Big Technology Podcast - Kamala's Tech Agenda, Drunken AI Spending, Meta’s Open Source Necessity

Episode Date: July 26, 2024

Ranjan Roy from Margins is back for our weekly discussion of the latest tech news. We cover 1) Kamala in, Biden out 2) Kamala's tech agenda 3) The importance of Lina Khan 4) Do VCs regret backing Trump after the past week's developments 5) J.D. Vance couch meme 6) Twitter at the center of the story 7) AI CapEx goes wild 8) Google's resilience through the first AI wave 9) OpenAI launches a search engine, SearchGPT 10) Elon's AI supercomputer 11) DeepMind's math-solving AI 12) Meta's need for open source AI 13) Southwest Airlines does away with open seating, --- Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. For weekly updates on the show, sign up for the pod newsletter on LinkedIn: https://www.linkedin.com/newsletters/6901970121829801984/ Want a discount for Big Technology on Substack? Here’s 40% off for the first year: https://tinyurl.com/bigtechnology Questions? Feedback? Write to: [e-mail]

Kamala Harris is in. We sort out where she stands on tech. Is runaway spending starting to catch up with big tech and the opening eyes of the world? Meta goes all in on open source with its new Lama model. And Southwest Airlines is ditching open seating. This is going to be interesting. All that anymore is coming up on a Friday edition of Big Technology Podcast right after this. Welcome to Big Technology Podcast Friday edition where we break down the news in our traditional cool-headed and nuanced format. Well, a week after we started talking politics, we've had a swap at the top of the Democratic ticket.

We're going to talk all about how Kamala Harris might approach the tech industry. But we're going to do that briefly and then get into the meat of our show where we talk about the capital expenditures of companies that are working in the open AI field. Meta's new Lama 3.1 model. And then, of course, the Southwest Brujah and plenty more. So joining me, as always, on a Friday here is Ranjan Roy. Ron John, welcome back to the show. I'm back in New York. I have my sure mic in front of me, and I'm ready to talk about this week's news. Okay. Well, we are set and ready. And last week, we talked about how Joe Biden was at the top of the ticket, but we don't know for how long. I think you presciently said, you know, probably within 48 hours, it could all be different. And lo and behold, 48 hours later, it was all different. And Biden is out. Kamala is in. And the big question here, because like last week, we talked a lot about the differences.

between the candidates and their approaches to tech is how Kamala would be different from Biden on tech and then whether she'd actually be all that different from Trump on tech. I think the big question here really comes down to is she going to keep Lena Kahn or not. I mean, there was a story in the New York Times talking about how, you know,

what her approach has been to AI regulation because she's been the AIs are, I mean, it's been very pro-business. Everything that she's done has been extremely pro-tech and pro-business. But, of course, the question is, do you keep the same leadership at the FTC? That's the big question for me. I'm curious how you're thinking about this.

Yeah, I'm thinking very similarly because I think whatever decision Kamala makes around Lena Khan will represent exactly what her overall stance on to the whole technology industry will be. And remember, Kamala, it kind of reminds me of the Obama era administration to tech, definitely hands off, definitely much friendlier. That's been the way she's generally approached things. And I think the Biden administration has been very, very aggressive, and I think it's a good thing in terms of the way they've approached antitrust.

And I think that could be a big change under a Harris administration. Yeah. And I think that's the difference when you have a senator from California versus a senator and then vice president from Delaware. You know, say what you will about California's anti-tech regulation. And there's certainly been policies. that have driven some in the tech industry out, but not many, but to places like Austin and Florida, you look at Harris, and she understands the economic engine that these companies are.

Well, it's not just the economic engine. You have, I mean, remembering that people are influenced by their surroundings, when you come up in California politics, and when your brother-in-law is the chief legal officer of Uber and someone who's been very, very aggressively and successfully anti-regulation, it has to influence the way you think. I can't imagine it wouldn't. So to me, the most interesting part of this will be, I'm so fascinated what's happening internally right now.

Like, what are the discussions? Who's she going to keep on? Is she just because Reid Hoffman coming out and saying that she should get rid of Lena Kahn and here's $7 million that I'm donating, does that stuff influence her? Or, you know, to me, me, the Biden administration's messaging around controlling big tech and working for common people

and data privacy and all these things, I think could be powerful in this election cycle and I think still will resonate. So where she stands on that, I think is going to be very, very interesting. Yeah. And I think that's just going to be something she'll figure out over time right now. She's probably just like, how do I stand up a presidential election? She's got memes to me. She's got memes to work at right now. It has been interesting watching how memeable she's been. Coconut memes, J.D. Vance Couch memes, you name it. Okay, the J.D. Vance Couch meme. Do you understand the J.D. Vance Couch meme? I do. I have spent a lot of time thinking about and analyzing the J.D. Vance Couchmeam.

Okay. Let me set this up because the only thing, I saw a video of J.D. Vance being compared to a couch. And then the AP apparently ran a story saying, no, J.D. Vance didn't have sex with a couch. That's what I've heard about it. All right. Would you like me to walk you through the story of the J.D. Vance Couch meme, Alex. Yes. All right. So someone totally shitposting tweeted that J.D. Vance in Hillbilly Elegie, the book he wrote in 2016,

had written about how he had stuck a rubber glove in a couch and had sex with it. And this was in Hillbilly Elegie. Now, this is completely untrue. or it's not in the book. This one person posted, this one absurd, ridiculous thing. And we're going to talk about Twitter more. And this was just Twitter at its finest, that like this absurd, ridiculous thing. And I think a great meme is great when it's just so ridiculous, but there's part of you that you're like, I could see that being true.

That could be, especially within hillbillology in the book that, like, And the memes have been good. They have been great. So I didn't fully understand what was going on, but I understood that there was this rumor because the AP came out with a debunk, which the AP eventually removed from its site. And I think it should have because the headline was like, no, J.D. Vance didn't have sex with a couch.

And it's like, what are you even doing? No, no, I actually thought, to me, that bothered me a bit because it kind of shows how the absurdity of fact-checking when people are too online. Like, it's, okay. It's embarrassing for the AP to do. put that out there like no one's thinking about that outside of like terminally online people on Twitter and my yes including myself much that way you'd count there's not a name count and i don't

include a fact check i don't anyone who is enjoying a good jd vans couch meme is not actually thinking he had sex couch and even if you are you know i mean maybe uh you're not the type of person if you believe that you're not reading apie fact check that's not your home page on your It was really an embarrassing moment for the AP. But you are right. The memes are good because I did see a photo of J.D. Vance and his wife sitting on a couch and someone wrote sort of great picture of my favorite thruple. Yeah.

J.D. Vance just has defined what intersectional means. Is that someone wrote? That was one of them. Oh, no, I've been enjoying this. But I think the Kamala side, I mean, to me, it's been fascinating because there's part of me that the coconut memes, the unburdened by what has been memes, like this whole energy on one side is amazing to watch because we have not seen that level of kind of like social media

enthusiasm since Trump in 2016. Like we haven't seen that level of just like energy around a candidate and just absurdity and people having fun posting online. So I think I think it's really interesting as powerful. But obviously, the big question is, is this just what we are all looking at? But in reality, majority of people don't even know the coconut thing or understand it and don't care. Of course. And they won't influence them at all. And maybe they'll actually view it as a negative. So I think that that's, we'll see where that goes. But one thing that is directly applicable to our world is that the questions we had last week about some VCs coming out and support of Trump and, like, how this would change things for them.

There's been two things that have happened that have been really interesting. First of all, Kamala coming in has changed the calculus on the election. So if they were all trying to bet on the winning horse, it went from an almost certainty to still likely, but that sort of changes the, like, back the winner mentality. So that's interesting. But I think even more profound is that we've started to see, remember we talked last week about, like, wouldn't it make more sense to be neutral? And, you know, you've mentioned that, okay, maybe they're going to try to attract some, you know, founders with ideological similarities.

But for me, it's just like, all right, you're playing in politics. You're going to, everybody knows politics is a dirty game and you're going to get dirty. If your job is to give founders money, that is a risk. And I think we're starting to see it both in terms of the supporters of Trump and the supporters of Kamala. because David Sacks has been nonstop, who's the VC, who spoke at the convention, didn't do a great job of it, to be honest. He has been saying that the Democratic Party engaged in a coup by taking Biden out and putting Kamala in. And so Parker Conrad, who was the former CEO of Zenefits, Wild Sachs, was the CIO.

He said, let me tell you, coups are this man's specialty, because Sacks apparently led a push to get Parker Conrad out of Zenefits and then took over as CEO. So this is basically Sacks drawing incoming fire. And he says to, he, he quote tweets Conrad and says, you were sanctioned by the SEC, nobody else, only you, but you spent the last decade trying to shift the blame onto others for your own poor ethics. Okay. So for a little bit of context, like Zenefits had some compliance issues.

This was reported about by some of my colleagues inside BuzzFeed, but basically they had a macro program that would like. make it look like some of their benefit administrators were sitting through courses that that they weren't and that was part of the reason why Parker Conrad was ousted. So Sachs is like now, you know, basically public target there and, you know, fighting back. And then Paul Graham, who's I think more on the Democrat side, writes to him, do you really want the full story of what you did to Parker to be told publicly because it's the worst case of an investor maltreating a founder that I've ever heard, and I've heard particularly all of them. I was talking recently to another investor

about whether you're the most evil person in Silicon Valley. He thought about it for a few seconds and agreed that he couldn't think of anyone worse. This is Paul Graham talking to Sacks. We're not going to litigate the whole Zenefits thing or the, you know, the big, the coup one way or the other, but I do think that this is something that like, once you get into politics, the opophile gets dropped and people take this stuff very seriously. And, you know, they are, They will throw punches they've previously held. And I think that's what we're starting to see spill out this week. And it could get uglier.

Yeah, I think especially Paul Graham coming out. And I don't think I would qualify Paul Graham as a Democrat. I think he's a pretty independent thinker. Some views I agree, some I don't. But again, very respectable founder of Y Combinator. And I think like this is a perfect example. And I that David Sacks trying to get involved, trying to become, hyper aggressive on this, trying to get a spot speaking at the Republican National Convention,

you're right, you're going to take incoming fire on this and we're seeing it happen. And I actually, you said we're not going to litigate Zenefits now and we are not, but I will admit I'm excited to see what comes out because I don't think this is ending. And I think like David Sacks, I mean, has spent a lot of time being very, very aggressive in the things they talk about and the way he communicates. And I think it's the, it's going to come back around and it appears to be right now. But what do you think about the fact that Twitter has been so central here in this, in this moment? Obviously, people said that, like, you know, Twitter was done for once Elon took over. Clearly, it's the center of everything. Now, one

interesting thing that happened was that Biden, when he posted his, I'm leaving letter. on Sunday. He posted it on Twitter and everyone's like, whoa, Biden posted the news to Twitter. Why did he choose that platform? Actually, he didn't choose that platform. He put it on Twitter. He put it on threads. He put it on Instagram. You put it on Facebook. I'm sure other platforms. But the only one that people paid attention to was Twitter. So clearly there's a vitality that has remained under Musk, even if the business is suffering, which it clearly is, like Peter Thiel recently said that like this is not a good financial decision to invest. But I'm curious what you think about the centrality of Twitter in this moment

and how it's felt especially post Biden dropping out. I mean, I can tell you I've not been on blue sky for a while. Maybe it's where the reelection is going. Maybe it's all really going down on blue sky or even threads. I actually just opened up my threads again and it's just the most random out of the moment posts. Twitter is back. Twitter is back with a vengeance. I'm spending way too much time on it as I'm sure you are as well. And it's becoming where all of these things are happening again. And I think that's, it's, it is such a weird thing because there's a few parts of it. Again, from a business

standpoint, does increased engagement actually result in improved financial performance? One thing I had seen once was like if you, the costs of actually serving a heavy Twitter user, because the ad inventory is so poor, could they actually be losing money on total utilization, which would be amazing that you're actually costing Elon money. He's paying for you to use Twitter. I don't think that's happening. I mean, that was certainly hyperbole,

but I think the quality of advertising that I'm seeing on there is not getting better. It's still the same random stuff. Wait, are you saying that telling advertisers to go F themselves actually led them to leave the platform? No, I mean, if you think about it, like, they, I have to imagine all measures of engagement are skyrocketing right now. Everyone I know who is off, Ezra Klein is back talking on there.

Like, everyone, people are on there. That's where the conversation is happening. And they still can't get the advertisers. And it's clear that, and it's even the fun stuff is happening on there, the coconut memes and everything else, and they can't get them. So I agree that I think that was a problem. But the other thing that was interesting to me was, so I was on Twitter the moment that I refresh my feed

and I see the Biden post almost like one minute old. And the weirdest part of it was I had to do like five checks to believe it. And I still wasn't quite sure that because verification has been so screwed up, that what is real, what's not real, that Biden, even it's from the account at Joe Biden, I'm looking at it, I click on the letter, start reading it. I'm like, oh my God, I think he dropped out, but I'm not sure. So the verification, veracity side of Twitter has gotten into such a ridiculous place

that it's almost turning into just like even less of an information source and more of just a pure entertainment platform, which maybe is not the worst thing. Well, it's still vital for news. And there has been this moment where, like, people are like, okay, does this mean that threads is now important for news because Biden posted there as well? Let me read you the engagement numbers because they're crazy. So the Biden post that he's leaving on threads has 2.6,000 reposts.

The Biden letter on Twitter that he's leaving has 352,000 reposts. So just to mere 350,000 more. And this has even reached Facebook. I mean, this is from an information post. Even the head of product for threads, Emily Dalton Smith, was dissatisfied with how the policy played out in the apps and the app in the wake of Biden's announcement. This is from the information. She complained that there was no trace of news in her For You feed. It feels broken that a topic overwhelmingly, overwhelming my following feed does not show up at all in for you, she said, in the post.

And this is the point. And just think that, like, if threads, it's a double-edged sword and it's tricky, right? If threads does not engage in this content, it's going to be totally irrelevant. Like, to not be a player in this moment, it's a real-time network. That's what this text, a text network is a real-time network. And to not play in this way, you just see ground. And you were given, effectively, if your threads, you were given a gift by Biden to start this up and did nothing with it. Yeah, I think they've made it clear that they will not go.

into news, they're going to deprioritize it. And thread still feels like an experiment that they had its moment. They were playing with it. I mean, I genuinely think that Twitter is back and will beat threads, especially for real-time information, and that's what it does better. So, so I think, I'm curious to see where threads goes, how much it's mentioned to the next earnings calls. I mean, I'm not even sure if Zuck, I get like WhatsApp or Instagram message broadcast from him now, but I'm not sure if he's even threading these kind of things. So I think Twitter is winning on this one. All right. I think we've done enough politics. Let's talk about what's going on with Gen AI. So you posted, in our shared doc, you posted a story from the

information asking if Open AI is a good business. What do you think? Okay, so the information this week reported, and this is very, these numbers are pretty insane, that the company could lose as much is $5 billion this year, more than double the $2 billion that was estimated lost in 2023, and this is despite generating what's estimated between $3.5 to $4.5 billion in revenue this year. So let's say in the midpoint, you generate $4 billion in revenue. You're losing five, so you're spending $9 billion. And that's taking into account that Microsoft charges OpenAI a discounted rate for users of its cloud. So they're burning money. But also the same article had a quote from Sam Altman's when earlier actually was in 2022 talking saying whether we

burn 500 million dollars a year or 5 billion or 50 billion I don't care as long as we stay on a trajectory where we create more value for society that as long as we can figure out a way to pay the bills if we're making artificial general intelligence if even if it's expensive it'll be totally worth it. I think this is going to be a problem for them. I think that people are going to start asking questions around the numbers. We've talked about this a lot. Maybe that Goldman Sachs report was the inflection point

where people start to really pay attention to the numbers. But I think the way Open AI has been spending is not, it's going to be an issue at some point. And then you start to question, what's their competitive mode? And we're going to talk a little bit later, about my love affair with Claude and incredible experiences learning to code with Claude Sonnet 3.5. But, like, this stuff is so interchangeable. The moats are so low. And if the moat is enterprise, then Google Cloud and Microsoft Azure and everyone else has a competitive advantage against OpenAI.

So I just don't see what their advantage would be other than having hit product. after hit product, which maybe SORA is finally released and blows us all away. But I think this number, if it's real, is going to be an issue. Yeah, I've been saying all along that open AI is in the hits business. And by the way, it's not just open AI. So this is sort of the moment, especially we have like all the big tech companies reporting their capital expenditures. And this is the moment where we're starting to see these questions that have sort of bubbled up over the past couple weeks really, really become spotlighted. And this is from a thread that Michael Spencer, who is an AI watcher, shared on threads.

So fair enough, but he did put it on threads, but I found it because he commented on one of my LinkedIn posts and linked to it. So anyway, it just tells you enough. We've done the threads segment, but just a pile on here. So he said Alphabet's Q2 earnings report yesterday revealed a 91% surge and capital expenditures to $13.2 billion and $2.2 billion. spent on AI development alone, and he's saying generative AI is not sustainable. And he talks about then, he brings up the open AI spending $8.5 billion annual revenue, maybe $3.5 to 4 or even

$5 billion loss this year. And then he says anthropic situation is potentially even more dire, reportedly making one-fifth to one-tenth of open AI's revenue while burning 2.5 in compute costs alone. And he mentions that an analyst said we're unlikely to see material AI revenue. until 2025 or even 26. And he says my analysis, nobody is going to be around but big tech by 2030, most likely.

And he says most foundational model builders outside of big tech, Google meta, XAI, won't be able to fund their activities given the weak revenue generation. So this does seem really that we're in a moment where the ROI is starting to become front and center

and the funding is going to run out unless like commensurate revenue is going to start or even, you know, revenue that starts to touch the size of this spending hits. And it's just not here yet. Wait, two things on this. First, especially not to go back to the politics situation, but it is interesting to think about an anthropic, a mistral, and we've talked about this, is the bet for a lot of investors that they do get acquired for a hefty sum by one of the big tech companies?

And obviously, if it's Lena Con in charge, that's a very different story than if it's not Lena Con in charge. So I think who does win the election and whose head of the FDC does influence the direction the company like Anthropic goes. Wait a little bit, before you move to the next thing, I do think that the bet was that they are all racing to build the best model. And it's starting to look like it's just going to be a size thing. And it wasn't clear that that was the case in the beginning. So maybe some of these models are doomed. I mean, I don't know if that's the case, but I do think, oh, it's like anthropic doomed. There's a chance.

Don't take away Claude for me. I mean, either. Yeah, but I, yeah, yeah, I think that's exactly it, that, I mean, there's this moment that, like, 20 bucks a user, what does that have to scale to? especially when then if Gemini is free and gets better, then do you move away from Anthropic to someone else release a chat bot, Facebook, and we're going to talk about it, and meta goes open source. So, yeah, the competition is so heavy. Honestly, are we just in ride sharing circa 2015?

Because it's kind of what it sounds like, where these companies are losing insane amounts of money. The goal is to build the best model and monopolize or at least gain market dominance. And meanwhile, you and I get to probably use Claude for either free or even at 20 bucks. It's a bargain for what it can do. And maybe the compute costs we're running are actually more than $20. So maybe this is kind of. Yeah.

No, no, exactly. So it kind of feels like ride sharing 2015 right now when Uber's in New York City, I could go from downtown to uptown for like eight bucks. and now cost me $30 to $40 to just go across town. And the subsidized life is over. So enjoy those queries while you can, I guess. Let me ask you this. Have you seen what's happened to the S&P this week?

Oh, I think these questions, to me, the fact that Google's overall ad sales were growing, sorry, their ad sales slowed, but in the AI revenue was reportedly up, I think the fact that people were not looking at only at the top line, but this is the first time that the stock actually ended up down, even with pretty good numbers, because people talked about the capital expenditures. So people are actually asking about ROI, about investment, what is this actually going to look like, which in the past, we never saw these questions coming up. For real. I mean, the S&P had its worst day since 2022 this week. it had multiple days. I think one day down 2%, another day down one. Usually doesn't happen like that for the S&P. These are significant drops. And the S&P, of course, is made up largely of these big tech companies who've ridden the momentum or on AI to the point where they are. And the

S&P is still up something like, I don't know, 20-something percent year-to-date. Oh, no, it's only up 13 percent year-to-date now after the fall this week. It was up like 18 percent. So this is Only 13% in July. I know. No, it's still a great year, but compared to where it was and where it was going. And I do think that investors are pulling back largely from these big tech companies because they're starting to read the same things that we're reading and see the same things we're seeing and being like, okay, and what do you see?

Like the smaller companies are getting funding. They're also 2000s doing well. S&P 500 or not. Yeah, I think it's finally being asked. I mean, Nvidia is going to lead the market the same way it let it up. And as it's starting to fall, I think there's going to be a lot of follow-on impact from that. And I think it's a healthy thing. I mean, we've talked about this a lot.

For me, again, as a believer in Gen AI and this AI future, things have gotten out of control. Expectations have been inflated. And I would rather that a little bit of air is let out now rather than, you know, we see some kind of total collapse later. So a nice, healthy correction is not the worst thing. I want to talk to you about the Google thing also because, yes, Google stock is down almost 3% this week after it basically beat all of its earnings. Maybe YouTube was earnings expectations.

Maybe YouTube was a little bit slower than expected. But should we take a moment to appreciate how well Google has handled the moment? And maybe some of that was luck because people haven't gone to Bing instead of Google. and that search business remains strong. I mean, this quarter it grew 11. It grew at search business, 11%. Last quarter it built, it grew at 13%. And we're not talking about small numbers, right?

This is from the Wall Street Journal. Alphabet reported Google generated 64.6 billion in ad sales April to June, an increase of 11% in the same year, right? And so, yes, it's spending more, but it has to spend more, and it's good that it's spending more, as opposed to giving up.

but it's growing its business at an incredible clip. It's no longer, you know, sort of a lagger, I would say, in AI. And the competitive pressure from AI is just not hitting it the way that we expected. So is this a moment where we should like stand up and say, holy crap, Google did it? I'm sorry, Sundar. I'm sorry for ever saying you should be fired. I don't think we called for his firing. I just think we discussed it very serious.

We discussed it. We said, should Sundar stay or go? And the fact that that was even in conversation probably six months ago, I agree. They've, even in aggressive measures, I don't know if you saw, like, in part of their deal with Reddit, now only Reddit search results will only be allowed to show up in Google and not other search engines like Duck Duck Go. Like, they're still playing hardball. They're playing aggressive. They are pushing them, whether that's good or not is a different question. but uh look at all the money they're making it's great no no whether it's good for the internet and

society is a different question but it's good for google no i'm kidding no no they do seem thoughtful though they're that they're pushing hard and yeah on in in the whole generative AI space i mean they are like amazon is actually the one i think on the enterprise side a w s and i mean i know your conversation from a couple of weeks ago was actually the first time I heard a clear vision of what they're trying to push and how they're thinking, but I don't know whether they're just avoiding that public conversation. And in one-on-one sales meetings, this is all being communicated. But Google certainly is telling their story, showing numbers on the cloud side, and it looks in a decent space, at least relative to

where they were six months ago. Yeah. I think I'm writing about that this week in big technology, just like Google survived. And it's a big deal. The other, yeah, so but hold on because open AI is coming after them. So this is from a story from the Wall Street Journal. OpenAI is launching a search engine. It's called search GPT, taking direct aim at Google. And this is from the story.

Open AI is launching a test version of its long-awaited search engine, which it says will cite sources of information, including news from business partners like the Wall Street, Journal and the Atlantic tool called Search GPT will summarize the information found on websites, including news sites, and tell users to let users ask follow-up questions, just like they currently can with ChatGPT. It's its most direct challenged yet to Google. Since the release of ChatGPT in 2022, Quot the tech company flat-footed, Google this year widely roll out its own AI search feature that synthesizes information for multiple web sources. And of course,

told people to eat rocks for a minute, but then ultimately that faded. And we haven't heard anything about AI overviews since. So do you think this is a real challenge? I mean, couldn't you have like asked Chechip-T about the news already? No, no, but they would actually, they were very careful and cautious about the types of answers and saying I can only search up to a certain date or I can't find real-time information. I think this could be big. I think this could be a threat. but to me this is almost more of a threat short term to a perplexity yeah again when I I've seen even more and more still when it's like informational I go to perplexity when it's something more kind of analytical or I know is going to be more of an in-depth conversation I go to a clot or chat GPT

if it's something more Gemini actually is not bad in the informational space but I think this was definitely perplexity is the leader in this space, and it could be a huge threat to them if it's done well. But I still wonder, this is not an enterprise product. This is a consumer-facing product. And OpenAI as a company, is this what they really need to be doing? Or do they need to be building out their API and working with companies and building out an enterprise sales team and figuring out how to get people to trust them versus another cool consumer-facing? thing. Maybe both. Maybe both, but I would also say that this is probably tied into what we were talking about before, which is it spending all this money and losing money. And it might

be scrambling now to try to find new business applications and search is a fairly straightforward, I mean, with AI, maybe not, but there are keywords in there. It seems fairly like a straightforward business to add on to like your more experimental businesses. And I'll add one more thing, which is that I am scratching my head and saying, is this a sign that Open AI is fundraising? And I'm sure all these companies are always fundraising, but are they seriously fundraising now? Because you sort of show that, hey, we have a search business. And in your PowerPoint presentations, you say this is the total addressable market of search. And we've now tacked that on to our broader AI market.

And so we have near term and long term revenue potentially coming in. Search GPT. Give us the cash. What do you think? that's not a bad that's not a bad theory because there is part of like the reporting that came out was you know they're burning tons of cash and their runway could be shrinking i think i saw like 12 to 18 months out there wow um so i mean if you're losing 5 billion and i think they've raised 12 13 um obviously it will catch up with you at some point uh so that i i like this theory i think that could be the case you throw out a couple of different business models to add slides on in the fundraising deck and maybe that can help you. I do think, though, when we say search is a good business the way it is for Google, I don't think AI search is a good business yet.

Because again, we've talked about this. No one has figured out the model perplexity said we'll put in some like suggested results and links and stuff like that. But we don't know if this is a good business. We know it's a in-demand business and someone will or could come up with the right business model, but definitely we don't know what that is yet. So another story that I think is really worth mentioning here when it comes to talking about whether these companies are going to have a lead, whether this is going to be commoditized,

right? Because that's what we've been talking about. It's open AI, anthropic, going to be able to survive. Is it big tech? How about this? So Elon Musk said that he's built what he calls the most power. AI training cluster in the world. And we got into it a little bit on the Wednesday show with GitHub, the GitHub CEO.

But I think that, you know, I've thought about a little bit more. And I think that this is a massive news, totally underappreciated. And Elon says he's going to use it to create the world's most powerful AI by every metric by December this year. And it sounds like typical Elon Musk bluster, but this is from Tom's hardware. He's going to call it the Memphis supercomputer. To put the Memphis Super Computer, Supercluster, sorry, Memphis Supercluster, to put the Memphis Supercluster compute resources in some context, going by scale, it's easily outclasses anything in the most recent top 500 list in terms of GPU horsepower. So the world's most powerful computers, such as the frontier, which has 37,000 AMD GPUs, Aurora, which is 60,000 Intel GPUs, and Microsoft Eagle.

14,000 Nvidia H100 GPUs seem to be significantly outgunned by the XAI machine. And we're going to talk about the meta open AI, sorry, the meta open source model fairly soon, but their Lama 3.1 was trained on 16,000 Nvidia H100s. So Elon Musk is putting together a supercomputer that's going to do multiples of that 100,000 GPUs and 100,000 H-100 GPUs. And you put that all together. If this is really a scale game, does he have a chance later this year to create,

by far, the best, most capable large language model in existence? I think he has a chance. I don't think, like, we've talked a lot about this, is scale the only game? And I don't think it is. I think making a bold claim, especially when your main company's earnings aren't coming out well and your stock drops 12% the next day is not unheard of from Elon.

And I think like it sounds exciting, but I genuinely don't know if it is. Because I think it's still, again, is this purely a scale game? What kind of large language model can truly be generated? Because even there's been a lot of talk around, like does Moore's law apply? within for generative AI.

And I think it'll be interesting to see, one, does this really happen? And two, what the actual impact of this would be. I mean, who knows? Maybe XAI does build the world's most powerful LLM and then GROC gets a sweetheart deal and turns into the chat bot that we all use and beats Claude and chat GPT and everyone else.

But as with many other things, Elon, I'll believe it when I see it. Yeah, okay, that is definitely a fair caveat. Yeah, but I'm also surprised this didn't get more run this week, especially, like, think about what would have happened, what would have happened if Open AI or even Meta would have said that they're building a supercomputer. I think, but I think it's a genuine, I think it's a genuine fatigue among the press with Elon, that it's like, this could be exciting, but again, the moment it's real, the moment it's seen,

the moment he's showing it to reporters or showing the output of what they can produce, people will be excited, but otherwise, I think people are just taking things with a grain assault. Yeah, I guess he's earned that. And he's also talked about whether Tesla should invest $5 billion in XAI. Most of his Twitter followers say yes. So maybe that happens. I don't know. I would not be happy with that if I was a Tesla shareholder. That stuff I can't even look at, because it's like from just a capital markets standpoint and just the fact that, this stuff is happening. And again, like as your company, your revenue is slowing, your margins are collapsing. You're talking about stuff. You're throwing up Twitter polls about whether you should

divert more money to like a like internal deal basically is, is bothersome to me. But listeners can understand where I generally stand on at least the kind of dealing, business dealings of Elon. Yeah, Tesla, even with a run up in the most recent month, is down 11% on the year, and SMP 500 is up 13%. Okay, one more bit of AI news. You know, we're in this kind of moment of we're not quite at disillusionment yet. We're not quite at that trough, but we seem to be peering over the edge. But that being said, we're in this moment where we're going to have, I think, less and

less enthusiasm about AI coming from the public, but also more and more impressive breakthroughs. And I'll put this development from Deep Mine in the latter category. right, an impressive breakthrough, where they have created a model. It's called Alpha Proof, and it specializes in math reasoning. And they also have this new one called Alpha Geometry 2, which focuses on geometry. And this is from Bloomberg. The program aged four of the six problems featured in the International Math Olympiad, an annual competition in which students tackled topics such as algebra and geometry.

Okay, here's why I think this is significant. I think there is a reasoning element here, which is that like with a typical large language model, you ask something and it just like combs its massive amount of information and then spits back an answer. With these math problems, you actually have to work through step by step. That's how it works.

And the fact that this thing was able to answer four of six questions correct is, I would say, extremely impressive. What do you think? I think it's definitely impressive. I think, like, again, that barrier between showing genuine reasoning and where we are today, where it's clear that LLMs don't actually think they're basically analyzing and, you know, just predicting next letter, next word, I think if that could be shown. We've talked about this a lot, agentic AI in building AIs that can actually, you know,

perform a network of tasks and be able to reason at every step of the way without having a predefined choice or behavior given to them, if that gets solved, it really will be massive because that will be the kind of like step change in terms of how effective this could be at solving problems. So I think this stuff is interesting. I kind of like how you said. To me, I'm actually excited by the fact that overall enthusiasm is going down while real improvement is happening. Honestly, that's where I want Gen A.I. to be right now. You're a trough man. That's really where you stand. I'm a trough guy. I'm a trough guy. The trough is where we do our best. That's true. The trough is where we're all at our best. Respect the trough. Respect the trough.

Respect the trough. We should get to hats. I'd say respect the trough. Again, for people unfamiliar, trough of disillusionment is part of the Gartner hype cycle. Get your Gartner right, people. Yeah. All right. Let's let's talk about the meta open source model and Southwest Airlines cancelling or ditching open seating right after this, which will take a quick break. And we'll be back momentarily. Hey, everyone. Let me tell you about The Hustle Daily Show, a podcast filled with business, tech news, and original stories to keep you in the loop on what's trending. More than two million professionals read The Hustle's daily email for its irreverent and informative takes on

business and tech news. Now, they have a daily podcast called The Hustle Daily Show, where their team writers break down the biggest business headlines in 15 minutes or less and explain why you should care about them. So search for The Hustled Daily Show and your favorite podcast app like the one you're using right now. And we're back here on Big Technology Podcast Friday edition, breaking down the week's news. I'm here with Ron John Roy of margins. We're going to talk a little bit about this meta open source models. This is from the Verge. Meta releases the biggest and best open source AI model yet. It's called Lama 3.1. It's the largest ever. open source AI model. It is bitrained, as we talked about, on 1600, NVIDIA, H100 GPUs.

Verde says meta isn't disclosing the cost, but it's safe to guess. It's in the hundreds of millions of dollars. It has a very big context window, 128,000 tokens. It has support in English, Spanish, Portuguese, Italian, German, Thai, French, Hindi, and it's available on AWS Azure, Google, Oracle, and then there's companies like Scale AI, Dell, and Deloitte that are going to implement it. So basically, the argument here is that this is a frontier model. It's one of the most capable AI models. It's open source. It's largely free to use. And meta is releasing it. And they are basically saying we're committed to open source. So you might be paying, you know, open AI per token when you use its models, but you'll pay us nothing and you can customize it.

you want. Talk a little bit about the significance of this move, Ron Chan, and where do you think this leaves the AI industry? Where do you think it leaves meta? I think this is a huge deal. I think this is kind of like masterclass, brutal Zuckerberg at his best. Because what's happening here in my read is choosing open source is their path forward. And I want to get into the blog post Zuckerberg wrote about, called Open Sources the Path Forward. But basically, meta realized that trying to compete for like consumer or enterprise dollars when you don't have a consumer subscription business, when you don't have an enterprise cloud business would be a massive uphill battle for them against everyone already in the market.

So instead, go open source, undercut everyone, make this thing free, And the fact that Lama 3.1 is outperforming GPT4 in a lot of cases, it's as good as anything else out there. And it is free, it's just a massive deal. And it's going to, especially at that enterprise level, I think, pose a genuine threat. If people start thinking about things that I actually have an open source route to this, especially if it's more customizable, if I have more control over it, that really is a shot across the bow of, everyone, whether, I mean, for open AI, it's a huge problem. But even Google, Microsoft, Amazon, it's a problem as well. That's right. And I think that like there's been some discussion of like

meta's open sourcing this model because it wants to sort of mess with Google or mess with Amazon. It can do it for free. But it's also very expensive. And Zuckerberg in this post, and they've always been pro open source, but he talked about why it was so necessary for them and did it in like the most direct language possible, which I thought was like very interesting and worth, worth sharing. So he says, one of my formative experiences has been building our services constrained by what Apple will let us build on their platforms. Between the way they tax developers, the arbitrary rules they apply, and all the products, product innovations they block from shipping, it's clear that meta and many other companies would be freed up to build much

better services for people. If we could build the best versions of our products and competitors, we're not able to constrain what we could build. On a philosophical level, this is a major reason why I believe so strongly in building open ecosystems and AI and AR VR for the next generation of computing. So basically he's saying, I don't want to be approved by any other platform. I don't want to be relying on any other platform to build. And I am going to just build the underlying technology and open source it and everybody in the open source movement will help me improve it and gatekeepers be damped and i think that this is pretty interesting all right this is this one got me pretty hard and first thing i'm going to say is zuck's public PR image has been

pretty amazingly transformed from sitting in front of congress pale and like nervous to wakeboarding and wearing like a t-shirt and a chain and looking cool and like hot zuck memes going across the internet. So that that PR makeover has been successful. But I honestly think this kind of like developer PR makeover, no one's talking about and might be the bigger story. Because when I was reading this, I was like, okay, some of the things he has in headlines, why open source AI is good for developers. We need to protect our data. This is coming from Facebook. He says, we need to control our destiny and not get locked into a closed vendor. Anyone who is an advertiser understands that meta is a closed ecosystem. And we want to invest in a ecosystem that's a standard

for the long term. And the quote you used around our formative experiences were being constrained by Apple. Anyone who was in tech in the late 2000s early 2010s, if you remember Farmville and the all the Facebook apps and the thriving developer ecosystem that they built with the promise that you will all get to build and do, you know, Zinga going public. Like Facebook was the ecosystem and then they just shut it all off. They just cut off everything, decided that they didn't want it. it was the most arbitrary, I mean, for them good, but for all their entire ecosystem bad. So the idea that suddenly you have been pained by Apple,

but you care so much about open ecosystems when you run meta, I think is a bit rich. But this is different though. Yeah, sorry, go ahead, caveat and then I'll. Well, last thing, but talking to a number of friends or developers, they're like bought in. They're like, oh my God, did you read this from Zuck? This is, it's beautiful.

This is what I believe, too. And so people, so it's working. And that's why, you know, we talk about credit to his, like, cool guy PR team. Credit to whoever's writing these blog posts and who's coming up with this story, because it's working. Well, here's what I. I think that it's different. And this is maybe a maturation because he really has been messed over by Apple. Apple shut down their internal app.

People couldn't like book conference rooms for a while because of, Apple's anger with them. And of course, you know, Apple is a nemesis of Facebook. So I would just say that, like, I think he's generally angry about this. And they've also given up the ecosystem play within Facebook. They've made a lot of changes to Facebook and Instagram over time. Now, there's still a walled garden. I won't argue that. But they're a different walled garden than they used to be. They're not really all about developers building experiences there because they realize that the platforms are not going to be the entire internet. They're going to, they're specifically used for a certain thing. And when it comes to AI,

in particular, if you open source it, you know, it's not going to be everybody living within the meta ecosystem. The idea is, here's the underlying technology and you build on top. I get that argument. I get the idea that, again, they are a different ecosystem than they were before. To me, still, you know, you can't forget that that did happen. But I still think that the idea that like, it's this, I don't know, the idea that AI is something significantly different, If at any point it is in Facebook's existential business interest to then close off the entire ecosystem and somehow change everything around and try to actually bring everything in house and try to screw over developers, I think they would in a heartbeat.

And I think this is, I don't know, the idea that this is like a genuine vision versus I think it's a brilliant cutthroat business move to go after all your competitors. editors in a way that I think is could work like could work really well you take your money from an Instagram and Facebook or their advertising is back people are spending tons of money on it you're still you have you what is it is like 2.2 billion people use their products every day so I think you're in a pretty good position funnel reality labs lost 3.4 billion dollars last quarter you can afford that. You can build some open source tools and then kneecap your competitor. So I still think

that's what it is. And I think it was a good blog post. I just don't buy it. Okay. So I can tell you I'm fairly certain that Zuck is freaked out by the consumer adoption of AI tools, or at least was in the beginning, which has led him to build these AI tools, but that's also tailed off. So I do think that maybe you're right. Maybe part of this is just like him seeing a competitive threat from Sam Altman, who he did Needle in interviews talking about how he's building closed models as open source AI and open AI. And Elon Musk's like, yeah, Zuck, you tell them. And yeah, go ahead. No, this is, I enjoy watching this. This is like business battle at its finest. And this is where Zuck is in the cage match that we never got someone to be feared. Because I think like the

way they've approached this entire thing is brilliant. And I agree. I think, yeah, they probably, they sat down, they saw a threat and they had to decide how, is this going to impact our core business? Is it going to affect messaging, social networking, photo posting? And I think they realized it wouldn't that TikTok was the only real threat there. So then, okay, what do we do about it? It could at some point pose a threat. So let's, let's mess with their business. Yeah. Okay, one more thing about this and then we're going to talk about southwest briefly the china thing uh i know for a fact that like everyone is freaked out like not everyone but a lot of people within the tech industry are freaked out at the fact that meta has open sourced llama and a lot of the chinese

a i companies are building on top of it and learning from it and may eventually exceed it and you think about the amount of computer that's going to be available in that country if they they want to pool their resources, they can have some central planning that can get them into a really good place. And this is what Zuckerberg says about China. Our adversaries are great at espionage. Stealing models that fit on a thumb drive is relatively easy, and most tech companies are far from operating in a way that would make this more difficult. It seems most likely that a world of only closed models results in a small number of big companies, plus our geopolitical adversaries, having access to leading models while startups, universities, and small

businesses miss out on opportunities plus constraining american innovation to close development increases the chance that we won't lead at all and i read that i'm like okay good argument but we'll see like i that that is a great argument i'm telling you whoever he has media training him on these answers i don't i think this is actually zuck writing this stuff but go go ahead and nickleg and it never yeah nickleg okay maybe it's nickleg but it's more involved than i think that you're you're you're realizing he's very hands-on in that company and especially with messaging i think he's involved but i come on that that is so good that if he if he has the time to write this i'm worried about him not taking care of the rest of the business because what a brilliant answer it's like

probably i would imagine it's nick clegg sorry go ahead yeah yeah we're i mean it's about small businesses universities like you know the it's about the little guy and helping them um yeah I think they're doing well on the communications front right now. And if Lama, I'm going to try to experiment with it soon. If it gets interesting, it could completely change the economics of even how I'm using these things. Yeah. Oh, I got to give you this opportunity to talk about your cloud experience. Speaking of how I'm using these things, I got to say, so Claude, the other day I was doing a revenue scenario analysis.

in Excel and something that would have involved hours of building a model. I was like, you know what? And it's taking like a funnel in e-commerce. And I was like, let me just put this into Cloud and say, hey, Claude, I'd like to build an app that allows me to take these numbers and do revenue scenario planning. So it'd be like, I want to change the headline revenue number. I want to change different steps in the funnel.

And again, normally you'd have to build a complex, find it like model around this. And instead, Claude's like, let's build an app together in React. And so it starts giving me code. And I start testing it within the cloud. They have this new window called artifacts. And I'm like in there testing it, saying, actually, I think this is a bug. You're not doing this right. We're improving the code.

Then for me, the magical moment. And I will expose myself a bit that even covering tech and being very involved in tech. on the development side I've always known basic HTML and CSS and can do some things and I'm somewhat comfortable but anytime I get into the command line and I have to deploy any kind of code or like deal with GitHub repositories I would just get stuck and this has gone on for years I would just be like all right I don't have the latest version of Python installed I don't know how to do this this is getting too complicated I literally would just give up it just walks you through step by step

copy this copy this copy this and then even so I'm testing things something something breaks I upload the code base to Claude

Claude finds oh you added an extra bracket on line 233 no let me update that for you here is your new code copy this into this file

upload it and then I'm like okay the actual core of the app is working then I'm like let's style this here's a color scheme can you add a button here that says this then it writes the CSS file like it's doing all of

this and then i deploy it and it's live and i can even share a live internet link with other people like i've made an app from in 90 minutes just playing around with this and again super basic simple thing but suddenly i'm i personally i'm like i want to build apps i want to test this stuff things that i never would have wanted to do before and i would always just be like all right i'm going to pass this off to someone more technical, and that to me, I'm going to say, I've been trying to document my wow moments. Right now, I think I have first time using chat GPT, first time using mid-jorney, first time using Suno for music. I'm going to put this up there as my fourth most magical generative AI moment where you're like, this changes everything. This really, to make people, to democratize the ability to actually

not just code but to actually build things is completely different especially if you want to just experiment build small experiences and I think I heard the GitHub CEO saying like he built an app that only tracks all of the flights he's ever taken

so something that no one would ever want to build and for him it was in the time to drink a glass of wine for me it was 90 minutes here he is the CEO of GitHub but but I think like this is it was crazy for me it was it was fun it was exciting so i'm the why i hope anthropic doesn't go out of business and clot is here to stay so how do you think this changes everything you said it changes

everything yeah no it's to me so chaty pt what it changes is anyone who could not string together words anyone who is a non-native english language speaker can now market in english can now have business communications in English, like, like the, it just completely transforms what skills are necessary at a basic level. And anyone for the same thing with images, the way you can now create, you can run an entire e-commerce business, you can do anything around image generation. Now with code, especially, and not just code, because GitHub co-pilot I tried to use, but it's so developer focused. Now you have someone walking you through how to build something. So if you want to anyone like I already was thinking about what stupid little things do I want to have an app for

that I can actually run on my phone that I do I you know like what weird little you can make apps for your family you can make experiences you can make just like cool digital things and then so that's obviously just making cool little toys but then people actually getting into launching businesses it really was that next layer of okay this really makes it more accessible for for anyone anywhere. And again, that idea you had spoken of with the GitHub CEO of could you have a one person billion dollar business doing everything, it doesn't seem that far fetched to me,

given how these tools give you like supercharge you doing things that you never would have been able to do otherwise. Amazing. Yeah. I'm sure we're going to hear more examples of this. All right. So finally, let's talk about the Southwest airline situation where they were hit, I think, in this crowd strike,

thing, which has been like this sort of shock moment for so many airlines, like Delta said, they're going to do away with unaccompanied miners being allowed to fly on their planes. And now Southwest is like, yeah, screw it, we're going to be a normal airline. And this is from the journal. Southwest Airlines is ditching, open seating on flights. It will soon aside in seats on flights and sell some with extra legroom, making sweeping changes in a bid to broaden its appeal to passengers and boost revenue. the push comes as it fends off an activist investor looking for an overhaul of its leadership

and operating strategy. And of course, that investor is Elliott Management, which effectively pushed Jack Dorsey out of Twitter. And this is from the journal Open Seating has been a hallmark of Southwest flights since the airline got its start more than 50 years ago as part of a business model that produced decades of uninterrupted profits and democratized flying in the process. But Southwest executives said the company needs to adapt to what today his customers want, making what might be its biggest ever shift. Just a quick rant for me. I saw someone say, listen, people don't want to fly on a bus with wings. They want to fly on an airline. I'll just say my, you know, okay, I'll just say, knowing that Southwest Airlines is going

to do away with open seating plans just lights up my heart and makes me so happy. I've flown the airline exactly once and was boarding towards the end. And it was the weirdest situation I've ever experienced on a plane where as you went on, you looked and the rose just filled with one person on each side, right, aisle and window. And then the middle seats were just left open all throughout the flight. And basically the people coming in had to come in and just file in a middle seat, middle seat, middle seat. I think it gave you the illusion that you were in control, but you weren't. And I found it to be a miserable experience. So rest in peace, the open seating plan of Southwest Airlines,

you are not loved in life and you will not be missed in death. I agree with you on that. I think open seating on flights was another Southwest gimmick that I think it's time to move on. And hopefully for anyone out there who has not enjoyed the service of airlines, especially Delta customers, hopefully things start to improve pretty soon.

They won't. Not if Pete Buttigieg, or transport secretary has something to say about it. Well, yeah, I may be even still. I'm flying in a couple of weeks, and I'm already, you know, nervous of the delays or what sort of cancellations might arise, but such as the peril of flying in the summer.

All right, Ron John, welcome back to the U.S. Great speaking with you, as always, and we'll do it again next week. See you next week. All right, everybody, thanks so much for listening, and we'll see you next time on Big Technology Podcast.