Big Technology Podcast - OpenAI’s $6.6 Billion Fundraise, Meta’s Quiet Election, Imaginary AI Friends

Episode Date: October 4, 2024

Ranjan Roy from Margins is back for our weekly discussion of the latest tech news. We cover 1) OpenAI closing a $6.6 billion fundraise 2) A look at the investors, including Softbank, Tiger Global, Cathie Wood, and Altimeter 3) Will OpenAI need to raise again shortly? 4) Is an IPO next? 5) Investors can get their money back if OpenAI doesn't convert to a standard for profit company 6) What's next for OpenAI's relationship with Microsoft? 7) Meta's quiet election 8) Is social media's collision course with politics sad? 9) AI girlfriend apps on the rise 10) AI slop has taken over the internet 11) Ranjan's weird Facebook experience 12) Depression and bewilderment --- 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]

Open AI just raised $6.6 billion at a $157 billion valuation, the largest VC funding round in history. Meta's had a remarkably quiet election and our imaginary AI friends, now the future of social apps and perhaps our lives that's coming up right after this. Welcome to Big Technology Podcast Friday edition where we break down the news in our traditional cool-headed and nuanced format. The Open AI fundraising is finally in and we have a bunch of other. fun stories to speak with you about joining me as always on Friday is Ron John Roy of margins. Ron John, welcome. Great to see you. Good to see you. Open AI did it. They did it. Yes. So it finally happened. They've raised $6.6 billion in new funding. The investors, it doesn't include Apple. It's just a billion from Microsoft. But they have SoftBank involved. They have Thrive Capital.

Invidia is also investing $100 million. So, and they also have Tiger Global. We can't forget. Them and Kathy Wood, of course, is in, and so is Altimeter Capital. So that's how the round has shaped up. Any idea that they had potentially been struggling to fundraise, which is something that I admit I've brought up on the show, that's gone to bed as the largest VC fundraising in history. So I will admit I got it wrong that there would be some trouble, of course, hottest name, probably in tech in a decade. They figured it out, but it did come from some untraditional sources. shall we say, the big names like Sequoia and Dreson Horowitz didn't participate. Rontan, what do you make of the final shape of this fundraising?

This is exactly what I was hoping for. And I admit myself, I think both of us have been throwing a little bit of cold water on this fundraising round and saying maybe if they did not raise this, and this would be the moment, the moment, the inflection point for generative AI, especially. given all the other questioning that's coming out about the technology and its adoption. But it happened. And boy, did it happen. I mean, this is like the megalopolis of funding rounds right now in my mind because it's not just huge and ambitious, but it's also, I don't know, it's ridiculous. It's Tiger Global putting in $350 million. Kathy Wood putting in 250 mil, altimeter putting in 250 million. It's every like high growth, high momentum, late stage fundraising name that I've written plenty about, you've written plenty about that dominated

the headlines in 2021. They're all in it. Even the UAE's new state-backed sovereign wealth funds in it. It's like the who's who of who bets big late stage. But I mean, now even more so, I think this is, everyone is going to have to watch what happens right now. Okay, so there's plenty of details involved here that we're going to get to, including Open AI asking these investors not to invest in competitors and a potential chance for them to get their money back if certain conditions aren't met. But I need to pause here and say, Ron Jen, you said this is the round you've been waiting for. It seems like all of these names that you're listing are like red emoji flag for you in terms of things that you're worried about. I know, but this is, okay, for the last few years, you know, when Tiger and Ultimiter

and certainly Arc and Kathy Woods Fund, you know, when they kind of created this whole late stage momentum style of fundraising that was completely in vogue for a couple of years that seems to have quieted down, this is back and in such a big way of betting big and almost willing these kind of IPOs into existence and these valuations into existence simply by the value of putting so much money into it, that I thought this was kind of going away, but now they're coming back and saying, we're going to go as big as ever. If this fails, then I'm almost happy, because I've always felt this is a somewhat problematic mode of funding, and it's not good for overall capital markets. So the reason I've been waiting for it is

this is the biggest bet on this style. Wait, wait, sorry, did we even mention SoftBank? My man, Masayoshi's son is here. He's in. Like, everyone, this is like the Avengers of this kind of funding, bringing everyone together in the most absurd, massive funding round ever. That's why I'm excited about this one. What does it say about OpenAI that they had to resort to this type of fundraising?

I mean, the truth is, it was either going to be this type of fundraising or, you know, a big fundraise from tech giants like Microsoft, right? Like, you're not going to necessarily, you cannot power this round with A16Z money, can you? Well, I mean, at least having A16Z and Sequoia as part of it, I think they could definitely start getting towards this kind of number, but I think it's still so telling it. That's why this makes this the most interesting fundraise in my memory is because they don't have. Again, remember, Sequoia, A16Z, all of them have been raising large, at least billion-dollar late-stage growth funds. This is exactly the type of investing they have been pushing towards, away from just being seed and early-stage specialists.

This is where they're supposed to be, and they're not. And the names coming in are Masayoshi-son, Kathy Wood, names I would never want to bet the future of my company on. But to raise this kind of money, I mean, I agree that to get this volume of capital, at this stage, given the type of company Open AI is, they did not have other places to go. They had to go for this kind of money. And we always talked about how it was going to be a difficult fundraise. Like there were going to be some fundamental problems here, including the fact that they were not yet converted to a for-profit, including the fact that they have the weird governance.

I mean, let's not even, we got to get into that as well, but just to start. And let's just start with the fact that, okay, they're raising this $6.6 billion, but when are they going to need to raise again, right? Because this is, again, coming from Ed Citron, who's been again tallying this up. We talked about the subprime AI crisis a couple weeks back. The company's expected to lose $4 or $5 billion a year. And it's going to raise $6 billion. So what is its runway? Like 14 months? So is it just going to be in this perpetual cycle of needing to go out and raise these type of sums? And if you're so interested in whether this bet is going to work, are we going to get the answer, like in somewhere, sometime in mid or late

2025? No, I agree with you. I think so the timelines are really worth kind of going over in detail. As you said, losing potentially five plus billion dollars a year, I mean, and they're actually saying that they're going to potentially invest more in frontier models, which is the single most expensive thing you can do. And let's not forget that they haven't really built out their entire actual business enterprise infrastructure yet, which is a very expensive thing for anyone to do. So if you think the runway is around 12 to 14 months and which it could very well be, but then they have two years to convert themselves into a for-profit entity, which is already, let's recognize so weird that the biggest VC funding round of all time is for a non-profit that is promised to, that is promised

to convert itself. And if it doesn't, the investors will be able to collect a 9% interest rate on their money. So basically, they invested, if it's 16 months from now or two years from now, and Open AI has not converted, you get your money back at a 9% interest rate, which doesn't make any sense because how would that money still be there in two years? I mean, almost by definition, they're going to have a burn rate greater than their revenue and be losing money. So it just is so odd and interesting that especially those timelines, they just don't match up in my mind. So basically, they get to this end of this timeline. They're not converted. Those companies aren't getting that money back. No, of course not. I mean, there's simply no, unless, like,

There's no capital left. Where in the world is that, I mean, maybe the Norwegian sovereign wealth fund or like, I mean, I just don't know who. Open AI, the nonprofit is crushing it and decides it wants to stay there and just say, all right, we give you your interest. We move on. Yeah. Well, no, but where, what's the, what capital is left? There's none in the world. Well, they're going to have to, it's crazy to say it, but they're going to end up having to be profitable at some point.

Or they're just going to IPO unprofitable. which many tiger global soft bank altimeter like you've arc investment management these are the companies that perfected the art of bringing unprofitable companies to the public markets passing it off to the public investors at large and watching the stocks collapse so they've they've gotten pretty good at that uh that game can it work at this scale that's a different question is this funding around also just the i don't want to say end but a new a really important shift in the relationship between Microsoft and open AI now Microsoft invested at 13 billion 10 or 13 billion in open AI and basically gained a

trillion in market cap because of it somewhere in that range and now the new round comes up and they're investing little less than a billion dollars and letting others lead the round while building their own capabilities internally now it's kind of weird to say all right they're going to invest nine hundred something million dollars in the biggest fundraising biggest fundraising round of all time and sort of distance themselves from the company but it just like reading the vibes

that's sure like what it feels like to me I think yeah I think they are I mean again they just had a massive co-pilot launch event where they were kind of highlighting all the different features that they're integrating into Office 365 and across the Microsoft ecosystem so they are clearly embedding on their

own horses right now. I also, one thing I've been wondering is do the cloud credits that were investing, you know, does that actually somehow still multiply in 10x and what you've already made when you invested at 13 billion or 12x now it's at 157 billion? Like there's so many layers to the existing funding that's gone into the company and even the current round that are so unconventional that it just who makes money where and when is still unclear to me from every all the reporting well do you think Elon should get money uh if he helped found the thing with 50 million dollars and the IPO shouldn't get an ownership stake i mean i say yes actually i say yes too just throw him in there because i want him next to tiger and masayoshi son

can you imagine that what the scene would be like if all those characters ringing the bell at nasdaq On opening day. I mean, you know what? That almost has me rooting for the success of an IPO. If you line everyone up, take them all out to dinner, bring them the next morning, ring the bell. That would make my year, I think, just for the entertainment value. So there's another aspect to this also, which is that opening I really is. Like we talk about, you know, are they going to be able to pay that money back with interest?

It's not going to happen because they really are going to convert themselves to this. traditional for-profit or maybe some sort of B-corpor or whatever it is, they are now starting to act in less like change humanity for good and more like, again, ruthless cut-throw company where they're asking these investors not to invest in competitors. That includes super safe super intelligence, which is Ilya's company, which is unbelievable, anthropic XAI. What do you make of the fact? I mean, isn't that kind of like a show of weakness that, You're asking these groups not to invest in competitors?

I actually, oddly enough, was not as bothered by this. And apparently Uber had done this very aggressively in the past and, you know, especially in the helping kneecap lift a bit. The reason I'm actually not against this is this is more how traditional VC worked in my mind. Like in your portfolio, you did not, or you rarely invested in direct competitors. you kind of bet on which you thought was the winning company. So the idea that they're saying, you know, invest in us, but don't invest and try to, you know, fund competitors that could potentially ruin this investment seems actually kind of like

a reasonable thing. In this case, I mean, it feels a bit petty given it's all kind of the same group of people for the most part. All open AI co-founders. These are all open AI co-founders. Elon Musk, open AI co-founders. Elon Musk, open AI co-founders. under Ilius Sutskever, Open AI, you know, chief scientist, Dario Modi, the guy who built the GPT models.

Like, it's basically, the people who have been with an open AI, don't give them any more money. And they know that they're going to need the money because that's what it takes to scale these models. Yeah, no, no, it feels a bit personal, especially, yeah, and I think the founder of Anthropic was even saying just how expensive, again, the next generation of frontier models is going to be. So I think it, to me, again, from just a traditional capital markets perspective, I actually don't find it that ridiculous. But you can only imagine, given these are people who are close to each other, that this stuff is personal. So the information also has a little bit more information about what opening has been telling investors. And I wanted to get your take on this, Ranjan. So this is the revenue forecast.

So they think next year the revenue is going to triple to 11. point six billion triple and then double again to 25.6 billion the year after. Do you think reaching that level of revenue is feasible for this company? What? It's a tough question. I think because we still don't know. I genuinely believe again, as someone who is bullish generative AI, we don't know what the market, total addressable market looks like and how it's going to play out. Like, we know general things around, you know, like cloud spend and software spend. And if generative AI kills all of SaaS, then obviously all that spend gets moved over to one of these companies.

So I actually don't think that the idea that a company like Open AI could or should make $26 billion in two years is actually that crazy. Whether Open AI is the company that's in a position to do it and, like, execute on that, I think is, it's pretty ambitious. And again, it keeps coming back to the idea that, and we've talked about it a lot, they are not built and run like a traditional enterprise software company. They don't have the salespeople and the customer success people and just that entire infrastructure. It's a research house with an API on top of it and a great consumer product, which those things will not make that much money.

It won't be the consumer product because we've had this tweet in our doc for a couple of weeks now and time to read it. So to Nehai Pura, who's a former meta employee, shares great stuff on X. He said Anthropics is expected, and it's from CNBC,

Anthropic is expected to hit $1 billion in revenue this year. And that's a 1,000% increase year over year. And this is the breakdown of their revenue. Third party API through Amazon is 60 to 75% Direct API is 10 to 25%. And then Claude subscription is 15%.

Now, I expect OpenAIs chat GPT to have a much higher percentage or at least a higher percentage of chat GPT subscriptions than Anthropic because Claude is much less known except to listeners of this show. That being said, it is going to have to come through this API, right? It's going to be the enterprise that's going to be spending triple, quad, you know, double, whatever it is and not the consumers. Is that the right read? What do you think?

I think this is actually an incredibly important thing. There's even one part of it that I thought more interesting than obviously at the high level, more revenue coming from your enterprise or more business-focused API calls versus consumer-facing subscription revenue is going to be a lot stickier and just a more promising business model. What was really interesting to me is that they specified via Amazon, that meant third-party API. That meant their model being used in other cloud infrastructure and not developers where direct API is only 10 to 25%. So the majority of their API usage is being used in other cloud

infrastructure, which if we remember Amazon, who we don't hear very much in this whole conversation about, they originally, the promise was you can try all these models very easily in AWS infrastructure. So it looks like the Anthropic models are actually very active in that space, which is probably the smartest place to be versus people going directly to you. So I think the big battleground could actually be in AWS in Google Cloud. And it looks like Anthropic is doing pretty well there. Yes. Okay.

Let's close with a couple of quotes from Brad Gersoner, the altimeter investor who was at the Geekwire conference this week and was talking about the investment. First of all, about Open AI's ability to ship amid the chaos. Look in the middle of all this chaos in nine months at OpenAI, and what's happened? The numbers have accelerated. They are launching new products. That's not a sign of fragility. What do you think about that?

Yeah, I agree with him there. I think the corporate drama has not been the issue that we've all been worried about. They keep shipping. They're growing. Whether it's going to realize $157 billion valuation, I have significant. questions about. But overall, I think they have shown, to their credit, that amidst all these departures and dramas, they're still shipping, growing, and raising money. Now, this is the most revealing comment that Gersoner had to me during this conference. He said that Microsoft's co-pilot

as a consumer brand is going to have a hell of an uphill battle against ChachyPT. And he said, the good news for Saty Unidale is that he owns a lot of Chachipat. That is not the language I would use to describe a partnership that they're going to have to battle each other and it's good that he owns a good chunk of it. Yeah, I really think the Microsoft connection to OpenAI is effectively over with this fundraise. I think we've all been seeing the writing on the wall for a while and I think now it's clear that they are direct competitors and Microsoft has been very smart about kind of managing their potential financial returns on this, but otherwise they're direct competitors.

they're not friends anymore. Maybe they're getting discounted cloud credit. Actually, wait, because in Ed Zetron's piece, he had talked about in the numbers he had kind of calculated, how that Open AI is still getting effectively, like paying 33% typical cloud costs when they're using Microsoft Azure. Imagine if one day Satya says, by the way.

That's true. But that being said, the amount of money that Microsoft put into OpenAI protects Open AI, I think, more than it protects Microsoft, because they don't want that to go to zero. I mean, if they, their future is co-pilot, right? Like, in generative AI, that's how they're competing not against Open AI, which is huge in the startup context, but is still tiny in the big tech context. So like, if Microsoft ever started worrying about Open AI competing and actually hurting their

generative AI. future that is going to be you know push them against google and amazon and meta i think they could just turn that dial and say it's time to pay full price yeah but copilot is powering uh but oh sorry but gpt4 like open a i's technology is powering copilot at least for now so at least for now for now but it cannot be they're not betting the future of well they know they have to insulate themselves I'm going to say this. They just invested nearly a billion dollars in Open AI reupping. I think to me it's not the end of the relationship, but more of a careful distance.

And if you're Saty and Adela, that's smart. And if you're Sam Altman, maybe that's smart too. So it gives you an opportunity to go and play with Tim Cook's products and see what you can do on that front. All right. That's fair. That's what it's all about. So we had the vice presidential debate this week. We're not going to talk very much about that.

But one of the things I was thinking about, one of the things I've been reading about, and one of the things I'd be writing about this week is that, isn't it interesting that we're legitimately one month, I was going to say four weeks, give or take, we're one month away from an election. The election, the U.S. presidential election is on November 5th, 2024. Today we're recording October 4th, 2024, one month and one day away from the election. And we're not talking about Facebook at all. Mark Zuckerberg is not donating anything. He has a different company names called Meta now.

We're talking about their tech products like Orion. Instagram and threads have de-emphasized politics, and they have effectively ceded the game to TikTok and X. I personally think that this is a great move for Facebook and that politics and social media just don't mesh. And that there was this original dream that if we connect with people and learn a little bit more about who they are, then therefore we're going to start to see them more as people, less as caricatures, and then we'll have a more

thoughtful politics discourse. But the opposite has happened, right? The loudest, most outrage baity voices have risen to the tops, to the top. And politicians have decided that the most divisive content is going to be the content that will play the most on social media platforms. And here we stand ahead of the 2024 election. And Facebook is just like, Yeah, we don't want anything of it. We recognize what happened. It was a mistake. We would rather you take a look at shrimp Jesus, which we're going to talk about in the next segment,

as opposed to anything that these candidates are saying, go do politics somewhere else. I think that's a good thing. I'm curious what your take is on it. Yeah, this is such an interesting area to me because this is something I've thought and written about a lot. Meadows' role, especially after the 2016 election in 2020 around misand-relevant. information in general and the whole topic with COVID and vaccines. So for them to have just quietly moved away and us not thinking or hearing about them at all is pretty incredible.

And I think there's two things that they did in a very smart way. I genuinely think the Mark Zuckerberg rebrand worked of the kind of like the cool guy with his chain and they all wear matching chains now. And, you know, I mean, they've pulled this off where he's not some, like, the caricature of kind of a world dominating, terrifying figure. He's just like the cool guy with the next generation of augmented reality glasses. But the other thing is, from a technical standpoint, they have done this. They've actually done this. And my Instagram feed is now, like, Instagram Reels, the way they've approached Instagram Reels to compete with.

TikTok has been fairly genius. Like right now, reels comprise, I think I saw it was 46% of all Instagram feed content and 50% of that is not even from accounts you follow, but suggested for you. So it's because, and I'll admit in my social circles, people just share reels nonstop or millennials. And they've done this very well. And it's not political content. It's all just kind of pure entertainment content. And so they've made the algorithm work in that way. And I think that that's a success for them, the fact that they're still growing without like relying on needing to be the center of politics.

The only thing that I worry about for them in this, they don't drive culture anymore. Like Instagram and Facebook are no longer the drivers of culture. X is for politics and any kind of like, you know, more. reactionary type culture. And then TikTok definitely is more the driver of culture and trends for Gen Z or even products. And Instagram is the follower. Whether that in the long run could actually hurt them, I don't know. But right now, they have really found this kind of Goldilocks spot of growing, getting more and more engagement, but still being this kind of light entertaining content. Yeah, it's definitely like, this is the question is the relevance question because it is a risk.

right you become less relevant like we all we've talked about multiple times how threads is less relevant however like there are you know if there's anything we've learned over the past few years is that there are formats that just do not work with social media like it is the it incentivizes the opposite of thoughtful conversation on these platforms and them moving away from it has been good they have started this is according to the times they've recommended more content about like you're talking about sports, cooking, animals, celebrity gossip, and they've left all that other stuff to Elon Musk. And that being said, like, this is like one of the parts of the Times article that I disagreed with. So the Times was basically like, even still, politics is on Facebook

and the company can't get away from it. And it's like, yes, of course politics is going to be part of any social network. You can't prohibit people from talking about politics. They're going to do it in groups, they're going to do it wherever. But you can decide not to recommend it. And if you shrink the total pie, I think you're doing, you're doing a good thing. So I wouldn't paint it in this negative lens. No, no, I agree. I think like WhatsApp groups are probably still raging with, or even Facebook groups are probably raging. But what they did smartly is, as you said, like a stop elevating or pushing by algorithm that kind of content and focusing on Moodang. And they, I mean, probably that's the culture.

The pygmy hippo for those unfamiliar. Everybody knows Moodying for those unfamiliar. All right, I'm just checking. I'm just checking. It was a good S&L skit from Bowen, Yang, if he's happened to see that one. But, no, no, I mean, like, that's the culture that they're creating. And they're probably very, very happy with it. And especially going into this election, I think they are in a very, very sweet spot right now.

the only risk becomes, well, actually, when we'll get into the AI slop risk, but just as a starting point, no longer driving culture, because I still think social platforms succeed the most when they're the ones that people go to because they're like, what's actually happening now? What do I need to know now to feel cool and informed? And that's where X and TikTok are still dominating. But the question is, what type of culture do you want to drive? And this, I might end my piece this way. the eating the dogs and eating the cats thing that Trump said on stage in the debate about immigrants eating dogs and cats and pets in Springfield, Ohio, that actually originated in a since recanted Facebook post in a group, right? So it began on Facebook. But it was something that was amplified and took life on X. It didn't spread on Facebook. It spread on Twitter. It spread on TikTok. And eventually it made its way to the debate stage. And I don't know, I almost, there's almost something that I admire by Facebook saying, listen, like, if that's the culture that we're going to drive, we don't want to drive it. I think that's still giving a bit of

credit to say that it's actually like a kind of more like goodhearted. No, it's not. Obviously, it's business oriented. They don't want to deal with this stuff anymore. Yeah, exactly. I think they saw where they became the great villain and all focus was on and advertisers were put at risk and now again, you're an advertiser choosing between X and Instagram. I think we know where everyone's going to end up. And in the meantime, focusing on, I mean, their ad platform, especially after iOS 14.5, was potentially completely at risk because they lost the ability to do a lot of the targeting they used to. And now it's back and it's the place to for every to put their money. So yeah, I think from right now at this moment, they're in a good place,

but there are risks still sitting there in the form of AI slop. Yes. One more thing on this. I do think that there's something sad about the fact that social media, like I was on the ground floor reporting on this stuff, like doing, I started making my way into journalism reporting about the digital side of elections. And there's something sad about the fact that, like, I was on the ground floor reporting on like there was this initial hope that social media would give voters new access to candidates and make candidates, you know, much more human and bring people together and just hasn't happened. It's sort of like it's almost a coda on everything that we've seen happen with the Internet

over the past 10 years. I was, you never believed it? Oh, no, no, I was a big, the 2009 Iranian revolution, it's like social media revolution. I was the biggest believer in all this. I think that, I mean, this is why I become so jaded about its effect on our politics because I genuinely believe this would transform democracy and just access and make the world a much better place. But I gave up on that a while ago.

We have about 10 minutes left and there's so much I want to cover from imaginary friend apps to AI slop and plenty more. So we'll do that right after this. 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 2 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 of writers break down the biggest business headlines in 15 minutes or less and explain why you should care about them. So search for The Hustle 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. talking about the latest news in the tech world. And I just, this segment started off with me finding a tweet from Nikita Beer,

which then was sort of the times picked up and wrote a whole story about it. And there's so many interesting facets here. So Nikita, who used to work at Facebook, he sold a company to them, so then sold a company to Discord. He goes, rest in peace, social media apps 2005 to 2023, as of iOS 18, friend-based contact sync apps are basically dead on arrival. The number of people consenting to the new permission has nosedived, so there is no way to get density in a meaningful way.

Basically, it was like you share, well, you allow your context to be shared with this app. People are saying no. Now he says, but as one door closes, another opens up. From the handfuls of dashboards, I've seen the retention rates on AI companion slash girlfriend apps are extraordinary. Instinctually, I thought the entire space was bullshit. But it's now clear that the ability to craft relevant, empathetic responses to any user input is a powerful lever to activating users and entrenching them into a relationship, albeit artificial. For better or worst, LMs might be the new contact sync. If we can't find our friends on apps anymore, people will find something else to talk to.

Now, we initially plan to speak about AI slop here, but I think just starting with this idea of, of people finding Girlfriend, AI Girlfriend or AI Companion Apps to me is so interesting. And obviously it's a rising format, Nikita has his finger on the pulse of what's hot here. So instead of, for instance, apps where you're talking with human companions, you might be talking to a social media companion.

And there's also this new thing that I tried out recently, an AI social network, it's called Social AI. It's by Michael Samen. who's a young developer, or less young now, but he was hired at Facebook at 18 years old to build apps for the youth. And basically what this social app does is it allows you to effectively tweet. You put your tweets in. And this whole group of people respond to them with like different types of interesting responses. So I personally thought that the app just didn't hit the mark for me. could see stuff like this starting to become more popular.

Ranjan, what do you think about these type of apps? Okay, so on the AI companion, well, yeah, separate from AI companion, but AI social network, I think it's almost an interesting, like, postmodern art experiment or something like that. Like, it's not like a real app, but it's more, because it's almost making us question, and not to get too philosophical here, but on regular social media, when you're on Twitter, and you post, I mean, there's some number of bots, even if they're real people, like posing as some kind of performative bot responding. There's some kind of real people responding, especially with generative AI, they're probably completely automated bots responding.

So it's some mix of all of the above in existing social media already, and it's moved in that way for the last five to 10 years. So is it very different to just have the whole thing where it's just AI responding to you to make you feel good. And as he had said as well, it solves one of the biggest problems is right now the fact that follower-based social media is going away on Instagram, on Twitter, on TikTok, on everywhere, that you can't just open account now, open a new account and get engagement. It's nearly impossible. So like the idea that you're just going to post and people are going to respond, so they've

lost the like it's the cold start problem. where any new user is not going to find any engagement. So it could start killing these platforms from the inside. So maybe this is the way you do it. Imagine you get on there, you make a funny joke, and 20 people instantly respond supportively of you. That's what social media. Isn't that what we all want anyways?

Yeah, well, I am using the app. And it doesn't, I mean, using, I tried the app. And it just doesn't feel the same. You're using it all day long. Come on. Well, I do speak with my, I have like my imaginary diet coach that I built on Claude that I speak with legitimately every day and have begun sharing more and more, which is nuts, which we can talk about more on another episode.

Or maybe I should write about because it's gotten crazy. You're not falling in love with the diet coach, though. No, no, definitely not. It's dispassionate, but it's very, it's just, it's super helpful. Yes, it is. I mean, we talk about some very intimate stuff. food and and then but but i will say that that max brings up max read who wrote about this brings up an interesting point which is that why does it feel so weird to have a group of bots

encouraging you but less weird if you're speaking with a chat sheet pt which is just a singular one i agree i think like in the digital realm how much of all of this has been quote unquote real I think is it's been moving in a weird direction for a while. I think it's interesting. Like, I mean, we connected on Twitter. That's right. Initially.

And even John, my co-writer at Margins, we connected on Twitter. Like, many, many real life connections. But I feel all of those were made in like 2017, 18. Like, that's the golden age of real life connections being made on platforms like this. Nowadays, are they being made or is it just more, you know, 20 bots responding to you just to get a bit of dopamine? Maybe that is good enough.

So I think, and I agree that is that very different than talking to a chat TPT? No, it's not. It's like, I mean, especially if you're trying to get some kind of validation for something, it's definitely not different. It's just serving a different purpose. If I want information, if I want diet coaching, no one has any problems with that. So if I want to make a joke and have a bunch of people laugh at it and support me, maybe that's weirdly not that different.

Right. So I think this is going to be a space to watch, especially the AI girlfriend space, which we haven't really spoken a lot about, but maybe I should get somebody on the show who's actually building this for people because it does seem like the developer energy is going to go in this direction versus the other. And so as we start talking about this, another thing has come up, which is that, wow, the internet is just filling with AI slop. And we've talked about it a little bit on the show, but there was another great story in New York Magazine about the rise of AI Slop. And to me, the thing that really stood out was how prevalent it was.

It's magazines getting so many submissions that are just, you know, basically AI generated. It's books on Amazon telling you whether you could forage for mushrooms and what are poisonous or not. That's AI generated. And of course, Facebook pages, right? So what's replaced some of the politics? It's AI slop on Facebook where people like prompt ChatGBT, GBT, give me an engaging photo. And then ChatGPT will give you prompts to create an image generators and you'll make them, you'll put them on Facebook. They'll go viral.

Facebook wants this stuff. It's not banning it. It has its own image generator. That's how you end up getting like shrimp Jesus, which was like one of the most viral images this year. And it just continues to spiral to a point where we're not going to know what's going to be, what's real and what's not. I have an interesting thing to tell you on this, Alex. That's here. I'm a little nervous.

No, no, no. This is, this is when I show this to people, it's like the most fascinating thing. So I, as I had said earlier, had had some issues with Facebook and its impact on our politics back in. So I deleted my account in 2017. I recreated an account maybe a year ago. A brand new account, so it's starting from scratch. I also realized I was losing touch with my relatives in India and more second-degree family and kind of loose connections. So basically, my profile right now, which Facebook knows about me is, I'm a around 40-year-old male who has relatives in India, and that's about it.

It serves me. My feed is psychotic. Like, it's literally, like, I just opened it. Yes. So top post Russian Discover Girls, Hello, Friends. I am 51 widow from Australia looking from a boyfriend. a kind of older-looking, completely AI-generated woman. Like, I am in it.

This talk of AI slop, that's my whole feed. And it's crazy because I go into the comments. And this is, again, that AI social network, it's happening. Like a lot of the comments, I weirdly spend a lot of times clicking on the comments, which probably then the algorithm thinks I'm really engaging with this content. And that's why I get served more. Yeah, you're just researching. You're not just enjoying it.

and something you like to do. There's like a whole world of where it's like old celebs and it'll be Selma Hayek, but it's a clearly AI generated photo of her. And so, but you see some of the accounts look completely fake. And then you click through and they have like two connections, two friends or whatever. Some of them are just, it's great like real life older guys who have like their grandkids in their profile picture and list where they work, we'll just write the most creepy, obscene things in the comments.

Are they hitting on the AI, the AI image? Yes. Oh, yes. It's like, would love a great time with you or like a beautiful lady. Do you think they know that they're AI? I don't. I don't. Or do they, actually, that's a good question.

I don't think they know it's AI. But would they care either way? Or is this actually the AI girlfriend promise coming? to life. I don't know. This is one of those things, but knowing how what I see in my feed and knowing how just, again, psychotic in my mind it is, but it's so fascinating. And I wonder just how much of Facebook blue is this right now is something that like I'm in the trenches on. And it's real and it's happening and it's terrifying. Have you heard of the dead internet theory? Oh, no. What is this? So this is from the story.

It's this idea that humans are just a tiny minority online, and the bulk of the internet is made by and for AI bots, creating bot content for bot followers who comment and argue with other bots. That's the dead internet theory. Maybe we're living in it. I think, hold on, I literally, I just scrolled down from the account of volleyball girls. There is just, it says sports beauty, and it is a AI generated image of a, volleyball player with a gold medal that does not, you can even see in the metal I'm looking,

it's kind of like a warped in, yeah, warped text in AI. I mean, I think it's happening. I think the, it's definitely happening to an extent because remember, the platforms are still incentivized that that's a win, that that's a win from an engagement standpoint. I guess advertisers at some point might backlash, but not if the same people commenting beautiful. a full girl are actually going and buying a my pillow or whatever else so my god that is like the bleakest description of online life but unfortunately i think it's true yeah and and i guess the social networks are good with it that's the crazy facebook blue i mean but the mark mark mark suckerberg and meta have shown what diversific portfolio diversification of product can be we're going to move

just the complete dead internet AI slop world to Facebook blue and be totally happy about it and make Instagram kind of cool and entertaining it still is like it's not driving culture but it's still the place that every cool kid and fashion brand and anyone wants to be and any celebrity so yeah they've shown and WhatsApp is just message people you know and it's still the dominant force of that. So this is the episode I feel where we have really talked up meta in a both a good and a bad way, but at least they're firing on all cylinders right now. It feels weird to be speaking positively of a company that's sort of replaced political

discussion with images of shrimp Jesus and guys hitting on fake AI woman. But I don't know, I guess that's the world we live in. He has his head in his hands right now. I'm struggling to say anything positive right now. This episode has depressed me. I think I have been watching this for the last like six to eight months. And again, now that I'm saying it out loud, I'm realizing that because I'm so fascinated by the comment threads on these posts,

it clearly is just amping up the algorithm on just what is the most twisted of this kind of content to serve me. And it is. So I got desensitized a while. It is crazy how much of the internet is like this at the very beginning stages of generative AI. At the very beginning. And as the models get better, we're just going to be swimming in the slop. So on that bright note, Ron John, have a great weekend. And have a great weekend, everyone.

Ron John, thanks for joining. I'm not sure if I'm going to thank you, but it was great to be. Let's nod here at the end, an acknowledgement that we were present. And coming back to center. Okay, everybody, well, this has definitely gone off the rails as we anticipated when we looked at the dock as we began. But we appreciate you joining us week after week on Wednesday, Liz Reed, the head of search at Google, is going to be here to talk about what AI search is going to do to the web and how Google will transform in this era. So we hope to see you then. And then Ronan and I will be back next Friday.

Thanks for listening. And we'll see you next time on Big Technology Podcast.