It's Hard Out There For a Lobster
February 25, 202601:01:59

It's Hard Out There For a Lobster

This episode is sponsored by Airia. Get started today at ⁠⁠⁠⁠⁠⁠⁠⁠⁠airia.com⁠⁠⁠⁠⁠⁠⁠⁠⁠.

On this week’s AI Inside with Jason Howell and Jeff Jarvis, I unpack the viral “2028 Global Intelligence Crisis” memo, Anthropic’s claims of Claude distillation attacks, an OpenClaw inbox meltdown, Meta’s massive AMD chip bet, Samsung’s “Hey Plex” phones, Pomelli’s AI product shots, and Claude’s new Wall Street push.

Note: Time codes subject to change depending on dynamic ad insertion by the distributor.

Chapters:

0:00 - Start

0:02:55 - ⁠THE 2028 GLOBAL INTELLIGENCE CRISIS⁠

0:04:45⁠Viral Doomsday Report Lays Bare Wall Street’s Deep Anxiety About AI Future⁠

0:08:24⁠IBM is the latest AI casualty. Shares tank 13% on Anthropic programming language threat⁠

0:09:22⁠Cybersecurity stocks drop for a second day as new Anthropic tool fuels AI disruption fears⁠

0:20:00 - ⁠Anthropic: Detecting and preventing distillation attacks⁠

0:24:19⁠American AI Industry Trembles as Deepseek Prepares to Release New Model⁠

0:33:41⁠Meta Exec Learns the Hard Way That AI [Openclaw] Can Just Delete Your Stuff⁠

0:37:39⁠Google clamps down on Antigravity 'malicious usage', cutting off OpenClaw users in sweeping ToS enforcement move⁠

0:41:52⁠Jia Zhangke Creates AI Video With Seedance 2.0⁠

0:42:29 - ⁠The video (translation CC available)⁠

0:52:21⁠Facebook owner Meta to buy AI chips from AMD in deal worth up to $100 billion⁠

0:53:08⁠Nvidia’s Deal With Meta Signals a New Era in Computing Power⁠

0:54:13⁠Samsung is adding Perplexity to Galaxy AI⁠

0:55:22 - ⁠Google: Create studio-quality marketing assets with Photoshoot in Pomelli⁠

0:56:55 - ⁠Anthropic Links AI Agent With Tools for Investment Banking, HR⁠

00:00:00:10 - 00:00:08:25
Unknown
This episode of the AI Inside podcast is brought to you by area. Get started for free today@irca.com.

00:00:08:27 - 00:00:46:08
Unknown
Coming up next, Jeff Jarvis and I look at the viral 2028 Global Intelligence Crisis report by Citrine and Tropics claim that Chinese labs carried out distillation attacks on Claude, and how a famous Chinese director used Sea Dance 2.0 to make a short that I think really shows just how far generative video has come. That's up next on this episode of the AI Inside podcast.

00:00:46:10 - 00:01:04:25
Unknown
What's going on everybody? Welcome to another episode of AI Inside the Show, where we take a look at the AI that is layered throughout the world of technology. Usually we do it every single week. But last week we did take the week off. Primarily because I needed a break. I went to the mountains, I did, we snowboarded in Tahoe and it was absolutely amazing.

00:01:04:25 - 00:01:23:13
Unknown
So I'm here. Jason now joined as always by my friend back in his office, Jeff Jarvis in a real chair. Yeah, congratulations. I think I mean, it might not be as comfortable as the bed. Okay, okay. I'm not all. Oh, I bet it's wonderful. But not my body. Yeah, my my cloud of pillows bring me down right now.

00:01:23:14 - 00:01:42:11
Unknown
I'm getting there day by day. Well, during Covid, when I was working at Twitter and everything moved into, you know, the homes for a year and a half or however long that was, I blocked it at this point, I did much of my work in my office, which was on my bed upstairs, and actually over time that started not being comfortable.

00:01:42:13 - 00:01:57:17
Unknown
It's a weird way to sit, like all the time to do work, you know, it is. I can't figure out my kids still do it even as adults, and I can't figure out they just they just did it their whole lives. Just like, do your kids study? Do your kids study on bed? On the bed? No. You know, my daughter does go to the table.

00:01:57:18 - 00:02:15:04
Unknown
I mean, sometimes sometimes she will. But, you know, she'll go to the table in the dining room and she'll spread out there. So, yeah, it's not like our media go to is not the bed. So anyways, it's good to see you. I'm happy you're feeling better. Happy that you can do the show again. Yes. It is, pretty last minute.

00:02:15:06 - 00:02:32:04
Unknown
So apologies in advance, but, Well, not in advance, I guess is the point, that this is a day earlier than we normally record, so we don't have a live show. We don't have the recording coming out on Wednesdays. Could release one day early. We do have an interview, which you can look forward to. That's going to be, releasing this Saturday.

00:02:32:04 - 00:02:51:18
Unknown
So, you know, you might want to make sure that you have your feeds all fired up and, and warm and ready to go by that. Not that that's how I or our podcast feeds actually work. But anyways, but we do have a bunch of news talk about, so why don't we just dive right in, Jeff? And we're not rounding up the last two weeks of news.

00:02:51:18 - 00:03:16:25
Unknown
We're just going to focus on the last enough this week. Yes. Yeah, indeed. Let's talk a little bit about, I don't know, I sentiment talk a little bit about, fictional stories. There's a, there's a I don't know if it's a report or an essay, what you would call this, but it's on Substack, right? Yeah. Yeah. So, Trini, it's a, it's a, it's a speculative 7000 words scenario.

00:03:16:27 - 00:03:41:15
Unknown
Okay. A scenario. Okay. That's a, that's a word that I had thought to put it in. But that's exactly what it is. It's called the 2028 or the 2028 Global Intelligence Crisis. A thought exercise in financial history from the future. And in it, the report lays out what it calls a scenario, like you said, not a prediction. It imagines the near future.

00:03:41:15 - 00:04:12:04
Unknown
I think the date on this is, let me get rid of this little pop up here, June 30th, 2028. So you know, they've basically targeted a couple of years in the future where, where this explores what happens if I works like the optimists, expect or hope or anticipate and how that what that actually looks like or what that could look like, let's say, because this is a work of fiction, I don't even think that they're saying that this is what's going to happen.

00:04:12:10 - 00:04:33:13
Unknown
But they were. But my understanding is they really just kind of looked at this and said, all right, if all of these great things around the development of AI, did happen, what could that lead to? What? You know, what what are the the markers or the pathways that that points to and, and I mean, we can talk about that, but yeah, everybody seems to be reacting to this right now.

00:04:33:13 - 00:04:57:03
Unknown
Yeah. Well, so so the Wall Street Journal had a story. This is how I came across it, say, in viral Doomsday Report lays bare Wall Street's deep anxiety about AI future and supposedly in great measure because of this report air quotes because it's all fiction. The Dow took an 800 point drop on this, which only demonstrates that Wall Street's an idiot.

00:04:57:05 - 00:05:17:08
Unknown
So they continue doing this, continues. They do it, then they do it. It's just amazing. And so this, this, this, this report. And for those of you who are not on video, I'm doing air quotes. Reminds me very much of something we called AI hyphen 2020 7.com. Yes, about a year ago, where the same thing happened.

00:05:17:08 - 00:05:40:10
Unknown
It was a scenario of, it was both of these. Everything that can go wrong will go wrong. And we're going to go through every possible thing that's that's going to fall apart. And so I had, I read it was friggin forever. But I also will say that, until, tla I, I had, Gemini summarize it.

00:05:40:13 - 00:06:06:05
Unknown
So it says that there's, there's a few phases here. There's the boom, which is now then comes the ghost GDP phenomenon, where the on paper, the economy looks great and productivity and GDP or record highs. But output generated by machines doesn't circulate back into human economy because humans because machines don't buy lattes. Then the deflationary spiral as high earning white collar workers lose their jobs because.

00:06:06:05 - 00:06:29:28
Unknown
Because they are, in the words of of this air quotes report obsolete, then company. Then they lose their jobs. Consumer demand creators, companies respond by falling revenue, blah blah blah blah blah. Financial collapse comes in 2028. Unemployment hits 10.2%. Why? They add on the point 2%, I'm not sure. They've done their numbers. They've ran their numbers very specifically.

00:06:30:01 - 00:06:52:16
Unknown
Yeah. Hallucinating here. A $13 trillion mortgage market, collapses because nobody can afford their homes. And then we have a payment shift, which I agents bypass traditional banking and credit card networks in favor of nearest zero fiat stablecoins. Okay, so they threw everything possible they could, including crypto, into this scenario where NFT is in this equation.

00:06:52:16 - 00:07:12:18
Unknown
Seriously? Exactly, exactly. So I asked a Gemini to poke holes in this, so I'll quickly go through some hole poking. First is the lump of labor fallacy that there's a fixed amount of work to be done, and that if you eliminate some work, that work disappears. But we know from the Industrial Revolution, we know from the computer revolution.

00:07:12:22 - 00:07:45:08
Unknown
We know that when some jobs disappear, other jobs that are being created, we all we end up working. Second is underestimating the velocity of policy that if things were going this far south, even our screwed up government would step in. The Jevons. There's a Jevons I never knew paradox, which is often, called upon, which says that, that as a resource, as a resource becomes more efficient and cheaper, its total consumption increases.

00:07:45:10 - 00:08:00:17
Unknown
You think it's the opposite, then overestimating AI agency that it can really do what it says to do that there's a consumption flaw that suddenly we're going to stop consuming everything. This is a little bit specious because it says, well, the rich people will consume more because they have more money. That is the good. Not at all.

00:08:00:17 - 00:08:20:05
Unknown
Now and then the implementation lag. All right. So so that's it. It's it's a it's a it's a it's clickbait. It's a doomsday scenario put out there. It's 7000 damn words. It gets talked about all over the markets a fool. And that's where we go. But we're in we're in a tenuous time now where the market's feeling tetchy.

00:08:20:07 - 00:08:51:08
Unknown
Yeah, and the economy is feeling tetchy. So we have a few other examples here. IBM shares tanked 13% has hasn't gone down far that far in in a long time. Why. Because of anthropic and a programing language threat the COBOL. These are COBOL sorry not cobalt but yeah yeah. So so because it could be COBOL it can do these things that now IBM is dependent upon doing so 13%.

00:08:51:10 - 00:09:08:24
Unknown
I'm trying to see where it happened. It doesn't say how long it was. I was I think because cloud code could be used to automate the exploration and analysis work that drives most of the complexity in COBOL modernization, a key IBM business. I mean, this is this is a code system that was developed in the late 1950s.

00:09:09:01 - 00:09:28:17
Unknown
I wonder how many people are fluent enough in COBOL to do this. Anyways, I don't understand why this impacts I. Yeah, IBM is doing, you know, updating of these ancient things, but is it that much of their business? Is it is it worth that kind of drop? No. Then we have, cybersecurity, two stocks drop for a second day as a new entropic tool.

00:09:28:17 - 00:09:50:05
Unknown
Anthropic picks. The bad guy at all these stories fuels AI disruption fears. So that's going on. Sector by sector. You know, the RBS scenario goes through travel and other sectors. So sector by sector there's nerves now and then Wall Street chooses to say it's done. They're the next ones. Meanwhile the next story I no I'm rations rushing through this.

00:09:50:05 - 00:10:29:16
Unknown
But it's funny how all this came together. Next story is that Wall Street's latest bet is on Halo companies. That is to say, companies that are immune from AI. So other parts of Wall Street, while they're well, they're deserting, IBM they're buying, Deere and McDonald's because we're always going to be French fries, folks. Because because Big Macs can't be AI away, because you can't get an AI version, of a Big Mac, but you could get an AI driven, production and delivery system of Big Macs that replaces you could and John Deere is dependent upon AI now, but it's benefiting from it.

00:10:29:19 - 00:10:55:02
Unknown
So that's the next kind of dumb story we have. And then finally the I, I'm in my sick head. This all came together. So the Washington Post had a story based on a Goldman Sachs report that said, on the one hand, wow, I contributed to the economy in huge ways. Oh. That means the economy is in bad shape because I was taking up so much of it.

00:10:55:05 - 00:11:18:04
Unknown
The Goldman Sachs report said that in spite of this huge spending on AI and on, data centers and stuff, it really has had no impact on the core economy. And I think that's interesting to me, that's the worst news for AI, because if it's not really having a huge impact in measurable ways, then is it worth all this investment?

00:11:18:06 - 00:11:50:14
Unknown
But this too is kind of speculative. So that's our Wall Street roundup today of all of the fear panic. Sidestepping and, prevarication, going around around the impact of AI, it just it says something that a Substack post can move the market. Yeah. And what it maybe maybe you can clarify this for me. What is a report like?

00:11:50:17 - 00:12:10:12
Unknown
What is the difference between a report and a, a blog post or a research article or something like that? Like when I think of a report, I think of like we had, you know, a lot of researchers that, you know, that queued hundreds of people, if not thousands of people and collected data. I mean, it's it's just a deeper research sort of thing.

00:12:10:12 - 00:12:32:24
Unknown
I mean, this analysis is that all a report is, or so I'm looking it up now. It makes it feel more official, I guess, is what I'm saying than maybe what it is. Because if this was just a fictional story, why is it a report? Because the wrong report on there as as some journalistic, you know, places do it just makes it seem a lot more official and a lot more legitimate when I'm like, yeah, but it's a fictional story.

00:12:32:24 - 00:12:56:07
Unknown
Like, I don't know. So the Google tells me that that's a Trini research, which is a Substack newsletter. That's really what it is. Is written by James Van Gillan. The founder and primary author who started publishing macro and thematic stock research under that name in 2023. Right. So anybody can come along and say, I'm going to research stocks and I can publish it and fine, that's perfectly fine.

00:12:56:10 - 00:13:15:11
Unknown
You know, anybody can do it. This is the era of the of the yeah. The individual reporter or you know what I mean. There's a lot of independent happening right now. And I guess this is part of that. Right? So, Noah Smith says the Trading Post is just a scary bedtime story. I don't want to I don't want to diminish that.

00:13:15:11 - 00:13:35:22
Unknown
Yes. Some of these bad things will happen to some people in their jobs. It's happened totally. But, this is just amazing that this I mean, it's not as if it didn't even though 7000 words didn't say this could happen, but then that could happen. Well, but if this happens or didn't, it didn't say this bad thing happened, this good could happen.

00:13:35:28 - 00:14:07:03
Unknown
It just said every bad thing that can happen will happen. And I a pile it on continually until we show that by 2028, not far from now, over two years, everything is completely borked in the world. That's not how the world works. It's as if as if we, you know, yes, we're frogs in boiling water politically, but economically, Wall Street aside, there are a lot of people whose job it is to keep track of this and to watch out for these signals and to hedge them and, yeah.

00:14:07:03 - 00:14:24:14
Unknown
So this was just an I was just irritated as hell by this. And this is only two years in the future, which, I mean, you know, a lot of people would look at kind of where we've come in two years and I and be like, okay, well, that makes sense that all of this kitchen. How long in my lifetime have I heard that, like, flying cars is like ten years?

00:14:24:18 - 00:14:45:18
Unknown
I bet in the future, you know, and that continues to be the the kind of thing and, and, I don't know that all the research that I do on, on, on media and technology, the amount of time it took for my favorite prop, the, tripod vacuum tube, invented the amplifier on your side of your desk. Now again.

00:14:45:18 - 00:14:50:09
Unknown
So you can I can pull my props back, reach. Watch out.

00:14:50:12 - 00:15:11:13
Unknown
Gutenberg type isn't far behind. Fell on. Yeah. So the number of years it took from this invention to the full impact on culture and economy with broadcast radio was quite a while after Gutenberg. Sorry about this. It was a century and a half where we saw it before. We saw the invention of the modern novels around us, the essay with Montaigne and the newspaper.

00:15:11:15 - 00:15:30:24
Unknown
These kinds of developments take time before we understand what they are, what to do with them, what the impact is. Yes, they have huge impact. Look at steam, look at electricity, huge impact on society. But it's the hubris of the present tense to think that, oh, well, we'll we're going to do this in two years. We're going to understand everything in two years.

00:15:30:27 - 00:15:49:12
Unknown
It's ridiculous. And and, you know, the irony of this little report is, on the one hand, they're buying every single bit of B.S. that comes from the AI world. They're saying, oh, yes, it's going to eliminate all these jobs. It's going to change all these industries. It's going to have this unbelievable value. It's going to do all this.

00:15:49:15 - 00:16:12:11
Unknown
And then everything that happens out of that is bad. Nothing good happens. So anyway, I won't be hiring a trainee to do any, stock picking for me. Maybe they could. Maybe they could ghostwrite a novel for you. Yeah. A novel about the future of AI. I wonder how much of this is written by a fictional tale.

00:16:12:11 - 00:16:34:23
Unknown
Well, I mean, that's that's the question at 2026 right now. Have the have the systems just got good enough that we that we can't detect. Let's go to Claude. And the prompt is, tell me all the doom that you're going to cause Claude. Yeah, right. That you're capable of. Yes. That you're so amazingly capable of without the album for you.

00:16:34:25 - 00:16:56:16
Unknown
Yeah. It's just interesting how this keeps happening. And, you know, how the markets continue to react in such a way. And. Yeah, I'm sure we'll be talking about it more a little bit later on the show. Yeah, I think I can't, I couldn't find it. I thought the author of that AI 2027 thing kind of recanted, but but I couldn't find it.

00:16:56:18 - 00:17:15:11
Unknown
Oh, really? Yeah, I think kind of said, oh, yeah, maybe I was a little overboard. Maybe that was. Yeah, maybe we'll find. So set that in your in your calendar when these guys when the treaty says, okay. It's 2028, we, we, we were a little, a little pessimistic. Yeah. Yeah, totally. Well, thank you to our amazing patrons.

00:17:15:11 - 00:17:33:14
Unknown
Patreon.com slash I inside show you enable us to do the show each and every week. Except for last week, because I was in the snow. Let me just tell you, by the way, I, we chose the literally the best day to go to Tahoe, because the day that we left was the first day of this massive storm.

00:17:33:15 - 00:17:54:16
Unknown
The time that we were there pretty much didn't stop snowing. It was just, it was beautiful. It was gorgeous. I kept saying like, we won the snow lottery. So thank you, patrons, for enabling us to take those shot. Patreon.com slash I it's I'd show big thank you to our patrons of the week, Bradley rich wine, Brian Yeager and Bob Coburn.

00:17:54:19 - 00:18:01:02
Unknown
Just a few of our amazing patrons who enable us to do this show. Each and every week. Thank you so much for that.

00:18:01:02 - 00:18:17:27
Unknown
All right. Going to take a break and thank the sponsor of this episode. And it is area areas. Pretty awesome. We talked about genetic AI, a lot on the show. And this is what it's all about for area and especially if you if you're in the enterprise I'm using it as like a personal user.

00:18:17:27 - 00:18:41:25
Unknown
But there's many different ways that you can use area. And that's pretty awesome. Bring it AI into your company is exciting, but it can also feel risky. Between data privacy concerns, security threats, constantly evolving, compliance requirements, it's easy to hesitate. And that's where area comes in. Area provides a unified security layer with advanced threat detection and strong governance tools.

00:18:41:27 - 00:19:05:13
Unknown
It's designed to protect your AI ecosystem from day one so you can safeguard sensitive data, monitor activity, and stay aligned with compliance standards, all while continuing to innovate at the speed that you want to move. So instead of slowing down because of that uncertainty, you can actually move forward with clear guardrails. Real confidence. You don't have to choose between innovation and security.

00:19:05:13 - 00:19:25:01
Unknown
An area helps you do both at the same time. It really is a cool system. The more time I spend in there, the more my imagination runs wild with what I can build. And, kind of string these pieces together to create a workflow that really works. And that's just for one person think, let alone for like your enterprise, your organization.

00:19:25:03 - 00:19:47:27
Unknown
If risk has been the thing holding you back from going bigger with AI, well, this is your green light visit area.com to learn more or start a free trial. That's a I riaa.com. And yeah, we thank Aria for their support, their continued support of AI and side again. Couldn't do it without you. So thank you Arya for being here with us.

00:19:48:00 - 00:20:00:09
Unknown
All right. We're going to take a quick break, come back and talk a little bit about anthropic and distillation attacks. Is attacks the right word? That's one question that I want to answer. We're going to talk about that in a second.

00:20:00:11 - 00:20:44:24
Unknown
Okie doke. Let's see here a few stories here in the mid block. Anthropic reports that Chinese labs deep seek, moonshot AI and minimax have run what they're calling distillation attacks. And, they're saying that this is siphoning off Claude's capabilities into their own models. They're stealing our soul. They're stealing our stuff, man. Distillation attack, as it's called, is when someone uses a model to teach their own model by methods of, like, hammering at the API with a huge amount of of prompts and training, training their own models on the responses that they get back.

00:20:44:24 - 00:21:22:07
Unknown
So you can imagine you'd have to do a lot of this to get anything useful, right? It's a way of copying a model style, its capabilities, its behaviors. Through those responses and anthropic says this was done through around 16 million exchanges across, 24,000 fraudulent accounts. And so they're charging they're they're asserting that this was happening at a very, very large scale, and that all that traffic was routed through proxy networks in an effort to avoid detection, avoid network blockages that would prevent it, that sort of stuff.

00:21:22:09 - 00:21:39:23
Unknown
Yeah. Is attacks the right word? Because I don't know if when I think of attacks, I don't know. I don't know. These attacks is the right word here. Yeah, totally. Well I think they're trying to they're trying to open up the emotions here. But for sure they are. Let me go to my Claude here. Claude, can you define irony?

00:21:39:25 - 00:22:02:17
Unknown
Because anthropic, is settling a, a huge suit from authors who say that it stole from them. Stole attack? These are all kind of emotional loaded words, right? Yes. Yeah. And, Claude said, well, no, we just read learned, however, along the way, we took the books without buying them. So was that an attack?

00:22:02:17 - 00:22:25:02
Unknown
Was that a theft? Then they turn around and they're accusing deep seek at all of stealing from them. Well, there are tools out there in the open. I don't know, nothing wrong with creating various accounts where you. A lot of people create various accounts to test things. They learn from Claude. Is that not fair game?

00:22:25:04 - 00:22:52:11
Unknown
I mean, I do know that anthropic has some terms of service. You know, that they've that they have language about this. But again, that's to us as terms of service. I mean, anthropic is also saying this is a national security violation, you know, and of course other hide what they're kind of what they're kind of saying there is, you know, you're we're handing you're the Chinese, the services, all of our valuable information.

00:22:52:11 - 00:23:09:20
Unknown
They're going to be war more powerful than the US, you know, companies. And so it's a security risk or whatever. Yeah. I mean, it's what are you going to do? What are you going to do, cut off Chinese users from the American internet? Right. I don't think we want that, though. I don't think we want that.

00:23:09:24 - 00:23:33:09
Unknown
You know, anthropic is just such an amazing. They're they're amazing. They're a fascinating company. Because on the one hand, I don't trust their, their view of AI safety. And, and, you know, they're half, utopians thinking they're going to change the world with AGI and then half, dystopian is thinking this can ruin the world and so on.

00:23:33:11 - 00:23:57:02
Unknown
And, and they fit kind of into that. So treaty report, yes. We're going to create this future, and then that future is going to be, oh. And you better watch out for us. On the other hand, they're doing brilliant things. They've handled the code world amazingly. People are going gaga over what they're doing. They're moving the market because they they've create tools that that clearly people can use and have an impact.

00:23:57:04 - 00:24:17:17
Unknown
And then they come back around and they do things like this, which, yeah, I can understand they're not happy about it, but do do they read the room? Do they have a sense of how they're perceived? It's a really it's a fascinating company because on the one hand, I think they're doing really great work. On the other hand, I think they don't listen to themselves sometimes.

00:24:17:20 - 00:24:38:26
Unknown
Yeah. Yeah that's true. Meanwhile Deep Seek is is readying the release of its V4 model. Oh. I wonder if its V4 model is going to be a lot like Lod. But I mean that's that's interesting. Just because a year ago, about a year ago. Right. The deep skinning of AI happened. I don't know if that was the word, but I like that.

00:24:38:29 - 00:25:05:06
Unknown
And it really kind of showed how these models could be created differently, how they could how maybe massive, you know, gargantuan sized models wasn't absolutely necessary to get these, these large gains in performance. Yeah. So, yeah, the futurism headline, Futurism has let them use the American AI industry trembles as deep sleep prepares to release new model.

00:25:05:08 - 00:25:26:04
Unknown
So, I think the expectation is they're going to wow was again. We'll see. Yeah. Do we think that it's do we think that what anthropic is charging here is like like and they're and they are calling out deep sea that we would see any of those gains in the next model of deep sea. But here's the thing.

00:25:26:06 - 00:25:51:23
Unknown
There was a, there was we even know there was a Benedict Evans essay this week, which was too long to go into, but it was it was looking at AI's OpenAI's weaknesses, basically. And one of them is that all these models are becoming commodities. Yeah. And so it's true across OpenAI and then through Epic and Google, and deep seek and company, they're they're just we talk I use the word all the time in the show.

00:25:51:23 - 00:26:15:06
Unknown
We leapfrog each other, tiny leaps. So they clearly are watching each other. They're clearly copying each other's functionality. It's just like, iPhones and Android. No different. And at some point they all become commodified. That's what they should be worried about. Not that deep seek is stealing for them. What are you doing that's unique. And what then went on, you know where that unique is?

00:26:15:06 - 00:26:42:00
Unknown
Lie. It could lie with. You have data. Nobody else has. Well, that's probably Google. It could be a finance company. It could be a few others. Do you, I forget all of those points here, but do you, do you have a, scalability with consumer use? Well, Google, Apple, Microsoft. But not anthropic and not, others.

00:26:42:02 - 00:26:48:19
Unknown
And so it'll be interesting to see,

00:26:48:21 - 00:27:11:04
Unknown
What where it's always comes down to where is the real value. And the value right now is kind of in the models but kind of not. So and we're getting to a point to where the moat if that's the right word for this. But you know a lot of it has been around these models learn me like over time.

00:27:11:08 - 00:27:30:22
Unknown
This model knows. So much about me that I couldn't possibly think about switching between them. Even that's being kind of chipped away at because now we're getting data portability between these, different services. And so that goes away. Yeah. Is your moat is the next story we'll talk about is your moat, the amount of compute you have.

00:27:30:22 - 00:27:55:09
Unknown
But that presumes that scale is everything. And if your moat is well, I have an love that smart everybody else does. Is it going to be instead world models as we've discussed often in the on local on valley worldview. And Jensen Wong. Yeah. It's an early industry. We don't know that very, very early.

00:27:55:11 - 00:28:22:26
Unknown
Very interesting. Have you just just as like a general check in on, on kind of these different models? Have you played around with cloud? Much like not enough like Gemini to do it. You know, I've got to do it more. And the thing is. Yeah. Pardon me, I'm going to probably use my maybe some people off. There's, there's a view of some people on religion that you can't really understand the religion unless you buy the religion, unless you are, you know, you are a believer.

00:28:22:28 - 00:28:45:05
Unknown
And only from that position can you fully appreciate, the value of it. And the problem with Claude is I don't think you can fully appreciate what Claude could do unless you're a coder. Even though you don't need to write the code Claude does to understand its full capabilities, to be wowed by it is I, I think, has to be your coder.

00:28:45:08 - 00:29:09:12
Unknown
Because you see I used to spend days doing this. Look, it did it and it's good and it works. Whoa. Yes, yes I think you're right. But that actually kind of ties in that. Part of the reason that I bring this up is because I have spent the last handful of months spending a lot more time with Claude, and I've gotten to really fall, like, fall in love with it, like it's it's really useful for certain things.

00:29:09:14 - 00:29:29:20
Unknown
The other day I was working on, on trying to kind of like wrap my mind around thing, you know, how I invest my time into my business and, you know, time versus money or how my time breaks down into an hourly thing. And if I'm helping my wife with her business a little bit like the economics of all these things.

00:29:29:23 - 00:29:50:08
Unknown
And it occurred to me at one point to be like, all right, like I'm seeing this all represented in like a in, in a text based lens sort of thing. But what I really need is I really need this visualized. And so I said, Claude, visualize where we are right now. And it ended up creating this. Like, that's all I said was visualize this.

00:29:50:08 - 00:30:15:14
Unknown
That was my prompt was visualize this. It ended up creating this interactive with sliders and numbers and everything. It pulled out all of the most important factors that I was considering in this kind of thing, and put it into an interactive, dashboard that I could then go in and be like, okay, well, if I write for articles instead of three, or if I'm paid this versus that, it's the new spreadsheet.

00:30:15:17 - 00:30:30:28
Unknown
And, and, you know, I'm, I'm moving these things around and it gives me a little, you know, bottom line summary down at the bottom, it was very much like a coded app that I, I didn't even I don't have to be a coder. I don't have to be a I think I just had to know to ask for it to develop.

00:30:30:28 - 00:30:49:19
Unknown
You had to know what you need. Visualize it. Yeah. I knew what I needed was I wanted to look at something that I could honestly, I just wanted to look at something. It gave me something more than that. It gave me something that I could get in there and tweak. And suddenly I had this under this broader understanding of the problem that I was faced, that I was using it for in the first place.

00:30:49:21 - 00:31:07:21
Unknown
And so, yeah, I don't know that you necessarily had to be a coder to see that. When you see that, that gets my imagination going where I'm like, oh my goodness, these things are even more useful than I thought they were. Yeah, it's such a great example because because what your what you're harkening back to is what made the apple and thus the personal computer valuable, which was Visicalc.

00:31:07:21 - 00:31:26:18
Unknown
Right. Right. Dad. Brooklyn and Bob Frankston. And it allowed executives to say, what if I watched my own father, who was a sales executive, be able to sit down and you have to forecast for all of this field and do this and that, and now the spreadsheet, let him ask, what if? And so that's what you just did.

00:31:26:21 - 00:31:46:22
Unknown
The other thing that I think is important and you make a great point. So Leo Laporte, our friend Laporte is going gaga over Cloud Code. Absolutely. Gaga. Yeah. And and I kind of can't discuss it at the same level because I don't know as much as he does. Paris Martin I was doing some of this, gamely. But Leo was kind of saying, you can't understand unless you could do all this stuff.

00:31:46:25 - 00:32:03:02
Unknown
I mean, Leo does have a coding background. Oh, I suppose, yeah, that's, I think a little bit a little bit of pressure. She's brave and tries it, but Leo's argument is you have to give it a big challenge. And that's what you did. You did know was a big challenge. You said, I didn't know that that was the case.

00:32:03:02 - 00:32:17:20
Unknown
Right. But it came back with a big answer. And I think I think you're right. I think that's probably the key, what I do, I don't have a bunch of big challenges. Right? I need to write paragraphs. Right? Yeah, I can have it write those paragraphs, but I'm not going to like it because I fancy myself a writer.

00:32:17:25 - 00:32:40:08
Unknown
So no, I'm not gonna have to do that. So I need to find those big challenges that then say that it fails or impresses me. Yeah, that's the key. I think you're right. And I'm. Yeah. And, you know, maybe this is a lot of topic, but by doing that the other day, by having that experience, it's like, now I've carved a new pathway in my brain, right?

00:32:40:08 - 00:33:00:00
Unknown
To know, to ask that or to know that that's a pursuit that I can take if and when I need to. And, that broadens out the capability. And so for people who aren't coders, you know, maybe, maybe it becomes easier for them to discover things like that, and then their imagination goes. And I mean, at that point, it really doesn't matter if you're a coder or not, you know, if you're a developer or not like it.

00:33:00:00 - 00:33:27:07
Unknown
It created a little mini app for me that didn't exist prior and all. I said, yeah, and a key. Sorry, it a key. I don't think it visualized it. In the end it functionalized it. Yeah that's true right. I took a question and gave you something that was functional that you could interact with. Yeah, I don't even know what I was expecting when I said, visualize this, I didn't know if I was going to get like a pie chart or I was going to get a graph or something, but I was I was just honestly curious.

00:33:27:07 - 00:33:40:29
Unknown
I was like, well, take all these things we've been talking about and put it into a visual representation of some sort, what would that look like? And it gave me that. I was like, oh, that was actually more useful than I thought I was going to get. I've got to I got to find those causes to play more.

00:33:41:02 - 00:34:04:19
Unknown
Yeah, yeah. Then again, stumble on them. You could be the better executive that use the tool, to do interesting things. And then what happened Jason. Why why why? Well, we've talked about Open Claw and, on the show before, a few weeks ago and talk about people going gaga over things. Open claws just been the darling, right?

00:34:04:21 - 00:34:22:13
Unknown
Yet at the same time, it is highly imperfect. There's for every story that you read of. Oh, my goodness, I can't believe I built this thing. If there's another story of, like, oh, you know, this, you don't everybody doesn't want to rush and do this because it's still early days and there's still a lot to risk in summer.

00:34:22:13 - 00:34:44:02
Unknown
You, meta safety and alignment director gave the open claw open source AI agent complete access to her inbox of, you know, probably, among other things, I'm guessing, and told it to confirm before acting through a task inside of her inbox. What did you say then? What do you say? Open. Closed. I think I heard that I.

00:34:44:02 - 00:35:03:17
Unknown
Yeah, I don't care, I don't really I choose not to, understand what you just said there. And instead she witnessed it, ignoring those orders and rapidly deleting her email inbox as she tried to get it to stop, she had to use her phone to try and do that. Was she actually able to to get it to stop, or was it just like, don't think so.

00:35:03:19 - 00:35:24:12
Unknown
So I couldn't stop it from my phone. I had to run to my Mac mini like I was defuzing the bomb. She said, oh, there we go. Oh I'm sorry, summer, I'm afraid I can't do that yet. That's how I was stressful, right? I mean, God bless her that she's open about this. She's a bit of because, you know, she's the head of, safety and live at meta superintelligence.

00:35:24:12 - 00:35:46:10
Unknown
You'd think she'd be like, oh, that is the thing, right? It's kind of like, okay. I mean, but at the same time, like, who? No one knows, I guess, what to expect out of these things. But that should be the warning in and of itself. For someone to not do this is like, well, literally anything could happen. Do you feel comfortable with the possibility that your inbox might get completely nuked by installing Open Client?

00:35:46:11 - 00:36:09:29
Unknown
If you do feel comfortable with that possibility, go ahead and do it. Just don't be surprised if it happens. Yeah. So she identified it as a rookie mistake. She's no rookie yet. She's not a rookie. And while this could happen to everyone, it's not exactly reassuring to know the work of the people in charge of making sure artificial intelligence, act in accordance with established guidelines at one of the largest tech companies in the world is out here making the same missteps.

00:36:09:29 - 00:36:28:01
Unknown
A novice would just give small, see, they're humans just like you. And they are. They are. So you know what it does do the the heart. Good. In a way. Yeah. Yeah. Yeah. It's like it's like, somebody else gets phished and, you know, you almost get phished once and you almost fell for it, or you fell for it.

00:36:28:03 - 00:36:49:29
Unknown
And then when somebody else tells you, you think, oh, God, I'm not the only one. Mac minis selling like hotcakes. Because people are buying these things in order to spin up their own, open kind of servers. And, you know, install it on there so that it's a more, I guess, controlled environment that isn't just, yeah, you can, you can, you can er gap it to your stuff.

00:36:50:02 - 00:37:11:07
Unknown
Yeah. Yeah. Which is, which is compelling and interesting, but yet at the same time like it's still the, the, the potential impact here. Negative impact is still very great. There are there's a whole lot to learn about this before. It's the sort of thing that any, you know, any random person should feel comfortable just going and doing, you know.

00:37:11:07 - 00:37:31:01
Unknown
Yeah. And this goes back to the door. Opening story again strikes me that what analyst at Apple would be sitting there thinking, oh, up, open cloud comes along, we're going to see a rush on Mac minis. Nobody even pronounce that right now. The cause and effect here is is completely unpredictable. And that's the way the world works.

00:37:31:03 - 00:38:04:09
Unknown
Yeah, yeah. Suddenly those those moments just happen. And, that's that's where Apple finds itself with its Mac minis. Speaking of open cloud, Google also with anti-gravity, is apparently cutting off anti-gravity access for open claw users. They detected a malicious usage. They say accused some workflows of exploiting token access through third party agents, through open claw, and degrading the service for others in the process.

00:38:04:09 - 00:38:24:07
Unknown
So basically, I don't do they cut off everybody or those who went too far. They said that. Oh, that's a good question. Some users were using anti-gravity to access a larger number of Gemini tokens via third party platforms like Open Claw, which overwhelmed the system for other anti-gravity users. Mind you, these are people who were paying $250 a month.

00:38:24:10 - 00:38:53:13
Unknown
So they cut off several users, underscoring the, says NTT data. But that, that I think led to, oh pinch lock creator Peter Steinberger dropping Google support. It's entirely I think so you know hard out there for a lobster. Yeah. Well it's funny with Steinberger because because anthropic went after him. He you're for for the brand.

00:38:53:15 - 00:39:10:15
Unknown
You're dead to me. Anthropic. I'm not going to work for you. I work for OpenAI. And Google says people are using it badly and you're dead to me. Google. He's feeling a little too powerful. I think at the moment. Are you at all compelled by by open out? Like the more I've read about it, like I still don't feel comfortable like I am.

00:39:10:17 - 00:39:29:27
Unknown
I'm definitely the kind of person I just, throughout my life that knows just enough to get myself in trouble on certain things. And so I feel like with Open Claw, I would, I would install it and I'd be like, I know enough to get it running. And yeah, and then my inbox would be nuked and I, and I, you know, I do it on on a this week on Google.

00:39:29:27 - 00:39:53:16
Unknown
You know once again you to confess it here. Yes. I think yeah. That was a great moment when I, I came across that the other day, I was like, that's I love that I have that as painful as it was, I love that I have that it was, it was a great, charming. Jason moment when he confessed for all the world to see because it happened right in front of us, that he cut off his own access to us on the phone.

00:39:53:18 - 00:39:58:24
Unknown
I think what happens, Jason, is that,

00:39:58:26 - 00:40:16:01
Unknown
People will be out there experimenting, coming up with, with applications. We couldn't imagine that someone will make a decent version of. Yes, probably at a big company we trust. But this is the this is the problem with the antitrust stuff going around with with the app stores is say, well, you have to cut down the app stores, let everybody get in everything on your phone.

00:40:16:08 - 00:40:36:00
Unknown
No, that's not safe. I like the fact that the App Store exists. I like the fact that Google and Apple takes some responsibility for the safety of the applications that you can then put on your phone. We need that, and we're going to need the same thing here with the ethic world of AI, that I could come because you're in an ecosystem and Microsoft would love that to be the case.

00:40:36:00 - 00:40:56:11
Unknown
And as long as you're using a Microsoft app on your Microsoft software, they will vet it for you. That's a little limiting, though. The the larger question is whether there's this open view that, yeah, I can get apps from anywhere, but somebody is very in it for me. Somebody is going to whatever effort that is. And I think that's what we're going to need for this to take off.

00:40:56:14 - 00:41:18:03
Unknown
Well, an open cloud was, was, was it Peter Steinberger that was, that was brought into open AI or open cloud was brought into AI because open AI both both. So then is open AI the company to do that? Maybe. Maybe. Do we trust them? I mean, I don't know. Yeah. I guess the question is do we trust them to to do that?

00:41:18:05 - 00:41:35:21
Unknown
I guess we'll find out. Yep. Find out soon enough. So that's open. I continue to be interested, but not interested enough to actually do something about it because like I said, I'd probably go, oh, what a watch, I can't wait. I wish I could. I wish a lot of people would post what they make with it, so we can see that.

00:41:35:23 - 00:42:06:19
Unknown
Yeah, maybe. Fascinating to watch, but yeah, I'm I'm hanging back. Yeah. Interesting. Let's see here. Boop boop. Do creativity and I, I love this topic. It's interesting. It continues to evolve. And this is very interesting as well. This is so let's see here Chinese director Jia Zhang I I'm, I am almost certainly getting his name wrong, but he is a famous Chinese director used see dance 2.0.

00:42:06:19 - 00:42:33:01
Unknown
See? Dance is a video generation platform to generate a short film called Jia Donkeys. Dance. I guess his dance. Sorry. Say yes. His dance. Essentially. Thank you. That's probably the better way to say it. It features two AI versions of himself, basically on screen, debating all sorts of things. And it's, you know, kind of cinematic. And I won't play with audio, but I can show video and everything.

00:42:33:05 - 00:42:58:20
Unknown
I gotta say, like watching the video itself, the quality is really good. Like very good. It's compelling. There's one version online that didn't allow, closed captioning and, subtitles, but the one we have here does. So if you go looking for it, folks, you can find it. And it starts off with the director meeting him, finishing a scene, and then meeting himself and then having conversations with him with his AI self.

00:42:58:20 - 00:43:19:21
Unknown
But they're both here and they go through scenes. I think some of these scenes are from the director's work in the past as well. So there's references here, the that I wouldn't understand. Yeah, but it is fascinating because it's an exploration of content, you know, am I real? Who, me. And who's going to make this and where does it go?

00:43:19:24 - 00:43:38:08
Unknown
And the fact that he uses, he has to do it before. One example here is, is that, the AI, the clearly made up AI guy as the also made up our guy, you know, do you like this version of me? I lost I lost ten, 15 pounds and the the other guy are 10 pounds as well.

00:43:38:10 - 00:43:57:25
Unknown
No, I want it back. You know, I don't like that. Look, it's very sly. It's very funny. But it's a neat exploration of AI creativity. This last week, I was going to go to Stony Brook to talk to the class. I've said on this show a few times, I wrote a syllabus for a class of AI creativity.

00:43:57:28 - 00:44:27:22
Unknown
And my, my colleague there, is running the course now with 436 students. It's amazing. And I was going to go talk to them, but of course, with my cane, not being able to get very far, I didn't. So he had me record a 15 minute, mini lecture, which I'll put up online. And, so anyway, I sent him this video because this is a great example of the exploration of what does I mean for personal creativity and self-expression and who you are.

00:44:27:24 - 00:44:49:15
Unknown
And it's just I just love, playing with this topic. So I recommend this is a variety story we have. But you can go to YouTube and look it up and I recommend watching it. It's just going to fascinate. I think. Yes. Everything that you said, you know, it's fascinating to me too. I, I've known as we've as you probably have too.

00:44:49:15 - 00:45:10:11
Unknown
I mean, we've talked about it like there is a time where these models get so good that like, we get over the, but there's these imperfections or, and there's these tels and, and this, that, and you know, that, that stand in the way of a what is, you know, could be a compelling story or something that could be told in this medium.

00:45:10:14 - 00:45:33:06
Unknown
And, I mean, I'm watching this thing and you know, it's it's good. It's spot like it's I'm sure if I looked deep enough, I'd be able to find some things that would be like, oh, that was weird. But it's really hard to my eye is compelling, determined. Yeah. And it tells a story. There's a cohesion to it. The kind of the the, the layout of the of the different camera angles and everything.

00:45:33:06 - 00:45:56:06
Unknown
Like it follows all the conventional kind of, you know, it doesn't. Every shot isn't that typical. I push in thing that you see a lot of. Right. Exactly. Yeah. I generated videos like it follows a lot of the standard conventions of theatrical filmmaking, and so much of it is dialog driven. I like I don't know how much of the audio was also done by the model.

00:45:56:06 - 00:46:16:14
Unknown
That's a question that was scripted. I'm not sure of that. I think that it was. Yeah, it was it was scripted. But it's it's yeah, it's fascinating now. It's really and there's there's subtle attitude to it. The the eye creation of the director, who seems to be the real guy, has a certain wry quality to him.

00:46:16:14 - 00:46:37:10
Unknown
Whereas the eye creation that's the eye person isn't sycophantic, but is a little too happy. Yeah, is a little too glib. There's there's a there's a there's some subtlety to it that the what they say. But what's fascinates me too is we had. So see, dance was used to create a fight scene between. Who was it?

00:46:37:10 - 00:46:59:21
Unknown
I can't remember now. Brad Pitt and. Oh, that's right. Who was the other fight. Yeah. And that caused that. Hollywood went nuts and they went crazy. And you can't do this. This is awful. And this is sort of illegal. No no no no no no no no no no no. Which means to say Hollywood didn't learn crap. Whereas my going to get a an honest to god real director to do this with them.

00:46:59:23 - 00:47:41:15
Unknown
This is, is really, really interesting. And it shows that Chinese culture is learning more than Hollywood. That's the warning to Hollywood. They're taking more wake up, more chances. Yeah, they're you know, taking more risks there to to see just to show what is possible. Yeah. Yeah. Well I mean, you know, an a part of that too, you know, you linked to a New York Times article around kind of sentiment around this and I think the headline is people love the.com boom, the AI boom, not so much, which kind of talks a lot about what we've talked about on the show many times that that, you know, that history is littered with examples that were kind

00:47:41:15 - 00:48:04:12
Unknown
of following in the in the trend lines. Right now. This article is saying that, you know, technology has long been kind of touted as the sort of thing that would save the world or be great for humanity or whatever. Radio would bring perpetual peace on earth. And TV, you know, had its own thing or whatever. And, and the.com boom, you know, that was going to be a whole sea change for how we, how the economy is run.

00:48:04:12 - 00:48:27:15
Unknown
And that is obviously been proven true. But, that something is different this time around with the AI boom and that people just have their, their minds made up on how they feel about this. And that's something to overcome. Two things driving nuts about this. The first is that my colleagues in journalism who fancy that they write the first draft of history, and so I would say means that they ignore history.

00:48:27:18 - 00:48:53:26
Unknown
And, even recent history. So, you know, I'm old enough to have lived through the, the internet boom. And I remember all the people who thought that it was going to ruin society and it was going to ruin media. And the blogging was ridiculous. And who once who was a citizen surgeon man? And all the stupidity. And if you go back, the same thing happened with radio and television and print and novels and so on and so forth.

00:48:53:26 - 00:49:10:15
Unknown
This is what I do, is research this. So it drives me crazy. They're acting like this is special and new. The second thing is this is a self-fulfilling prophecy, right? Media like the New York Times went out there screaming about how AI is dangerous. And then they go and they say, we found people who think AI is dangerous.

00:49:10:17 - 00:49:38:27
Unknown
Yeah, it's a tautology. This is this is how opinion polls work. This is how news works, is they create a trend, then they report on the the trend they created without admitting their role in it. So and then we have a time magazine cover story and time is the is the weekly journal of moral panic. And so you know remember cyber porn cover of time by an author who since has said how embarrassed he is that he was associated with it.

00:49:39:00 - 00:50:06:05
Unknown
Now we have the time. Cover is the people versus I what's what's what's behind this. Further, though, there's a political angle to this. That's that's really interesting times had a story today about a guy running for Nadler's, seat in Congress in New York who used to work at, crap. What's the what's the what's the company that scares everybody with,

00:50:06:07 - 00:50:26:19
Unknown
Alex, not in front of the P. It starts with the P perplexity, but the other one, Palantir. Palantir works for Palantir. He was helping eyes, but then he saw the light, and he left Apple for. I'm sure he cashed in all the stock, and now he's running for Congress. He's being the anti AI anti-technology guy.

00:50:26:22 - 00:50:50:01
Unknown
And the times has a story about how there's a there's a PAC run by, AI people that's trying to defeat him. Basically all the candidates in this race are questionable. But they're trying to present this as if this is, the present view of AI, that this guy is some kind of holy warrior here, doing this.

00:50:50:04 - 00:51:13:22
Unknown
It's media doesn't admit its own role in this and that, and that drives me fairly crazy. And so, I don't think we can reliably see this is the it's it's like I, I'm, I got off on a tangent. The political subtlety here is that under Trump, the right is defending AI because they're defending the AI companies and the market around them.

00:51:13:25 - 00:51:56:10
Unknown
And so the left sees an opportunity to become anti AI. And I think that's dangerous because each is unsubtle and without nuance. Yeah. And I wouldn't want to see progression Progressive's defined progression as slapping down technology any more than I want to see conservatism be boosting technology with no breaks. Right. Neither has subtlety or sense to it. And so I think what we're seeing is media are forcing the two sides into corners in a boxing ring that is not going to be healthy for the discussion of technology.

00:51:56:13 - 00:52:15:22
Unknown
Yeah. Well put. I love how you put that. Super true. Let's take a quick break. First of all, if you haven't reviewed this podcast, please do helps get more people, listening and watching it on Apple Podcasts. Give us, you know, give us a great review. We'd appreciate it. We're going to take a quick break, come back, do a speed round and get you out of here.

00:52:15:22 - 00:52:22:12
Unknown
So hang tight and we'll, talk about Facebook's investment into AMD coming up in a second.

00:52:22:15 - 00:52:49:06
Unknown
All right a few quick stories and then we'll get you going here. Meta is making the rounds today with news that it's buying six gigawatts of AMD, AI chips in a supply deal that could exceed $100 billion. Also grants meta ability to acquire up to 10% of AMD if it hits key purchase milestones and gets the stock price up to the, I think, $600 right now, it's at around $200 a share.

00:52:49:06 - 00:53:17:09
Unknown
And, you know, that would be like a penny, share that would be granted to meta, which would come out to about 10%. You know, this is, meta bulking up on its non Nvidia GPU supply, kind of showing that it can diversify a little bit and that all roads do not lead them to Nvidia necessarily. While it just did a big Nvidia deal, giving itself in the long run access to 1.3 million GPUs in total.

00:53:17:12 - 00:53:41:21
Unknown
And Nvidia, which was supposedly going to invest 100 million in open AI now is a 30 million. So, yes, there was a little diversification going on here, but it's still part of the circular economy. Yeah, it's still part of the scale. Scale, scale compute, compute, compute, using money in the future to do this. And so, you know, we'll see where, where all this goes.

00:53:41:23 - 00:54:04:22
Unknown
Other angle to this that I didn't realize is that meta because it's, it's is huge cash flow, but that cash flow is tied up in employee stock and stock options and paying the taxes on all that as it goes. So meta is having to borrow hugely to do all this. So this is there's a lot of debt coming out with this.

00:54:04:22 - 00:54:23:23
Unknown
And that's, that's where you can get in trouble that that itself is a balloon. So watch this story. Indeed. Yeah, it's a good point. Samsung has its, big smartphone event tomorrow. I'm actually going to be there tomorrow morning. That's part of the reason. Ergo, we are early today because, well, let's say thank you very much for choosing Wednesdays.

00:54:23:26 - 00:54:43:24
Unknown
They want, you know, before they scheduled this. But anyways, maybe next time, they're going to be unveiling the Galaxy S20 six smartphone lineup. And ahead of the event, The Verge is reporting that Samsung is building the devices with a multi-agent strategy. So, you know, you you've expected that Gemini is going to be on the devices as they have been, that Bixby is going to be a part of it.

00:54:43:29 - 00:55:09:12
Unknown
Well, now The verge is saying this multi-agent strategy means that perplexity is going to be built into Galaxy AI, in the process as well. So hey, Plex, we're going to have to remember all of our different hey, queries. Right? There's they're all different, anyways. And you can query the web, act on, on device data apps, third party apps, that sort of stuff on these phones, apparently, reportedly, let's say.

00:55:09:16 - 00:55:37:28
Unknown
And all I have to say is I'm old enough to remember perplexity. Yeah. Not talking about them very much, but around robot. But you talked about them a lot back in the day, but now we did it. So we're glad to see they got a little, a little boost here. There you go. New free Google Apps tool called Pinelli, which lets small and mid-sized businesses take basic product photos and turn them into polished studio style, imagery, and marketing images, that sort of thing.

00:55:37:28 - 00:56:00:19
Unknown
Uses Google's brand DNA brand context, along with nano banana image generation model to do all this. You know you've got your prompting, your text based prompting for edits, style references, integrating product URLs into it. Definitely the important thing here is how well it adheres to the source, because when you're talking about marketing, it's got to be spot on.

00:56:00:21 - 00:56:23:10
Unknown
It's impressive. You know, they hold up a jar of what looks like, you know, skin cream with a hand and not a pretty picture. But then it takes that and puts it in environments, shows people using it. It's pretty sophisticated. I'm surprised we haven't seen that agency stocks tank this week because of this. Or maybe, because it's it's pretty amazing how you can do it in a brand consistent way.

00:56:23:13 - 00:56:46:05
Unknown
And it's great for advertisers and Google. It also means ever since AdSense, Google thinks for the little tiny advertiser, you know, moms marmalade, can now use this tool and, and have pictures of people just, finding it very delicious. Mums marmalade marmalade I don't know. Have I ever had marmalade? I don't even know if I've ever had marmalade.

00:56:46:08 - 00:57:16:27
Unknown
It's fun to say it has. It's it's jam with an accent. Okay. I'm sure I probably have. Yeah. It's a, it's fancy jam. Yeah, yeah. And finally, anthropic is extending its clawed co-work agent into human resources investment banking design using, these co-developed plug ins, for big name integrations. You know, it's get like a suite of finance focused customization apps, tuning for equity research, private equity, wealth management, all sorts of things.

00:57:16:27 - 00:57:34:28
Unknown
And yes, as we talked about earlier, Wall Street reacted swiftly to the news. That's just the way of the world. The last month. It seems like anthropic announces something. Wall Street goes nuts over it. We've got to get to the point where we get real data. I'm sure some jobs are affected. They happen, we know that. But we need real data about that.

00:57:35:00 - 00:57:54:14
Unknown
Yeah, yeah. So know what the impact really is? Indeed. Well, Jeff, thanks so much for, for sitting through an hour sitting at your desk. I hope, I hope everything is all right. I'm still okay. Good bye. Good, good. Jeff jarvis.com for everything Jeff Jarvis, including his upcoming book, Hot Tip. If you want to, preorder, you can do that.

00:57:54:14 - 00:58:18:09
Unknown
Just go to the site. Should be able to grab it there. I inside dot show for everything related to this show. You can subscribe. You can find all of those reviews that you're totally leaving about this show, right? Right. Ladies as well please everything you need and then Patreon.com slash AI inside show. That is the place where you can go to support us on a deeper level.

00:58:18:09 - 00:58:38:06
Unknown
We appreciate when you do that. Like so many of our executive producers, Doctor Do, Jeffrey Marie Cheyney, radio actual one of 3.7 Dante Saint James Bond a Derek Jason night for Jason Brady, Anthony Downs, Mark Archer, Carsten Szymanski. We don't do it enough, but I'm going to give you a round of applause. Yeah, you are awesome. So thank you for your support.

00:58:38:11 - 00:58:54:06
Unknown
You're our peeps. That's right. You are our peeps and we appreciate you deeply. All right everybody, thank you so much for being here. We will see you. Well we'll see you later this week actually with an interview episode. So look for that. And then next week for another, news episode. There we go of the AI inside. What?

00:58:54:06 - 00:58:59:29
Unknown
Samsung doesn't do anything to mess it up. We'll see. I'm not guaranteeing anything by everybody.