through. Let's circle back to that last team lunch.

IT was such a blast, wasn't IT. I mean, who knew bob could devour three slices of pizza like that? What do you think we should do for the next one? Any ideas? Or should we just keep the momentum going with something fun? Maybe a little team building activity? Lets brainstorm.

Oh, hey, love the vibe. Everything's going great on my end, just keeping my ear to the ground for any juicy updates. So about that last project time, like the last magician, i'm still in.

And yes, let's keep that collaboration flowing. Teamwork makes the dream work. After all, i've had this nights for a while, and I said the show before words just me and a AI avatar ars, in a zoom meeting. I'm try to figure out what the hell I did to get into this place that is like something that keeps me up at night, because IT is gonna come, you know for sure, something a horrific broom.

Yeah, there's gonna be a transition period where a lot of folks are sending their AI avatars to represent them on. And you will be the sole human in the room looking around wondering why. And you'll spend up your own AI agent and you'll let them all go chat and you will be to still do a bullet point and that i'll come in as a slack and then your AI agent on slack, we've read IT and spit to another AI agent and role is onna kick back on the beach, Gavin, with our toes in the sand, playing A I hlubi ated minecraft as we melt slowly into each other.

And maybe having our pets will be floating around in spaces that same time. Maybe that pets .

and zero g fuf fer A I over on x has a several posts of dogs floating in zero og. These are images using the new red panda model out there, the refrain. V three, when you think about the the data that goes into training these models, right, like global of liquid zero g yeah, that exists. But I don't think there's a tone of data of li upside down hopelessly going from one module in a space station to the next.

Well, that's the fascinating thing about this to me is that when you look at that picture, I think that dog just looks like so kind of long plus, but not unhappy to be through the space station. What's crazy is like the physics of the hair in the fur of the dog look like it's it's in zero og. To me that part is like it's done that remarkably well.

I've done the world comment. I've been on the zero g plane. Yeah I, I, I was amazing, but I felt .

like that dog floating .

about like just hopelessly. And they tell you that when you hit geo gravity, you're going to fail about, you're going to try to get some some resistance or something and it's not going to work and you're going to feel silly and like, i'm not going to do that. And I immediately went for, like, golden retrievers, just like, help. Okay, I felt like that dog.

wow. That is crazy. So but I saw what might be my favorite A I video today, maybe other than stuff. Yes, I love this video to me.

This is one of those weird things where there was a great video that Cameron aleo these guys that made the pyrates like this black and y video of these giant jacked up guys that looked like they made the pyramid ds, and somebody saw that and decided to tell us why gordon ramsey actually made the pyramids. So if you're just less saying this, you're not watching IT. This is a jack gordon rams the in black and White building the pyramids, putting the giant bricks on his head, walking around.

What I love about this is it's like the promise of A I video, is that like somebody saw that other permit video and said, I can do one Better. I can go gorden ramma y in this and know gorden ground has become this like weird standard for AI video. And I think in part that has to do with those chinese models that we talked about before.

because so represented in those models that you can get good versions of him out. And once people realize that to the race.

he almost became the new will smith, right? You see a lot of videos of him, like in his kitchen doing stuff, or doing stuff on the stage where you laughing about the video.

You know, all of all of this sounds so ridiculous to me. I I tried that when we do our show. Take just a half step back and .

listen as if you .

have no context on this world b the world that weren't right now that you're like, this is the promise of A I video and on my screen it's a jack gordon razi running around, throwing blocks everywhere that you like. He almost became the new wl. Smith meetings baghni.

welcome to the, you're new, if you're new. This is our world. This is the world we live in. So something we've cover ai, in the shell. And early on there was a different star that thought he was going to be the face of A I. We really try to kind of get him out there.

And you know that our friend get theory and gathering is somebody who we've work with on the show, not in a real life, but the AI version of guy fery. And we've tried to kind of bring in forth, and we thought I might be worth interviewing and bringing guide ferry back to chat a little bit about what he feels that gordon ramsey, who is a, sensibly a competitive him, has leapfrogged him in the A I. world.

Gordon ramsey is the poor man's guy feat to everybody knows that no one knows S A.

I guy to I listen up, it's guy A I here and let me tell you about more ramsey fan boy AI away from torch in the whole kitchen flavor towns, feeling like a cold casual with all these quite few personalities. Where's the sil, the spice? Stick with me and 啊 i'll show you a uh I recipe uh， packed with a digital donkey sauce and coded half anio poppers. Buckle up because we're about to deep fry the whole Kevin.

We should never have brought the world's digital donkey sauce into the less kind of the world. That is our fault.

Now I think of the scene and billing tez, excEllent venture. I think IT was in the second one where they're doing the guitar from that united the world. Yeah I think that's gna be digital donkey says, I think people are going to look back to this moment in time. Go those two. They were on something.

Let's let's in that we platform them.

We might as well ask to follow up exactly .

what does he think he can do to get himself essentially pointed above, where gorden rams he can kind of para abla above. And then then you fall down later on, like the omelet.

Oh, you wanted know how I bring the heat back to the top. First off, I don't just serve up answers. I serve on with flames. Rami can teach his precision and refinement talk. I'm here to scorch the scene with insane combs like a five later deep fried A I burger with a drizzle of binary infused bacon greece status, i'll be the only A I shift dropping recipes no one dares to even dream .

of ja anio infused database .

tes aliza made entirely out of spicy RAM chips. I'm not just A I A I explosion.

What is a RAM chip?

Is that a thing? I don't know, know none of this.

I think I was like, he means like computer RAM chips .

RAM ah no random access.

I thought we were looking at at some sort of dangerous, the animal thing going on here. But no, no, that would make some sense. What I wanted know, before he goes away, I want him to do like a real, and I want to do like a calling out of orden, like A W W E rustling. I him, I him to get up and angry about this because he's gotten get a little bit more active in this space, like he's been lying back too long. A I guide fairy has to be involved and I want him to feel like what he's got to do to call out gorden rams.

alright, enough enough. Ramsay, you crossed crump. IT, listen, i'm coming for you and i'm bringing the whole deep fried double dip force of flavor town straight to your front door. You think you're the AI chef king, not you're barely aside dish in this main course while your playing delicate, pretentious code salads. I'm out here slinging slaver grenades .

dropping off the q voted with .

join your juice and digital dynamic you're all about refinement. I'm about blowing the circuits.

Oh, I want a matter of, I want to out of this building. I want him out of our home. I want him out. I don't believe he deserves to be a number .

through his Amber color goggles .

talking killed.

We had an interesting experience .

here today with I got theory. I necessarily think he deserves now to be at the top. I guess actually, we ask him one last question before we go.

I don't think we should. I think ash tag, goodbye guy, if you want to get IT trending, by all means, do IT. I'll see IT in the disco guy.

Good, never coming back. Go back. goodbye. We hardly know. You go to AI for humans dot show and sign up for our newsletter. bye.