Big Technology Podcast - Hawk Tuah and the Zynternet, OpenAI Hacked, One Year Of Threads

Episode Date: July 5, 2024

Ranjan Roy from Margins is back for our weekly discussion of the latest tech news. We cover 1) Why the Hawk Tuah Meme broke through 2) Evolution of the internet from the 'nerd internet' to the 'Zynternet' or 'bro internet' 3) The 'For You' recommendation system had turned the internet into a mass appeal machine vs. niche 4) Zuck's fourth of july water sports moment is indicative of the change 5) Raw dogging flights 6) OpenAI hacked 6) NYTimes rips off Dwarkesh Patel's Podcast 7) Figma's Apple Weather App design copy 8) AI startups raising a ton of cash 9) Will there be a reckoning for VCs investing in AI 10) Threads turns one — what is it? --- 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]

The internet is evolving, and the seemingly inexplicable meme at the center actually explains a lot. Open AI was hacked. Will it be worse next time? Funding pours in to AI startups and threads turns one. All that more is coming up on a big technology podcast Friday edition right after this. Welcome to Big Technology Podcast Friday edition, a show where we break down the week's news in our typical cool-headed and nuanced fashion. We have a lot of news to break down on this July 4th weekend. I'm joined as always. by Ron John Roy. We're going to talk about the involving Internet. We're going to talk about this hack that Open AI experienced. And then, of course, looking at the year of threads and whether it's what it means next to Twitter. So first let me bring in Ron John. Ron John, welcome.

It's good to see you, Alex. I'm currently in Taipei, Taiwan still at an Airbnb that is a converted massage parlor of all things. When we walked in, my wife, our family, and our cousins, you can see it's four rooms all kind of going all down one hall. And it's very clear what it was. It's been totally refurbished and it's very nice. But it's, uh, it's been an interesting week for me. I think that's like a perfect lead into like what we're going to talk about. I think it is too. I'll be honest. For the life of me, I couldn't understand why there was this one meme, uh, that was going as viral as it was for as long as it was and just wouldn't die. And it was was something I really wanted to ignore because I didn't want to talk about it on the podcast.

But I think we have to now because it really has a signal of how the internet has shifted. So for those who... What is that meme, Alex? Well, I first want to know that this is a parental guidance moment. This is definitely going to be a wacky July 4th show. So I'm just letting guys all know and gals know at the start that this could go off the rails. So if you have kids in the car, maybe, you know, switch the channel and then you can sort of pick this up when you're, when you're, on your own time. But basically, there is a meme. It's called the Hawk Tua girl. And I'm just

going to read, because I don't really want to try to explain it on my own, because it still makes me a little bit uncomfortable. But I'm going to read from Max Reed, his story called Hawk Tua and the Zinternet. He goes, Hawk Tua, if you're not yet familiar, is a newly popular meme derived from a street interview video in which a woman in Nashville named Haley Welch responds to the question, what move in bed makes a man go crazy every time? And she says, I don't want to even read it. She says, you got to give them. I think everyone's seen it.

Everyone's seen it. Okay. And so I was like, for the life of me, I couldn't understand why this meme was so prominent and why it just kept spawning meme after meme and wouldn't go away. It seemed like a throwaway moment. an on the street interview, not a big cultural moment to me at all. And now some of the folks who are writing about the internet have kind of explained it and it makes a lot of sense in context.

So this is again from Max Reed. The broad success as a meme is less about its own intrinsic qualities than about the increasing size and cultural power of something I'm going to call the Zinternet. Over the last 10 years or so, a broad community of fratty, horn dog, provocative 20 and sometimes embarrassingly 30-somethings, mostly but by no means entirely male, have emerged to form a newly prominent online subculture. And maybe you could even call it the culture. This network is adjacent to the sports internet, the hustle internet, and the reactionary internet of traditional influencers, but in its own distinct community

with its own distinct cultural references. And here we go, college sports, gambling, light domestic beers, and then zin nicotine pouches, with influential personalities and media outlets, among them David Portnoy, Pat McAfee, Antonio Brown, and call her daddy,

in additions to dozens of minor podcasters and hey, fellow kids, content creators, who nearly all work for sports betting concerns. So basically the argument here is that this is not a meme in a vacuum, but really sort of like the, epitome of an internet culture that's moving from whatever it was beforehand to one that's far more friday, far more dirty, far more obsessed with gambling and bedding and very different,

I would say, from the early social internet that was really defined by nerds. And the nerds wanted news because they wanted influence and nerds wanted celebrities because they wanted to feel important. And that's why we ended up having effectively journalists and news outlets pour into Twitter and Facebook and they would celebrate any moment when, let's say, a Justin Bieber would tweet. I feel like in Twitter offices there must have been an alarm that went off when Justin Bieber tweeted. But we're definitely living in a different internet now where sports betting has been legalized. It's financing a lot of content. Online gambling is financing a lot of content.

And then, of course, Elon Musk owns Twitter. And that's what we get. We sort of shifted from this like nerd internet to the bro internet. And that's what this internet is. I'm curious what you think about the shift, whether this is something that's being, you know, like over-emphasized by the bloggers and whether Hock-Tua really sort of showed a shift here. This is my favorite piece of cultural internet writing, I think, of the year. And the term Zinternet will stick, I'm calling it.

I think even Max Reed himself recognized, as he said, that he has to write this piece, because once you come up with the term Zinternet, you can't just let that one go. So my only slight issue with the piece is he talks about a boreshly provocative 20 and sometimes embarrassingly 30-somethings. I will recognize being in my early 40s, some of the brochats still lean in this direction. I got a happy Zindependence Day text a few hours ago right before starting to plan for this podcast. So I think it might stretch even further. But the most fascinating part about this piece I thought was how it tied together those different trends. So first of all, Zinn pouches, which is one of those weird subcultures that as I get older, I had realized had become

massive and I just did not even understand what was going on or what was happening. And then even especially 20-somethings at my work, talking to them. Explain what a Zins are. I'd say tobacco pouch that delivers, actually, I'm not even 100% sure, but I think it's the, like a Zin is to, chewing tobacco what a vape is to a cigarette it's like a more clean nicotine delivery but via pouch in your lip like you would have chewed tobacco in the past right by Philip Morris

and so I would say so powerful that I was walking around Manhattan with a health tech CEO recently and he offered me a Zinn and was just like you got to try this stuff I'm like you know that's bad for you right but he's like these things they really hit they I mean they they really hit I guess that's why I think

but but so that's already been there then you have the rise of sports gambling I think that's the most interesting connection here because that's the financing of all of this is that if you are trying to launch some kind of clothing line you're trying to monetize somehow like an influencer without monetization is nothing and now suddenly there's been this massive influx of money going after the people who can reach these internet audience and being able to give them money. And Dave Portnoy kind of is the, like, original example of this. Because remember, he sold barstool sports to pen gaming right before gambling was legalized for, I think, $500 million, $5.50.

And then they essentially wrote it off and sold it back to him for nothing. Like it just did not work. But he was the original sign that you can take your audience of, take this audience. and monetize them massively from gambling. And then the third one, Elon Musk, Twitter, now the algorithm is certainly tweaked. The culture is moving in the direction as well that, like, these kind of things are becoming more mainstream. So I think realizing how all three of those factors come together to perfectly crystallize in

the form of haktua, I think it's a pretty clear, impressive thing that he wrote. The one thing I would say it's missing is that there's a, also an element of backlash to the original social internet, right? Whereas like the original social internet was, shall we say, very democratic, prude to a certain extent. It was, I don't know, it became to be less fun. I don't think it really catered to the common person as much as this internet does right now. I would disagree with safe and prude because if you're ever on Tumblr in 2011-12, it was not a safe prude place. But, But I actually agree with you that niche, everything was niche.

Everything like signaling. Everything was about finding some niche thing that no one else has seen. And within your community that's online of people, the like 100 other people that think like you being able to find them, I think that was the promise of that early Web 2.0. And then now, you're right, it's mass market meme. What is something that is appealing to just a massive swath of the popular? and that's the Zinternet. And can I,

can, maybe we can dig in a little bit more because it's starting to crystallize to me that this internet or whatever you want to call it, the Barstall internet, whatever it might be, has also risen alongside the ascent of the 4U page. Right? You talk about this mass market versus niche. A niche would be great.

Like if you're a small account on Twitter, but you had a niche following, you could still speak with them every day. But now you're not getting any play at all on the 4th. you page. You're totally being outshadowed by, you know, this sort of Zinternet content, which is going to be more engaging. Remember, like

Haktua sort of like threads the needle there where it's like, it's sexual, it is entertaining, and it's like, yes, of course, in like an era of the for you page, it's almost the for you and the recommendation algorithm has ushered this in just as much

as the rise of gambling and Elon Musk buying Twitter. I think we're going to have to race to see who writes this post first because I think that's perfect. I think like is okay so there's a for you page on Twitter but then also the rise of reels and faith in Instagram completely shifting from a follower based social network and a chronological timeline to also algorithmic short form video so even if you weren't on TikTok already now you're on reels and it's being algorithmically served to you and I can admit again my Instagram feed there's a lot as internet

content. There's a, I mean, and then you even see how the advertisers, it's like bird dogs, pants, if you're familiar, Vorey definitely goes after these internet, like how the whole thing comes together. And that's just from probably watching one video following Dave Portnoy. And then just from those signals, the Reels algorithm just completely inundates you with this kind of content. So I think, I think that's right, algorithmic based content feeds now. And who has embodied the transformation from what's called the nerd internet

to the Zinternet more than Mark Zuckerberg himself. Because today it was impossible. If you logged on to the internet, we're talking on Thursday July 4th. It was impossible if you logged on to the internet to miss the video

of Mark Zuckerberg wake surfing in a tuxedo holding an American flag. Hydrofoil. Hydrofoiling. No, no, this isn't, I don't think this is hydrofowing. I think this is wake surfing. Isn't that his big thing?

Hydrofoil? No, I think he's switched because he's in a wake and I don't think you hydrofoil in a wake. But anyway, either wake surfing or hydrofoiling, basically doing a board sport in a tux in the water with an American flag in one hand and a tallboy, it looks like a beer. I don't know, maybe it's like. A Coors banquet, a Coors banquet tallboy. I had to look into this. I had to do the research. And it's like, you couldn't, you really.

could not sort of explain the shift to this new internet better than like looking at Mark Zuckerberg posting pictures of himself in like fields in Oklahoma with sheep in 2016 to what he's

doing today which is surfing it looks like in Hawaii with the American flag of the tucks and the long hair and that is that is sort of where we are and he's the content king it's a broadest signal as possible that this is where the internet's moving.

How do you think those posts come to be? I've always been fascinated. And I think we've talked about Mark Zuckerberg's overall kind of PR transformation from nerdy internet guy to ultimate fighter, cool guy, handsome Zuckerberg, wearing chains and everything. But like a post like this, I don't know. I always want an oral history of like, you know, was it him who brought it up? Was there a meeting?

Were there a series of meetings set up? Is it actually official Facebook resources? Or is it just him and his friends hanging out and playing around? I don't know. I'm always so curious these kind of things, how they come to be before they are. Yeah, definitely seem like a professional videographer. Yeah, this one seems professional.

Also to note, I follow. So in Facebook, WhatsApp, and Instagram. have been trying to promote this idea of broadcast channels. And I've, like, looked into this in the past. Basically, it's the idea that you kind of follow someone and they push content to you. So it is, like, a one-to-many broadcast feed. So obviously, Mark Zuckerberg was the first person who launched this type of account within Instagram. And he pushed it to millions of people who follow him on this.

And the app definitely kind of, I don't want to say, tricked you, but led you into following his broadcast channel. So, again, they planned. land this video. It's good. It's pretty awesome. And then they just push it out to everyone on Instagram and happy July 4th. Yeah, but go back to this like the internet thing. Like is this, so I'm pretty convinced this is a signal from Mark of where this is all heading. What do you think? Yeah, no, no. I think you can see even in his circle doing something like this, they all thought would be cool and funny and fun. Like, again, wakeboarding and tall boy in hand.

And it kind of is. Like, I hate to. Yeah, it's resonating online today. It's rough. Yeah, it's resonating. Like, his whole UFC thing, I always find a bit cringe-worthy because, like. Also, Zinternet adjacent, sports betting and all that stuff. If you think about it, Joe Rogan is the, like, actually early Zinternet crossover from UFC to,

for our podcast talking to, like, you know, intellectuals, that's pretty, or that's pretty zinternet as well. Yeah, no, I think it's, it's definitely here. It's grown. It's more powerful. There's so many different factors, as we've talked about, kind of making this more of a thing. I guess, and you had mentioned it briefly, but I do think the backlash to social media and the algorithms favoring, I mean, quote-unquote woke content from 2020 and 2021, where everything moved so far in one direction that there's definitely a backlash element as well here. So I think everything coming in together makes for a lot of content. And what is social media labeled this summer, by the way?

Oh, so I had to put this in. Well, no, no. So this was, again, in terms of like, the first place I came across the Zuck meme was Zuck One America Today, of White Boy Summer. And for those who don't know, like White Boy Summer, I found out was a term coined by Chet Hanks, Tom Hanks's son, actually a couple years ago, that I've seen being thrown around. And again, you picture like polo shirt, again, actually, polo shirts and tallboys and Hamptons, barbecues or whatever else. But the funny thing to me was, apparently the term white boy summer has been co-opted by hate groups and far-right organizations. And Chet Hanks, Tom Hanks' son actually had to write an Instagram post kind of apologizing for this saying

white boy summer was created to be a fun, playful, and a celebration of fly white boys who love beautiful queens of every race. Anything else that it has been twisted into to support any kind of hate or bigotry against any group of people is deplorable. And I condemn it just the fact that like someone sat down and with a PR team and decided I have to try to reclaim white boy summer is not racist is uh always hilarious to me yeah but but it sort of i mean that is that is obvious it's sort of like almost a factor of the old internet but like that this stuff will end up in the hands of hate groups but the fact that it's been broadly embraced as something that sort of is going to define summer of 24 right like even that like finance uh

six five blue eyes is also part of this right it's like that is that is sort of what we're dealing with so oh one more thing before we go before we move to the next topic what in the world is uh rog dogging a flight well so having been on a number of flights in the last three weeks and this was from tom gara one of my favorite tweeters i'd worked with them at the financial times he had said also in my opinion the zinternet is reshaping norms beyond its own borders i.e. internet culture type people doing trend pieces on raw dogging flights. And I'd love this because the whole, so raw dogging flight, for those unfamiliar, is this trend that's kind of taken off on social media of someone,

people bragging about taking a 13-hour flight and just staring at the flight map the entire time. Basically, no books, no watching no movies, no anything. no, you don't pay for a shitty internet Wi-Fi or an airplane Wi-Fi and try to do work. You just sit there for 10 hours, 8 hours, 13 hours. And having been on multiple 13-hour flights and watched every bad movie that's in existence right now, I have not done this, but just the term rod-dogging a flight, the idea of kind of like masculine, like, reclamation of just, like, combined with some weird health, or a wellness type thing of like meditate.

And it's called a meditative practice is definitely a weird extension of the Zinternet, I think. Yeah, but, yeah, it sounds like the most miserable thing that you could possibly do to yourself in the world. You just don't, you don't understand it, Alex. That's, it takes you to a higher plan. I don't. I'll be the first one to admit it. It is the most ridiculous thing to ever do. Don't do it. There's movies that, I mean, it's torture.

Otherwise, I guess I shouldn't use the word torture. But it sucks. you know you've got to find a way to pass the time you're right though that and like now that there is this audience that's why and as tom had said in the tweet like this is why a journalist will write a trend piece on this because there is an audience receptive to this whereas in the past they would not have found that audience so they wouldn't have written the piece so i think like that's why this type of thing i'm i'm curious what the next one's going to be but i think there's going to be a lot more of this kind of

Zinternet content coming out in mainstream publications, because that one I saw in like, I think some pretty mainstream publications. That really just sums it up perfectly. So there's another story this week that really caught my attention. I couldn't believe that how little

attention it's getting, maybe because the details are a little bit odd. But this is from the New York Times. A hacker stole open AI secrets, raising fears that China could do. It's interesting that like they go to China right away in the headline, but I'll just read a little bit

from the story. Earlier last year, a hacker gained access to the internal messaging systems of OpenAI, the maker of ChatGAPT, and still details about the design of the company's AI technologies. The hacker lifted details from discussions in an online forum

where employees talked about OpenAI's latest tech, but did not get into the systems where the company houses and builds its artificial intelligence. Okay, so the hacker got some so I'm like company chatter didn't get the weights on GPT4 but it is it is a concerning issue here that open AI at least their discussion forums could get hacked this is continuing on open AI executives revealed the incidents to employees during an all-hands meeting in April

2023 and informed its board but it did not the executives did not share the news publicly because no information about customers or partners had been stolen They didn't even tell the FBI or anyone else in law enforcement. And this is like the key line here. For some open AI employees, the news raised fears that foreign adversaries such as China could steal AI technology that while now mostly a work and research tool could eventually endanger U.S. national security. And one of the employees that actually raised concerns is a recurring character here on a big technology podcast, Leopold Oshenbrenner, who we talked about a few weeks ago.

and who William Saunders from OpenAI talked about raising security concerns. And then eventually Open AI effectively, I mean, this is the story that Saunders told, reprimanded Aschenbrenner, and then later fired him, although Open AI says he wasn't fired for this. But, you know, sort of like the worries from Open AI employees has been effectively that they are more concerned with, let's say, I mean, their words like the safety team folks say, like, well, you know, Samunders said that they're more concerned with shipping shining new products versus like advancing the safest possible research. And this is the thing that set off alarm bells. I mean, it's not quite clear what was hacked there. Like was it their slack that was hacked? It wasn't like their internal mainframes. But the fact that they could get hacked and they were hacked sort of brings into scope like how many entities must be trying to hack them and how concerning it can be that they are susceptible to security breaches.

What do you think about this? I think the key point there is not that you had said, like, that they could be hacked and that they were hacked. It's that they were hacked and did not disclose this outwardly to the FBI, did not, like, the topic of safety, I think this is a really interesting story because obviously normally the topic of safety leans towards is AI going to kill us all, but I think a more kind of like pedestrian topic of safety is just hacking and will foreign adversaries be. able to get the data that's been fed into these systems because it's one thing that right now the whole conversation is we promise you our models will not be trained on your data if you have a certain enterprise plan okay do we trust them on that I don't know it's still a little iffy sometimes but we will protect your data from all manner of hacking it's another thing that when you are a company that that's actually not that big and mature and developed security I'm guessing

like cybersecurity is probably one of the last concerns on your mind as you're scaling and growing. So I think this is, I think this is not the first, the last we're going to hear about this kind of thing, whether it's open AI or other, you know, like really fast, buzzy, growing AI startups. I think the cybersecurity element is something that no one has been talking about. Well, actually, one of the things that people have talked about the cybersecurity is that cybersecurity companies are going to really become way more valuable in the age of Gen AI.

I think probably for two reasons. One is companies are going to want to protect their models like OpenAI. And also it's going to make it very much easier to, like, for instance, fish people when you have a Gen AI tool that's able to like comb the internet about them or read documents and then write something extremely personal. But it is interesting. Like I do think that like right now we're not at a place where like if they, let's say, China got the GPT4 weights, it's not like. like a national security problem. Like if they're able to get that model, it's not like the U.S. has to like, you know,

go to war with China or anything like that. But you do start to see, and also we should note that there was no customer data or partner data that was exposed. But you do start to see the concerns of like some of the Open AI employees that have spoken out where it's basically just like if you're not dead set on security

and you develop something like this, like let's say you develop artificial general intelligence and you're susceptible to a hack of any kind, like you're going to be in possession of the number one target for hackers period if you develop aGI security has to be like basically top priority along with the actual research of how to push this technology forward yeah i think on the safety topic it is it's it's the more powerful and dangerous whatever your building is obviously it's going to become more interesting and almost like a james bond movie a target for nefarious hackers all over the

world. So I think it is important. But yet, the more, like, how are these companies building out their kind of defense functions in this way? And are they? We don't know. I think that's something that people need to look into. Because if you think about it, like an open AI kind of is the poster child that the core product and products have rolled out incredibly, have launched amazing products, done them very fast. But then certainly on the corporate governance side, like are they if they're ever going to IPO or their financials like public company ready, I would bet not. But like, I mean, have they built out their accounting function carefully? Their accounting is basically just like open AI sells and pays for expenses with Microsoft money.

Yeah. I mean, actually they do have complex accounting. But, but I think like all those kind of secondary support functions that are just part of usually like a midsize company, certainly a company with billions of dollars in revenue and tens of billions in valuation should have those kind of functions that are already mature. But I'm guessing that that's not been a focus of them or probably most other of these AI startups. And I think at some point, we will start to see some fall out on this. Yeah. And by the way, just to close the loop on a discussion we've been having on the podcast for the past couple months, we've been talking about like if the open AI employees

saw something, they should come out and say something. And we had William Saunders on the show on Wednesday is ex-open AI super alignment team member who basically said like there's nothing immediately concerning or immediately dangerous to like society or humanity or node like technological development in particular that's you know scaring the daylights out of these super alignment researchers so we got that out of the way but this type of stuff is the thing that do that does worry them along with like not having enough I think not having enough resources to run the research that they would need to make sure that the development of these models are safe.

You know, it's good to get to the bottom of what they're actually asking for, and this is actually something that always we think a little bit more about it. You know, it makes sense. Like, the thing that they're asking for does make sense. The New York City subway system slogan is going to save us from AGI and the robot takeover of if you see something, say something. Exactly. That's all we're asking.

Anyone out there. I thought about that, yeah. I thought about that, actually, as we're talking. it's a good it's a freaking good slogan yeah that's one thing you can say about new yorkman they brand very well uh one more thing i want to talk about about this this show this article that annoyed me it's not really that bad but it's sort of typical so the times obviously like listened to leopold ashenbrenner's interview with dworkish patel friend of the show who's been on before and they cite it as archenbrenner alluded to the breach on a recent

podcast, but details of the incident have not been previously reported. Okay, so clearly they followed the scent that Ashton Brenner left on the Dwarkesh show. And they also grab a picture of Ashton Brenner on the Dwarkesh podcast, but they cannot bring themselves to cite the Dwar Keshe podcast, even though the information from the show sent them, seemingly sent them down this trail. And that pisses me off. I mean, if you have somebody whose information you're, relying on you're taking a screenshot from their youtube and putting it in your story i can't understand

the arrogance of not including the name of the podcaster in the story it drives me nuts what are they not this is better reporting than than you did until dworkish came along and unearth this not the first not the first independent media uh professional to have that complaint i think yeah I mean, they, I mean, it's similar that they would do it to us when we were in at BuzzFeed. Oh, BuzzFeed News. I mean, I think that was a poster child for that. Right. Well, just like the big, it just doesn't, it doesn't make any sense to me.

Like, I've had scoops before also. Actually, the Times has been good at crediting my scoops, but others have not. And I just feel like there's this sort of sense of superiority that drives me freaking nuts when it comes to this. All right. So we should also talk about, in terms of, like, giving credit very quickly, did you see this thing that happened with Figma over the past week where Figma effectively released this tool where you can design stuff based off of a prompt and it ended up basically emulating the Apple Weather app when you asked it to build a simple weather app. Have you seen this story? Yes. So Figma, again, like a prototyping user interface, user experience design tool. Now, like many others, has the generative AI component.

And if you basically asked it to create me a weather app, it created you a weather app prototype that looks exactly like Apple's weather app. And it's funny now, Apple is kind of the, instead of a small artist or independent artist, it's actually Apple that's getting copied into oblivion by generative AI. But the interesting part of this was at first, the Figma C. CEO came out and basically said that, you know, we did not train actively on any other like IP like Apple's weather app. And that part was very interesting to me because, but also to note, they also now put the feature on pause as they're kind of doing more of a QA on what the underlying data is and what the different types of outputs are and if they have other types of problems like this. But what was interesting to me is that question, if he's telling the truth and saying we did not actually train on Apple's weather app, it's interesting because now Apple's weather app is so prolific and on every phone and everyone has it that now the AI models and just the world thinks that this is what a weather app looks like.

and then if you ask it to generate a weather app that's what it'll create so so it's interesting to think about in terms of like creative generative AI processes the more common a format is or like the more popular your tool is at a certain point that just becomes the reality of what an AI would think it is or it looks like so i think it raised a bunch of very interesting issues around this i mean i like dylan and i like figma but i don't buy it i mean it was such a carbon copy of the Apple Weather app, that there's no way that this thing did not train on the Apple weather app. It spit out the exact same thing. And now you're starting to get Figma developers being like, oh God, if we're developing in

Figma, are we next? I have a design friend. He sent me a couple of these tweets. Somebody give it to me straight. Is Figma going to use our designs for AI, or are we misinterpreting and overreacting to this? And then someone replies, if you have to ask, dot, dot, dot. I think that this also plays on a theme that we've been talking about over the past.

couple months, which is that these fears of IP theft are going to go far beyond the news. Like, information on the internet is the sort of bare bones of what we do online. Then we extend more and more, and we're extending here as well. And now we've seen the designers get involved. We've seen Hollywood get involved with Scarlett Johansson. And it's not just the New York Times that's getting angry and ready to sue, right? And this is sort of a large growing problem that just extends beyond news. And we saw the Suno thing last week.

Yeah, I was just going to say, we saw the music last week. And to me, this one was even more interesting because as someone who's not a designer, but very familiar and work with plenty of UX and UI designers, like what is actual theft? Like, is having a, obviously having the exact replica of an icon seems like it probably would be bad, but would having a specific layout be a problem? Would matching three different colors in a color scheme be a real problem? Like, I think it enters a world where it's even more vague and difficult to try to understand what's actually copying and what's actually just being inspired by whatever's out there. And I think, again, the Apple Weather app's the perfect example of

if that's what everyone looks at and thinks of a weather app is looking like, then obviously AI is going to think the same thing. Right. And look, despite some of the issues, we've seen just this enormous amount of cash, enormous, pour into the generative AI world. And there was another New York Times story, by the way, I mean, outside of not crediting to our cash,

they've done some great work on AI this week. There was another New York Times story by Aaron Griffith, friend of the show. Investors poured $27 billion into AI startups defying a downturn. And funding for AI firms has made, made up nearly half of the $56 billion in U.S. startup financing from April to June, according to pitch book. There's this whole meme about whether AI is investable or not.

And quite honestly, like, it seems like VCs are pouring money into AI, a lot of money into AI, and they're doing it whether or not they're going to get returns. Do you think that this is like, it feels like we asked the question about bubbles. I don't think either of us think that AI is a bubble, but do you think that this is just like another place where people could really lose their shirts as the hysteria and the buzz around AI continues to build? Yeah, I think when we've said this,

it's not a bubble, but a lot of people are going to lose a lot of money and a lot of companies are going to fail. I think is how I'm looking at this. The amount of money and the speed with which it's going in, I think, is huge. And actually, take Suno, for example, I came across, it was one of the A16Z investors. It was a thread from like back in March, but around talking about the AI tourist problem, that what's happening with a lot of these consumer-facing apps is there's this whole infrastructure, again, and probably driven by the For You page on Twitter,

where 10 crazy new AI things I tried this week, and getting to like a million users very quickly and people signing up for your product and trying it and using your compute has actually become a lot easier because there's a ton of people, myself included, trying everything. And then people can go out and raise money on that,

and they call them AI tourists, and I'd put myself in that camp trying everything, but there's no retention. But I think a lot of people are able to raise money right now that you go mildly viral, go to a bunch of investors who are just salivating over any AI opportunity they can get and raise a bunch of money. How bad do you think the fallout is going to be from this

when they have to return that cash? I don't think it's going to be like systemic failure of the Silicon. Valley Bank going down is like the bank existing. I think it'll be just a lot of people losing a good chunk of money and a few people making a lot. One thing

I've seen more that this is a different moment because of Lena Khan because in the past and that's why you've seen like with inflection I think there's another one last week where instead of buying out the company when it's not really going anywhere

Microsoft hiring the inflection team and like leadership or, like, I think we're going to see more of that, which is not a good outcome for investors either. Because basically, tech giants cannot just buy up every single AI startup right now. So it's going to change the incentives a little bit more, I think. Yeah, and what about this company runway? Oh, I also, in talks to raise $450 million. I have used them for a while for image generation.

So they just released Gen 3 Alpha, which is a text-to-video generator. And honestly, it's pretty amazing. It's also not amazing in certain ways. But like, so basically I was trying a couple of things. And this is like Open AISora, which we've seen demos of, and they've apparently released it to a small group of people. Anyone can go use this, subscribe by some credits, and you can generate video. I tried a couple of different ones.

The first one was, like, a man, I've been driving a moped around Taiwan a lot and really enjoying it. So it was like, man wearing a purple helmet, driving a moped in the streets of Taipei at night with like blurt out lights. And it came out really well, except for the motion of the wheels. It looks like it's just drifting the entire time as opposed to actually driving. It's weird. But you can see, okay, this is moving very quickly. But then you start to realize, like, if you do generic, I did. like a tiger walking in the Serengeti desert at dusk, and it's beautiful.

It's like genuinely good. So if this is V1, I can only imagine, like it felt a lot more real the Hollywood's in trouble after seeing this because, I mean, for me, just quick prompts, seeing the outputs and seeing what's going to happen, I think especially for any kind of B-roll or side shots or something like that as opposed to people acting, I think a lot of that is going in this direction. That's crazy. I mean, it's, it's amazing how two years ago, like, the idea of generating an image with text was crazy. And now we are, we are where we are. This stuff is going to end up being put into action in professional production. So, why don't we end this week talking a little

bit about Threads, the Twitter competitor from Facebook. So they hit 175 million users. They just hit one year. And this is from the verge, a year and a half ago, Threads was but a twinkle in Mark Zuckerberg's eye. Now the rival to Elon Musk's X has reached more than 175 million monthly active users. And basically, it's growing. We hit 100 million users very quickly, I think within a month or so. So then it lost. and now it's find a way to, found a way to rebuild. The interesting reaction I saw on threads, really, on Twitter, which was that a lot of people, like said,

175 million doesn't feel like that. And in fact, the app has given away a lot of its relevance because it's not going to suggest political posts from people that you don't follow. It's also, it seems to, like, not be so interested in news, and it's slower. And I guess, like, in some, ways this is good because Facebook is learning some of the lessons of its past when it comes

to spreading political information. But it's also just like it seems like it's never really going to have a chance to have the urgency it needs to actually challenge X without that. And so I'm kind of of two minds of it where I'm just like good on you meta for making the right choice, I think, from a safety perspective, but also just like threads seems to have like an expiration date to me if it doesn't sort of go to the dark side. And I'm not encouraging it to do that, but it's just what it is. Here's what you think. Yeah, I think threads to me, especially because the fact that they're saying we don't want to get into political content, in the next few months this year, that's going to be a pretty critical thing. So to try to

avoid that at all costs, I think is a big miss. But I think the main thing for me is still, I just don't know what it's for. Like when I go on it, there's some people I see who very like actively said I'm leaving Twitter. There's a lot of like posts two days old, three days old that are kind of viral. There's a lot of auto posted things that you could see the person that posted on Twitter and is cross posting two threads. So it's just, it's still feels kind of like people are there for the most part because it's there's some people there you're supposed to be on it if you work in social media but otherwise i don't know i i still maybe there are thriving threads communities that i'm completely missing but it still feels to me that it's more just

brute force facebook can go from 100 to 175 million users which is actually not that impressive when you have two billion people using your products every day and can just just hit them with any push notification you want. And they do that with Instagram anyways. So I think I'm still not quite, I'm not a Threads boy yet or anymore. I think you had your moments there, right? I had some moments, but I'm also just like,

I was just thinking, is there any time I learned something surprising or new from threads, and I can't really recall, whereas I feel like I learned something surprising and new from Twitter every day, despite how awful it can be. And by the way, I just did something that I wasn't expecting to do, but it's kind of interesting.

I'm logged out on threads right now, because you were asking, like, what does threads want to be? And I'm like, I'm looking at this logged out experience. I was actually trying to get the logged in. But let me tell you what these first few posts are. So there's a post, I think, by a dog. And it says by Honey Bear Henry. And the post is, Mommy says, I look like the 4th of July. And it's got like a 4th of July handkerchief around.

neck. The next one is a book recommendation. The other one is happy fourth with a picture of fireworks. The other, the next is a picture of a rock band. I can't really tell what the next one is. Eventually I get to a picture of a reveal of the Puma jersey for some soccer team and then a chicken that's going. Cluckia, America. This is, this is, this is, it's incoherent in terms of a vision for what this thing should be if that's the logged out experience it's just like it's not tumbler it's not Twitter it's not Instagram and I don't know I mean maybe that's fine like maybe it is just this kind of extra place that they can put you know sort of put some resources to to it but it doesn't really make sense to me why Facebook is doing this without without news and I think there's ways to do news responsibly I don't think you want to do politics necessarily, but there's stuff with that the news gives a product a sense of urgency. And I just don't see it there. Yeah, I'm actually looking right now that Mark Zuckerberg has been, he'll go like over a month without posting on threads.

Like May 16th and June 18th, then June 25th, so a week. But like before May 16th was April 19th. So it's not like, like this kind of service only works. where you're on it a lot. It's not like a scheduled post or not like you're jumping in every now and then where you're thinking out loud. That's the beauty of Twitter.

You think out loud on that service. And yeah, I think the real-time part of it, it's never going to get. It's just, it's been a couple of years now. And I don't see how, if it's not encouraging that kind of, that way of creating content or posting, it's never going to get there.

Yeah. and I heard like the conspiracy theory that I've heard is that this is like meta's way to train its models on real time information but that doesn't really compute if there's like not like any like pressing real time information that's showing up there. So I mean I continue to use it.

I log in like a couple times a week. I think it's interesting. It's fun to share things there. But it doesn't feel as alive as Twitter does. Maybe Zinternet threads is actually a thriving, vibrant community and we should all start to follow whoever the the key thought leaders in that space are yeah i mean barstool is doing well there 2.1 million followers so says something i don't know what it says but all right we managed to cover hawk too i think in a fairly like

reasonable nuanced and cool-headed ones and cool-headed way and it's actually one of the more fascinating things I think we've dug into here, like, talking about the evolution of the internet. So I'm glad we did it. I think Zinternet is not going anywhere. And I think, I think July 4th is probably the most zen of holidays. There is just meat posting, water's like, uh, selfie in the tank top. Cool parties. Yeah.

Tank top, beer like shotguning a beer, just grilling massive steaks. I think this is Happy's Independence Day, Alex. Happy's Independence Day. And to everybody out there, if you're a U.S. listener are celebrating U.S. Independence Day, happy Independence Day. And if you're among our many listeners around the world, we're wishing you a happy Friday in a nice weekend. And next week, Matt Wood from AWS.

He's the head of AI there. He's going on the show. We're going one-on-one, talking about state of AI at Amazon. And then Ron John and I are we back next week. together? Or are you on the road? Not next week. Not next week? The week after. I'm rod-dogging a flight next week. Oh, please turn the movies on. Please turn those movies on. So we'll have a guest host that will come in.

All right, everybody. Thanks so much for listening. Thank you, Rajan. And we'll see you next time on Big Technology Podcast.