Charlie Kirk与网络愤怒 + Trump的芯片政策大转向 + 本周AI动态

Are you watching Alien: Earth?

No.

Okay, first of all, it’s so good. It’s my favorite thing I’ve seen this year. Second of all, one of the subplots is that there is a character who is a synthetic humanoid person, and she has mastered the aliens’ language.

Mm.

So she can communicate using the language, presumably with some sort of machine learning. And it’s paying huge dividends for her, because that alien is eating a lot of people. He’s leaving her alone. That’s why communication is so important.

I do think that interspecies podcasting is going to be huge. I recently got a pitch. There was a PR person I know who was telling me, “We have this AI company we’re working with, and they’re learning how to translate whale song using AI. Would you like to have the founder on?”

I said, “No, but I’d love to have the whale on.” If the whale would like to come on Hard Fork, we would love to interview the whale. I have so many questions for a whale.

What would you ask?

Feelings on SeaWorld?

Yep.

Free Willy?

Yep. Yep.

Let’s talk about what it’s like at the bottom of the sea. Tell us about the most interesting krill and plankton you’ve come across recently.

I feel like you don’t actually have a lot of questions for this whale. I like the idea, though. In the old days, it was Dog with a Blog. Remember Dog with a Blog?

Yes.

And now it’s going to be Cod with a Pod.

I’m Kevin Roose, a tech columnist at The New York Times.

I’m Casey Newton from Platformer.

And this is Hard Fork. This week, our thoughts on the assassination of Charlie Kirk and the terrifying new reality of extremely online violence. Then, The Times’ David Yaffe-Bellany joins us to unpack a blockbuster investigation into a $2 billion investment in Trump’s crypto company and a controversial deal to send high-end AI chips to the Middle East. And finally, it’s time for This Week in AI.

1. Charlie Kirk Masters Rage Bait

Well, Casey, we’re going to start the show on a more somber note than usual. Obviously, many of you, like us, have been reading about the assassination of Charlie Kirk, which happened while we were taping last week’s episode, so we couldn’t really get into it then. We also just didn’t know a lot, but in the week or so since then, we have learned a lot more, and we’ve had a lot more time to think and reflect and try to figure out what, if anything, we could add to the conversation.

So we’re going to talk a little bit about what we’ve been talking about among ourselves, what we’re seeing, what we’re thinking about, and some of the connections between these internet platforms that we talk about on the show and these violent events out in the physical world—and what, if anything, can be done about that?

Yeah.

How are you doing? How are you feeling? What has your reaction been to all of these events over the past week?

Well, I think, like a lot of people, I was horrified by the event itself and have continued to be horrified as further events have unfolded. I was not a close student of Charlie Kirk’s work, but I know that he was a master of manipulating the platforms that I cover.

Over the past week, I have been thinking a lot about his relationship to the platforms, what has happened on platforms since his murder, and what it tells us about how media works in this moment and how politics and power work in this moment.

Say more about that, because I share your sense that Charlie Kirk is wrapped up in the internet, and these were the platforms where he made his reputation. Turning Point USA, the organization he started, was a major force in early-2010s online activism and did a lot of viral content, for lack of a better word. That’s how he became famous, how he became who he was.

It wasn’t by running for office or going on CNN. It was by making videos and posting them online.

Yeah, and specifically, I think he excelled at making what some of the platform nerds that I write about would call borderline content—basically saying things that come right up to the line of breaking a platform’s policy without quite going over.

For example, he speculated that vaccines may have killed more than 1 million people without backing that up. He said it was a mistake to pass the Civil Rights Act. He would criticize Jewish donors for funding, quote, “Radical, open-border, neoliberal, quasi-Marxist policies,” and then reject claims that he was being antisemitic. Maybe he wasn’t, but over and over again, you see him flirting with this line.

The reason that’s important is because platforms love it when you flirt with the line. It turns out that the most compelling thing you can do on social media is almost break a policy. That was one of the key methods Charlie Kirk used to gain notoriety, prominence, and eventually power and money. He did it by mastering this particular technique.

The thing about this technique is that, as effective as it was for Charlie Kirk, I think it’s really corrosive to our politics. The more it succeeds, the more, when you open up TikTok or Instagram Reels, you’re seeing something designed to upset you. It’s designed to make you uncomfortable, challenge you, and that kind of thing, which Charlie Kirk was so good at, is now the way a wide swath of our politics works.

Yeah, I think that’s a really good point. I remember this graph that Mark Zuckerberg once posted on his Facebook account. This was a much earlier version of Mark Zuckerberg, who had not yet gone full MAGA. He was sharing some of his observations about the kind of content that worked on Facebook, and maybe we could pull that graph up and put it in the show.

It basically showed that the closer you get to the line—no matter where the line is drawn, which was the interesting thing—the more engagement you get.

Yeah.

No matter where a platform decides to draw the line and say, “This content is unacceptable,” the closer you get to that line, the more engagement you get. He was sharing this not as a celebration of the phenomenon. He was just saying, “Look, this is what we observe happening on our platforms.” It didn’t seem like he knew what to do about it then, and it doesn’t seem like anyone knows what to do about it now.

What platforms said—and both Facebook at the time and YouTube said—was that they were going to downrank this kind of thing. They said, “We see you edgelords out there. We’re not going to let you get away with this, so we’re going to try to restrict the spread of this borderline content.”

But as we look at the media landscape today, there’s a lot of stuff that’s just right on the border. It’s stuff that you can’t believe people are getting away with. People are always finding these new borders, and they’re trying to exploit them.

They’re not always political. Do you remember the milk crate challenge on TikTok?

Yes.

People were stacking up milk crates and then trying to walk across them, even though it was an incredibly rickety and dangerous structure. At the time, TikTok didn’t have a policy for that, but then TikTok watched a bunch of kids falling off milk crates, and they had to say, “We’re not going to promote this sort of thing anymore.”

This is just the nature of platforms that are looking for the most engaging thing for you to look at. They’re always going to be drawing your attention to something that is just on the brink of maybe not being okay.

In the wake of the Kirk assassination, I’ve been struck by how the entire cultural moment feels like so many people are making borderline content. It’s people making edgy jokes online. Some of those people are being fired, some are being celebrated, and some are getting angry letters from politicians.

In a way, this whole thing would have been a perfect story for Charlie Kirk to comment on and say some really edgy things about. You really cannot separate him from these platform dynamics that he mastered, which are now the primary mechanism by which his death is being discussed.

Unlike you, I have watched a lot of Charlie Kirk videos. A couple of years ago, I made a podcast called Rabbit Hole about the world of online radicalization and extremism. As part of that, I spent a long time watching the videos and listening to the podcasts of the people who were big parts of that story, and Charlie Kirk was one of them.

I should say, I don’t think he was the edgiest or the most radical by any means. In fact, if I had to put him into a bucket of online political influencers, it would be the debate-me-bro genre. That was what he was known for: He would show up at a college campus and have debates, and then clips of those debates would be posted on YouTube or other platforms.

And sometimes the video title would be “Charlie Kirk Destroys Liberal.” There was always this sort of very violent verb in there: destroys, humiliates, owns. This was sort of the blood sports of that era.

I’ll never forget this conversation I had with a left-wing YouTuber around this time. He was saying, “Yeah, we have decided to fight fire with fire. We cannot get any reach, any views, any distribution for our content on the left if we do not mimic the aesthetics of the right.” So they would do their own videos that were like “Liberal Destroys Charlie Kirk” or “College Student Owns Ben Shapiro.”

It was almost like they were just adopting the same kind of rage-optimization framework, but switching the sign on it. And that really struck me because I feel like the whole culture you’re talking about now—our whole society—has kind of absorbed this optimization strategy. We as a culture are optimizing for rage now. You see it on the social platforms. You see it from politicians calling for revenge for the assassination of Charlie Kirk.

You even see it in these individual cases of people getting extremely mad at the person who made a joke about Charlie Kirk that was edgy and tasteless, and going to report them to their employer and get them fired. It’s all this sort of spectacle of rage, this culture of destroying and owning and humiliating. And I guess, to me, what I’ve been feeling over the past week is that there just don’t seem to be a lot of checks on that. There don’t seem to be a lot of stabilizing forces right now, and the temperature of the internet and our society feels really high right now.

2. X Makes Civil War Feel Real

Yes, and I think it is particularly true on X. I think it might be worth drawing a distinction between how X handled this situation and how the Twitter of old may have handled a similar kind of murder.

In the old days of Twitter, when there was a very violent video spreading, you would have teams that would either try to take it down or ensure that it was hidden behind a screen. They would probably downrank it and try to ensure that it was not being recommended to people. Twitter was never perfect at that sort of thing, but at the Twitter of old, there was an effort to try to prevent people who didn’t want to see this content from seeing it.

Of course, Twitter also banned some of the most aggressive and violent right-wing accounts. People who had posted really racist and inflammatory stuff all got banned. Then Elon Musk buys it, turns it into X, and what does he do? First of all, he brings back all those right-wing accounts, lets them post almost whatever they want, and then, crucially, he decides to use the platform to push his own very inflammatory views.

In the immediate aftermath of this, before we know anything about the shooter, he tweets, “The left is the party of murder.” So today, if you look at X, it is easy to get the sense that we are on the brink of civil war and the right wing is gearing up for battle. In reality, I think the vast majority of Americans are not on the brink of civil war.

But if you were one of these elites who is still Twitter-brained and still checking the site 20 or 100 times a day, it is going to change your perception, polarize you, and make you think that things are worse than they are. So this is just a set of dynamics that I think is enormously consequential because the elites of this country are still on X, and what they are seeing is a sort of very bloodthirsty vision of America. That scares me quite a lot.

I don’t know, man. Don’t you feel like this is happening on every platform? Don’t you feel like the temperature is rising on Bluesky, TikTok, and Instagram? I feel like this is broader than X.

To me, obviously, X is a particularly potent example of a platform that has shifted its policies and its approach to borderline content. But I feel this everywhere I go online. I do not feel like there is a place on the internet right now that is being optimized for civil and respectful discussion and debate. I feel like it is all rage bait.

I basically agree with you, but I do think it matters that the leader of one of the platforms is using his platform, where he has the largest audience of anyone on the platform, to promote these views as aggressively as he has been doing.

But to your point, yes, the broader dynamics do hold. If you go on Bluesky, you will see a very angry version of the left. Now, for what it’s worth, what I have seen there has seemed far less violent than what I have seen on X. I have actually seen much more violent leftist content on X than I have on Bluesky.

But I think your point is a good one: This broader dynamic of rage bait everywhere, polarizing us all the time, is not just an X phenomenon. It does just seem to be how the app-based media ecosystem works.

Yeah, and to be clear, we don’t fully understand all of the motives at play behind this particular violent act. I don’t believe that anything is ever just the internet’s fault, but it does feel like everyone has absorbed the logic of the platforms. And there used to be people inside the platform companies who thought about this stuff.

Mm-hmm.

During the 2020 election cycle, there was an effort at Facebook to take down the temperature of the News Feed. There was this whole set of what they called break-glass measures. Basically, if it looks like the country is spiraling into some kind of civil war, there are these levers we can pull and these knobs we can turn inside Facebook to bring down the level of discord and hostility on the site.

Not all that stuff was implemented. They very quickly figured out that it was bad for engagement to do that. But it is an option, and I wish there were more people at the platforms who cared about this and thought about maybe sacrificing some of their engagement.

As we know, this stuff is very good for engagement, and it turns out that last week X had more first-time downloads in the United States than on any single day in its history, including before Elon Musk owned it. That’s according to Nikita Bier, who works in product over at X. So as violent and awful and rage-filled as the platform is right now, that does appear to be working for them, at least by this one very shallow metric.

3. Bridging Algorithms Seek Consensus

Yeah. This makes me want to talk about one potential solution that I think has not been fully explored. It would not be a complete solution. It would just be something I’d like to see more platforms try, and it’s called bridging-based algorithms.

We talked about them on the show before. This is an idea from a guy named Aviv Ovadya, who’s come on Hard Fork. It’s probably been a couple of years since he was here, but he came up with the idea that is now at the heart of the Community Notes that you now see on X and on Meta’s platforms.

You know how, if you see something that’s wrong on X, you can add a note to it and say, “Well, no, actually, this is the truth,” and then link to whatever the truth is? The way that bridging-based algorithms work is that they show those notes to people across the political spectrum. They have various ways of figuring out where you’re aligned politically, and they only show the note if people who are more on the left and more on the right agree.

Mm.

They see a bridge between the two of you, and they think, “Well, if Republicans and Democrats both think this is true, this is likelier to be true, and so we’re going to show that note.” That is a way that, using an algorithm, you can try to create some consensus. You can try to build a bridge.

I’m not aware of any platform that has tried to do this in the core feed. It may be that this is less engaging and a worse business than the ones that we have today, but we do know that there is a way to build technological systems that bring people together because they are already in use today on some of the worst platforms on Earth. So I do think that there is an opportunity here for somebody who cares to try to bring this idea to more places.

4. The First Amendment Under Pressure

Yeah. Okay, so if that’s the platform piece of it, what are you looking at now, looking forward? Obviously, we’ve seen not just the culture-war side of this, but we’re now seeing a government crackdown on speech.

We talked on this show back during the early days of this Trump administration about people like Brendan Carr, the chairman of the FCC, who had explicitly stated that one of his goals was to take the fight over free speech to the broadcast networks, to the platforms.

We’re now seeing things like Jimmy Kimmel being put on indefinite leave from his ABC show after Brendan Carr went on a podcast and condemned his comments about how the right was characterizing Charlie Kirk’s alleged shooter. So what are you thinking about the free speech of it all?

This is just terrifying, full stop. This is what the First Amendment is supposed to protect us from: political speech, and in particular, offensive, tasteless political speech. This is why we have a First Amendment.

And over the past decade, I’ve watched Republican after Republican complain about policies on platforms that they felt did not enable free speech to the degree that they thought they deserved it.

Charlie Kirk himself was a big free-speech warrior, and he was part of this culture war arguing that conservatives were being disadvantaged on these platforms. When the Trump administration retook power earlier this year, at first it was essentially, “Let’s make things fair and balanced, so we’ll approve your media merger, but we’re gonna put some sort of minder at your company to make sure that you’re not too mean to President Trump.” That was already a bridge way too far for me.

Now, for the government to come in and threaten individual broadcasters because of political speech that they made—that’s getting pretty close to as bad as it gets. This is something that I think all Americans should be paying attention to because, trust me, you do not want to live in a country where you no longer have the First Amendment.

Yeah, or to put an even finer point on it, I’ve seen a lot of people pointing out that you really do not wanna live in a country where the comedians on TV cannot make fun of the president.

Yeah.

That is rarely a sign of a healthy democratic society.

Yeah.

I also have been fascinated and disturbed by this whole crowdsourced surveillance and snitching culture—

Mm.

—that the internet seems to have created. I’m sure you saw this, but in the wake of the Charlie Kirk shooting, there were these accounts that would signal-boost people reporting their Facebook friends, their neighbors, the people who ran small businesses in their towns, saying, “Oh, this person made a tasteless joke.” Then they would get retweeted by Libs of TikTok, or they’d send it to some Charlie Kirk “people saying mean things about Charlie Kirk” database that would publish it, and people would write letters to their employers urging them to be fired.

That’s not new behavior. That has been happening since at least the mid-2010s with Gamergate. But this kind of incentive structure, where social media and these influencers are incentivizing regular people to engage in this kind of surveillance and reporting—I have seen that before, but I have not seen it at this scale or in this organized a manner. So I wonder what you make of that.

To me, this is just another platform-incentives thing, right? This kind of surveillance and doxing is essentially a kind of video game that you can play on X, and people like to play video games, right? Because you’re playing with people’s real lives, it feels really edgy and cool and fun for those who are participating in this.

Candidly, part of this is the flip side of free speech, right? If you want to say that people have the right to express themselves online, one thing they can express is, “This person made a tasteless joke, and I think that they can lose their job.” I’m willing to accept that as a fact of life as long as I still get to engage in political speech. You know what I mean?

Mm-hmm.

I think we also need to start thinking about what kind of trade-offs we want to have in a world where these platforms are mediating so much of our reality.

Yeah. I think that’s right. I just wish I felt like there was any stabilizing force, anyone in any position of power right now who could take down the temperature, turn down the knobs, and encourage people to take a deep breath, either literally or spiritually.

I am feeling very anxious. I’ve just noticed my anxiety over the past week has really gone through the roof, and I think it’s because I feel this sort of ambient rising of rage in our society, and I want it to stop. I want people to take a deep breath.

People should absolutely take a deep breath. I would say, if you’re looking for reasons for optimism, we are now a week out from this shooting. There have not been huge upsurges of violence across the country. I think people are very sad and very anxious about the future, but so far they’re not taking to the streets with guns.

I think that speaks to the fact that the vast majority of Americans do not want to participate in a violent cultural war with people who disagree with them. When we get most of our information from platforms, we are always at risk of forgetting that because, on a platform, it always looks like civil war is about to break out.

So, if nothing else, I think this is a moment for all of us to think about our media diets and the way that whatever media we’re consuming is manipulating our emotions, and then decide if we wanna make a change there, right? Perception can become reality in too many cases, and it just might be a moment to find something a little bit calmer to help you understand what’s going on in the world. For example, a podcast hosted by two friends who mostly like to tell jokes about technology companies. Just throwing that out there.

I mean, you’re joking, but I am thinking about our role in this moment and what we do, and I wanna continue to encourage people to look around, to touch grass, and to not just see the world through this kind of refracted fun-house mirror of social media, because I think when you do that, it leads to some really dark places.

Yeah.

Wow, I couldn’t help but end this one on a dark note. Sorry.

I know.

It’s where my head’s at.

I brought it to such a nice place, and then you were just like—

I know, and then I brought it back down into hell. I’m sorry. I’m sorry. Let’s talk about something else. Well, Casey, when we come back, we’re gonna talk with DYB about the UAE, GPUs, and AI.

Well, Casey, there was another big story this week involving politics and tech, and this one did not get a ton of attention relative to the Charlie Kirk assassination story. But I thought it was a big, important investigation from my colleagues at The Times that deserved more attention than it got.

I completely agree with you. This one is a really extraordinary piece of journalism, and I’m glad we’re highlighting it today.

5. The UAE Crypto Chip Deals

Yeah, so this story involves 2 seemingly intertwined deals between the Trump administration, World Liberty Financial, which is the Trump family part-owned cryptocurrency company, and the United Arab Emirates. It raises a ton of fascinating questions.

Yeah, but it’s also a story about the future of AI and potentially AI safety. As you know, there are so many people all around the world who want to get their hands on state-of-the-art chips, but the United States has export controls in place that prevent some of our adversaries, or countries that we’re just worried about, from getting them.

Yeah, and I think it would be a really big deal in any other administration. I think it would’ve gotten months of airtime. It would’ve been all anybody would’ve talked about for weeks on end. There would’ve been hearings and calls for impeachment and things like that.

But I think we’ve just gotten so inured to the everyday business dealings of the Trump administration and the Trump family and the various associates that it sort of passed by without much notice.

Yeah, this is the sort of story that in the old days we would’ve been on Twitter saying, “This is not normal,” with a bunch of those clapping-hands emojis in between all of the different words.

Yeah.

So if you’re a person who likes to know when this is not normal, we’re here to tell you today: This is not normal.

This—

This—

—is—

—is—

—not—

—not—

—normal.

—normal.

So to talk about this very abnormal story with us today is one of the reporters who wrote it, my colleague and friend of the show, David Yaffe-Bellany.

David Yaffe-Bellany, welcome back to Hard Fork.

Thanks so much for having me.

So, DYB, this story that you reported, this very interesting and complicated story, starts as all great tech stories do: on a superyacht. Specifically, a superyacht off the coast of Sardinia, where you wrote about a meeting that took place this summer between the owner of the yacht, a man named Sheikh Tahnoon bin Zayed Al Nahyan, a member of the ruling family of the United Arab Emirates, and Steve Witkoff, President Trump's envoy to the Middle East.

Now, before we get into the contents of this meeting, I have to ask, DYB: What is a superyacht? What makes it different from a regular yacht, and how can I get invited on one?

Doesn't it just have guacamole and sour cream?

Well, did you guys read that New Yorker story from a few years ago, “The Haves and the Have-Yachts”?

That's very good, yes.

That's my main reference point for all yacht-related questions. I believe that there are actually very formal distinctions between yachts, superyachts and megayachts, and that has to do with the length of the vessel.

Oh, wow.

But clearly you guys aren't spending a lot of time on the coast of Sardinia.

I heard it wasn't about the length of the vessel, but how you use it.

It's about the motion of the ocean.

That's right. Anyway, we're getting off track here.

We're getting off track. Okay, so, DYB, let's talk about this meeting. What were Sheikh Tahnoon and Steve Witkoff there on this yacht to discuss?

The White House told us that this discussion centered on resolving international conflicts, essentially. But what our reporting showed is that the relationship between these two men spans both diplomatic matters like that and a business partnership between Steve Witkoff's family company and Sheikh Tahnoon's family investment fund in the United Arab Emirates.

Yeah. So your reporting, DYB, and that of our colleagues centers on these 2 big deals that took place earlier this year. Tell us about these deals.

One deal involved crypto; the other deal involved artificial intelligence. The first deal, which is the crypto deal, involved World Liberty Financial, the Trump and Witkoff families' crypto company, and an investment firm in the United Arab Emirates called MGX, which is essentially owned and run by the royal family of the Emirates.

The deal was that MGX would use the World Liberty Financial stablecoin to make a $2 billion investment in another party to this transaction, Binance, the big crypto exchange. And this was a huge deal for World Liberty because the company is trying to get traction for its stablecoin. As you guys know, the way stablecoins work is that if somebody is using the coin, then there are deposits backing it up, and the issuer of the coin can invest those deposits in ways that generate a yield. It's a very profitable business in the crypto world. So that was the first of these 2 deals.

You suggest in your story that just by getting that $2 billion infusion of cash, World Liberty will likely be able to make tens of millions of dollars in interest.

Yes, exactly. Companies like Circle and Tether in the crypto world have found this business model is really profitable and, frankly, not that complicated. You just sit on these deposits. You can invest them in fairly safe ways to generate a big return if the principal is large enough, which in this case it was: $2 billion.

So tell us about deal 2.

Deal number 2 is the artificial intelligence deal, and this is a tentative agreement between the United States and the Emirates for the US to allow the export of American-made chips to the UAE. These are the most valuable, powerful chips that fuel AI. They're developed by Nvidia, and without them, you can't really become a tech powerhouse in the way that the UAE wants to be.

The Emirates has been trying to get access to these chips for years and years, and the issue has always been that US national security officials worried that if we gave these chips to the UAE, they would end up in the hands of China, because there are really tight ties between those countries.

Yeah, tell us about some of those ties.

Joint military exercises and technology-sharing agreements. At the center of this is G42, which is another Emirati tech firm, also run by Sheikh Tahnoon.

The fact that the Emirates is getting these chips was a big deal and not necessarily an inevitability, given the posture of the Trump administration toward China and the resistance in the Biden administration toward giving free access to these chips.

And there was resistance in the Trump administration as well to giving the UAE those chips. Is that right?

Absolutely. This is something that we found in our reporting: There were disputes within the White House and people who were very concerned about the national security implications of allowing access to these chips.

I should be clear that this is a preliminary agreement, and there might be new security guarantees built in when those chips actually leave the US and go to the UAE. But at the moment, there are definitely concerns from some people in the national security community.

Right. David, let me just repeat back what I think I heard from you describing these 2 deals, and maybe draw some lines that are suggested but not explicitly stated in your reporting. So correct me if I get any of this wrong.

That's fine.

The UAE wants to be a global player in AI. They look at this technology and say, “We want to be one of the countries that's developing these frontier models.” But it has a problem: It can't get the chips that it needs to run these powerful models, because under the Biden administration, the export of the most powerful Nvidia chips was limited to the US and its allies.

The UAE, while technically not an adversary of the United States, also does some stuff with China, and so people worried that maybe if they sent some chips to the UAE, China would get access to them. So the UAE can't get all the kinds of chips it wants, and then it plows $2 billion into a cryptocurrency controlled by the Trump family business. Around the time of that deal, the Trump administration changed its posture about chips, rolled back some of these restrictions, at least in a tentative agreement, fired the main guy who was advocating for these export controls—and we'll get to this—and lets the chips flow to the UAE as it had wanted.

Am I missing anything important?

I would say that the main thing that we say in our story, which is important to understand, is that we did not find evidence of a direct quid pro quo here. What the story basically shows is that these 2 deals involved many of the same people, were negotiated in roughly the same time period and were interconnected in ways that people weren't aware of when these back-to-back announcements happened in May.

I'll give you an example of that from our reporting. An employee at G42, the big Emirati tech firm that wanted the chips, was simultaneously working for that firm and for World Liberty Financial while all of this was going on. And so you see this intermingling, and it raises all sorts of concerns about ethics rules and conflicts of interest.

It raises questions about whether US interests are being subordinated to the commercial priorities of the Emiratis and the families of the people in the Trump administration. All of that is at play in our story, but we certainly can't say that there was an explicit quid pro quo.

6. Trump Business and Government Collide

You can't say that, but just this character of Steve Witkoff, who is President Trump's envoy to the Middle East, this longtime golf buddy and friend of Trump's: At the same time that he is the envoy to the Middle East, where he is supposed to be representing America's interests, he is also a co-founder of World Liberty Financial, which is about to reap the rewards of this $2 billion investment. Is that right?

Yeah, absolutely. He still has a financial interest in World Liberty, even now. His financial disclosure form was released just a couple of days before the story came out, and it showed that he still has an interest in World Liberty, despite statements by World Liberty that he was going to divest. We were told he's still in the process of divesting.

Yeah. How unusual is it for an envoy like this to also have an interest in a business that is benefiting directly from a foreign government he is directly dealing with as a diplomat?

These sorts of overlaps between personal business and government responsibilities don't have much precedent in the US. This is a new thing in the Trump administration that's becoming increasingly common. These lines are blurring.

Even before the Trump administration started back in December, Steve Witkoff flew to the Middle East and was taking diplomatic meetings.

He appeared at a crypto conference where he met with a World Liberty Financial investor. At that conference, he was presented as the U.S. Middle East envoy, even though Trump hadn't actually taken office yet. This sort of blurring of lines started immediately, and it's continued throughout this administration.

Now, let's talk about some of the other figures involved in this story. In particular, I want to ask you about the role played by David Sacks, who's the AI and crypto czar in the Trump administration. He was also, as you report, involved in some of these deals. So what is David Sacks's role in all this?

David Sacks is the AI and crypto czar, and he's interesting because, similar to Witkoff, he has a foot in both worlds. He's simultaneously working in government, and he has this tech job at Craft Ventures, his VC firm. What we found in our reporting is that he was one of the driving forces behind the chips deal with the UAE. He was pushing for it aggressively in public and also behind the scenes within the White House, and that raised concerns from some of his White House colleagues about potential conflicts of interest.

The reason being, Sacks is a tech investor. He's still working as a tech investor. His company has invested in AI ventures in the past, and 1 of the original backers of Craft was another Emirati investment firm that is currently chaired by Sheikh Tahnoon. Those concerns bubbled up within the administration while this deal was being negotiated.

I should also say that Sacks received ethics waivers from the White House that basically said, "You're divesting from a lot of these stakes. What's left is de minimis and isn't going to affect your work, and it's okay for you to play in these areas that intersect with your business." So he did get that sort of permission slip from the Trump administration, but there were still concerns from colleagues.

I have to say, very little is funnier to me in this moment than the idea of getting an ethics waiver from the Trump administration. Just what is that process like?

Okay, I have 1 more character that I need to ask you about, David—another David, David Feith. He was Trump's senior director for technology on the National Security Council, a member of the Trump administration who was known inside the government as the export controls guy for chips—the guy who thought it was a really bad idea for the U.S. and its companies to be shipping the most powerful AI chips overseas. This was a view that he shared with some folks in the Biden administration and its National Security Council.

He was fired earlier this year as part of a clearing of the house at the NSC by President Trump. So what did you report about that firing and what may have motivated it?

That was a crucial moment during these chip negotiations, where a potential roadblock to a deal going through was abruptly removed. It cleared the field in a way that allowed David Sacks to eventually take control of these negotiations. So it was definitely a good thing for the Emiratis and their priorities that this happened.

The reason that he got fired is that Laura Loomer, the right-wing agitator, had a meeting with Trump in which she urged the firings of Feith and several other national security colleagues. The reason that Loomer gives for this is partly that Feith's father, Doug Feith, was aligned with the neoconservative, Bush-era Republican Party, which she advocates against. So she essentially argued to Trump, "You've got to get rid of all these people for that reason."

We should also say, for people who are not familiar with Laura Loomer and her work, this is a woman who has been banned from every major tech product, including—my favorite Laura Loomer fun fact—she's banned from both Uber and Lyft for being too racist.

Mm-hmm. Mm-hmm.

Do you know how hard it is to get banned from Uber and Lyft for being too racist?

Yeah. Think about how many racists use Uber that still have their accounts.

Anyway, so Laura Loomer, for some reason, gets a bee in her bonnet about this guy, David Feith, and talks to Trump. Then Trump fires him and a number of his other colleagues.

This is the part of the story where something—a light bulb—went off for me. I remember earlier this year when the Trump administration seemed to change its posture on AI and AI proliferation very suddenly. This was an administration that had a lot of people in it who were very hawkish on China, who saw us as being in a race with China to develop powerful AI systems.

If you are in a race with your biggest adversary to develop a technology that is dependent on powerful chips from Nvidia, you do not want your adversary getting those chips from Nvidia. That seemed to be a pretty accepted part of the Republican tech policy platform.

Then, all of a sudden, the tune changed. David Sacks and other people in the administration started talking about how it was actually good if these American Nvidia chips got to China and all of our other adversaries, because we wanted companies and governments to be building AI on top of our technology. That was sort of their America First strategy.

I could never quite understand why that change had occurred so quickly or what was motivating it. And David, your story helps me connect some dots that had been missing.

Yeah. That moment where the Trump administration's tone on this shifts is really important to what's going on. Again, this reporting is really difficult and complicated, and I'm sure there's more about these 2 deals that will come out over the coming years. We don't know exactly what went into all of the decisions behind the scenes, but I think what we were able to show is, broadly, the timelines of how these 2 deals developed and how they intersected.

Tell us—give us a little flavor of the behind-the-scenes reporting process, because there are a lot of bylines on this story. It's you, Eric Lipton, Bradley Hope, Tripp Mickle and Paul Mozur. What can you tell us about how this story came together?

This is 1 of those big projects that requires a lot of collaboration among different departments within The Times. Not only is it 5 different reporters on the byline, with a bunch of others who contributed as well, but there are 3 different continents. People in different parts of the world have access to different types of sources.

The reporting for this story really began for me when I was in Dubai in late April and early May, which is when Zach Witkoff, Steve Witkoff's son, went on stage at a crypto conference and announced this $2 billion deal. I was there in the room for that and covered it at the time. From the moment it happened, we wondered, how did this come together?

It was just so unusual for a foreign leader to be channeling money to the family of a top White House adviser. That's just not the sort of thing that happens in the United States. In a way, these interconnections between business and government are more familiar to the way business is done in the Persian Gulf, and have been for years.

One interesting thing that Zach Witkoff said on stage at that event is that we should really take a page out of the Emirati royals' book, essentially. We should try to be more like them.

That's when the reporting started for me, and then it became this really fruitful collaboration among people with different areas of expertise all over the world.

Did you get to try any of that Dubai chocolate while you were over there?

I saw the Dubai chocolate. I looked at the price tag, wondered whether it would fit within The New York Times' expense policy, and passed. I didn't get an ethics waiver to buy the Dubai chocolate.

Got it. How has the White House responded to this reporting? What did they say? Did they acknowledge that any of what you had reported was newsworthy or interesting?

Both the White House and World Liberty Financial told us that there was no connection between these 2 deals. One was government business, and 1 was private business. They weren't connected. That was a statement that both parties made.

The White House also told us that David Sacks didn't have any financial stake in the UAE chips deal, that he didn't know the key Emirati players in that deal before it came together, and that he was really just focused on advancing the administration's priorities.

The White House also told us that Steve Witkoff is working with ethics lawyers to make sure that he is in compliance with all the relevant rules, and that he's still in the process of divesting, 8 months into the administration. So that was basically the response. I should say, the Emiratis also told us that a G42 spokesman said they have protocols to protect against conflicts of interest and that type of thing.

Right. So let’s take a step back from this and zoom out a little bit. David, you’ve been covering crypto for a long time. You’ve been on this show many times to talk about the various entanglements that the Trump administration has with the crypto industry.

It seems to me like what this story really illustrates is that crypto has become the primary vehicle for influence in the Trump administration. Basically, people around the world are starting to figure out that if you want to change the Trump administration’s mind about something or influence its policymaking, one way you can do that—and potentially a very direct and effective way—is by buying a bunch of stuff that the Trump family has a business interest in through this World Liberty Financial crypto company. Is that true?

It has certainly become a pattern in the crypto industry that players who want something from the U.S. government—and virtually everybody in the crypto world wants something, because it’s such a nascent area, with new regulations in the works and all sorts of things that everybody’s pushing for—are now connecting themselves to the Trump business in some way, whether that’s investing in the World Liberty coin or attending the meme coin dinner.

Again, in none of these cases can you point to an explicit quid pro quo, but it’s just become part of the way that business is done in the crypto world: You at least consider some sort of collaboration with the Trump family apparatus.

It reminds me of the first Trump administration, when there were all these stories about the Trump Hotel and people choosing to hold their big events, banquets and dinners there, and booking big blocks of rooms when they would visit Washington, in part because they thought that Trump or someone in his family might notice, “These guys are really supporting our hotel here.”

This is just a much more direct route to getting the attention of the Trump administration. You can say, “I’m going to buy $2 billion worth of your stablecoin from you.” It’s really hard to book $2 billion worth of anything at one hotel, but you can go online and buy $2 billion worth of a stablecoin.

The hotel is a good reference point. It’s sort of like the Trump Hotel on steroids, and it’s way more global, too. Anybody anywhere in the world can buy the Trump meme coin or buy the various tokens associated with World Liberty. In previous reporting, we found that there are huge numbers of foreign buyers of all of those things.

And remember, foreigners are restricted from making campaign contributions. So this is a way for people in other countries to support the president without going through the channels that have traditionally existed and been restricted to them.

David, I’m curious: What has the reaction been to this story?

The reaction has been really interesting, and it’s been similar in certain ways to how previous reporting we’ve done on conflicts of interest within the Trump administration has been received. A couple of Democratic senators put out statements condemning these sorts of business dealings. You’ve got various good-government groups and ethics lawyers sounding the alarm over this.

But ultimately, in the Trump era, those Democrats and those sorts of ethics lawyers are kind of impotent. They don’t have the ability to turn this into a giant investigation, and so there’s a limit to what can happen.

We did a previous story about World Liberty Financial and the conflicts of interest associated with it, and that contributed briefly to a delay in the stablecoin legislation moving through the Senate as Democrats sounded the alarm. But ultimately, that legislation passed, and with Democratic votes. So it remains to be seen whether the left will take a major stand on this issue.

Yeah. Well, David, fascinating reporting. Really important work. I hope you keep going, and next time bring us some of that Dubai chocolate.

All right. Thank you so much.

Well, Casey, that’s enough politics for today because, as with every week seemingly, it has been a huge week in AI. So it’s time for our segment, “This Week in AI.”

Before we get into the big AI news of the week, let’s hit our disclosures. I work at The New York Times, which is suing OpenAI and Microsoft for copyright infringement.

Yes, and in “This Week in Boyfriends,” my boyfriend works at Anthropic.

“This Week in Boyfriends.”

Honestly, it would be a great segment. Boyfriends are always up to something.

7. AI Moves Into Everything

Okay. Well, Casey, the big AI stories of the week. Let’s go through them. Number 1: Business Insider tells reporters they can use AI to generate first drafts of articles. This was first reported by Oliver Darcy in his great Status newsletter.

This is about an internal memo that went out to the staff of Business Insider last week from editor-in-chief Jaime Heller, telling them that they could use AI not just for research and various other things, but for writing the first drafts of their stories, as long as reporters, quote, “Make sure your final work is yours.” Casey, what did you make of this?

I’m very curious what she means by “Make sure your final work is yours.” The AI did all of the research and wrote the first draft. What did you do, change the placement of a couple of commas?

Look, I’m very curious to see what happens with this one. The cynic in me suspects that this sort of thing is going to become a lot more common. As AI writing tools improve, I think more reporters are going to feel like, “Maybe I should let it take a stab at a first draft.”

I should say, for myself, I do use AI systems to edit my copy—not to do the writing, but to catch typos and some factual errors. So there is a spectrum of what you and I think is okay right now. This goes further than I would, but this seems like one of those frontiers that keeps advancing, and before long, a lot of other folks are going to do it.

You know, right now, I think that letting reporters draft the first versions of their articles with AI is actually kind of risky from a reputational standpoint.

Mm-hmm.

Look, the AI writing still has various stylistic flaws in it. You can kind of tell when you’re looking at something—

I can always tell. I can always tell.

I don’t think you can always tell, but I think you can sometimes tell, and I think that’s a big problem. If readers start to find evidence of AI-generated content on your website, I think they’re just going to start losing respect for you as a publication.

But I think probably at some point the writing will get good, and then we won’t be able to tell anymore. I think this will just become part of the background noise of the internet.

I don’t know. If you’re a reporter and you start doing this, I think you have to ask yourself the question: Am I participating in automating away my own job? There is not that far of a bridge from AI writing the first draft of my story to AI just having the job that I used to have. So Business Insider is playing with fire.

Well, it’s really interesting that Business Insider in particular is the outlet that’s pushing the vanguard here, because Business Insider and several other publications had to take down articles from their sites earlier this year when it was discovered that they had published fabricated stories that supposedly came from a freelancer who had generated them using AI.

So they want their own staff writers to use AI to create the first drafts of stories, but if freelancers use AI to submit fake stories, that will apparently not be okay.

That's very stupid. This whole thing—we're just in stupid town right now.

Well, Casey, to continue on the theme of AI and media, I want to bring you the next story this week in AI, which is about Penske Media.

Mm.

Penske Media, the parent company behind Rolling Stone and The Hollywood Reporter, is suing Google over its AI Overviews. The company says that Google is illegally using its reporting in the AI-generated summaries and, as a result, depressing its online traffic.

Penske claims in the suit that revenue from affiliate links—that's the kind of e-commerce link that supports a lot of the websites—

Mm.

—that recommend products online—they say that the revenue from affiliate links was down by over a third at the end of last year compared to its peak, and they attribute that directly to a drop in traffic from Google. Casey, what did you make of this?

The statistic you just pulled out is the most interesting part of this lawsuit to me. We talk to Google all the time, and we've been asking them for over a year about this exact possibility, right? What happens as you shift more people to AI Overviews? What happens to all the publishers that rely on our traffic?

And we reliably get this message from Google: “Casey, Kev, we send billions of clicks every single day. It's very important to us to have this thriving ecosystem.” But then you get down to the level of the individual publisher, and it's like, “Well, a third of our revenue is gone,” right?

So I'm very interested to see where this case goes. I can't say for sure that what Google is doing is illegal, but I do know that this is bad for the overall health of the web and that Penske is not the only suffering publisher here.

Yeah.

All right. Well, Kevin, this one caught my eye. Albania has appointed an AI minister.

Yeah?

This comes from the BBC. Albania has a new AI minister named Diella, which apparently means “sun” in Albanian—the sun like the star, not the son. Diella is going to be tasked with ensuring that Albania will become, quote, “a country where public tenders are 100% free of corruption.” So, an anti-corruption minister made out of AI.

What do you make of that?

Well, I was going to say that sounds like a terrible idea, but after the story we just heard from David Yaffe-Bellany, I'm ready to replace all of our ministers with AI too.

I do think that ChatGPT in general has a stronger moral compass than some of the people now running the U.S. government. I'll say that.

No, look, this is one where I think AI can be useful for a lot of things. I'm not sure it's quite good enough yet to put into a role monitoring public corruption.

Totally.

Do you remember when we had that guy on who ran for—what was it—mayor?

Yeah.

Using AI?

Yeah.

And he was just going to be the human face of the chatbot and do all of the votes that the chatbot recommended and put into place the policies that the chatbot suggested. I think there's probably some meaningful percentage of lawmakers today who are using ChatGPT to help them decide what to vote on and how to vote and how to run their piece of the government.

So, yeah, I think it is inevitable that someone will just say, “Let's cut out the middleman and just put the AI in power directly.”

Hmm. Well, at least in this case, do you know who I would have put in charge of overseeing corruption in Albania?

Who? A person of prominent Albanian heritage: Dua Lipa.

Really?

I think she would've done a great job.

I didn't realize she's Albanian.

Yeah, absolutely.

All right, what else is in the news, Kevin?

Well, our next story comes to us from the world of Roblox.

Ooh.

That is, of course, the hit game-creation platform, where, according to several accounts that track Roblox activity on X, a new game recently broke the all-time record for most concurrent players on the platform. Something like 23.4 million people were simultaneously playing a game based on the viral AI sensation Italian Brainrot.

Italian Brainrot, I think, was one of our favorite finds of the year, and you're telling me it's a video game now.

Yes. I guess you can build games inside Roblox. I'm not myself a big Roblox user, but I did spend some time on YouTube trying to understand this phenomenon yesterday. I watched a video about a guy who spent 24 hours playing Steal a Brainrot inside Roblox.

That sounds good for you.

Yes. The video thankfully was not 24 hours long.

Basically, this is a video game that a Roblox user—or one user in particular—has created where you go around stealing people's characters from the Italian Brainrot franchise, which, as you'll remember, you introduced me to earlier this year, was created on TikTok as a self-referential form of Zoomer humor.

So is this a play on Pokémon, except instead of stealing Bulbasaur, you're stealing Tung Tung Tung Sahur and Ballerina Cappuccina?

Yes. You can go around and buy various characters with your Robux. You can buy a Nubini Pizzanini and use it to do various things. You can then go around stealing other people's characters that they've bought.

This has become a huge sensation. 23.4 million people is massive, and I think this is a phenomenon that is essentially invisible to people who are over the age of 18. But if you are a teenager, this has been a very big deal.

I think it goes to one of the things that you've been saying on this show for a while now, which is that there is just a generation of people who love AI-generated content, who are not bothered at all by any of the moral or ethical or copyright concerns, who are just like, “Yes, I love Bananini Chimpanini,” or whatever.

Kevin, it's Chimpanzini Bananini. Please show some respect.

And I am going to devote many hours of my life to playing a video game in Roblox inspired by these characters.

Very interesting. We'll have to give that a shot.

I have one other wrinkle to bring you. This is a late-breaking wrinkle in this story—

Mm. Mm-hmm.

—which is that, apparently, the Tung Tung Tung Sahur character has disappeared from the Roblox game due to a copyright dispute.

Really? Who owns the copyright on Tung Tung Tung Sahur? Is he part of the Marvel Cinematic Universe now?

Well, according to Dexerto, this character was removed from the game by an agency representing the creator, Noxashed, who claimed that the Steal a Brainrot developer, Sammy, pulled the character while legal discussions over licensing were ongoing.

Wow. What more iconic move could you make than literally stealing Italian Brainrot from the Steal a Brainrot game? I think that person won the game, I have to say. Congratulations.

8. OpenAI Builds a Teen Mode

Kevin, here's a more serious one I'm curious to get your thoughts on. OpenAI is launching a new ChatGPT experience for teens. In recent weeks, as we've talked about on the show, there has been more scrutiny of OpenAI over the way that younger and more vulnerable people are using the platform. Some of those people have died by suicide.

There was a hearing in the Senate about these issues this week. On Tuesday, OpenAI said it's building a system that will try to identify users who are under 18 and will give them a more limited version of ChatGPT that will block graphic sexual content and, in rare cases, potentially involve law enforcement. So what did you make of this story when you saw it?

Yeah. I thought this was largely a positive development. I am very worried about the increasing frequency with which young people, especially, are using these chatbots as emotional support companions and forming intimate relationships with them. As we've discussed, for some set of them, this does really appear to be taking a toll on their mental health.

So I am all in favor of OpenAI doing what it can to determine which of its users are under 18 and to give those users a different experience.

They have also been under some pressure. Last Thursday, the FTC launched an inquiry into these AI chatbots, including OpenAI's, and the company is also facing a lawsuit from the family of Adam Raine, a 16-year-old boy who died by suicide after talking with ChatGPT. So this did not come out of nowhere.

They have been under a lot of pressure to do something for their younger users, but I think this is probably a good and overdue step.

Yeah, I agree with that. At the same time, I worry that this truly is not going to be enough. Right now, I think one reason why younger people and vulnerable people are talking and sharing their most intimate thoughts with ChatGPT is because they're confident that those communications are private.

Now that they know that those could potentially be shared with your parents, with law enforcement, my guess is those teens and vulnerable people are simply going to go elsewhere. They're going to go to other platforms that don't have those same protections.

So while I agree with you, I think this is the right thing for OpenAI to do, I think we need to look around the industry at platforms that do not have these protections and ask them, “What are you going to do?” And even after all of them do it, there's the risk that teenagers and vulnerable people are just going to start talking to large language models that live on their laptops and don't communicate with anyone, right?

So the horses are sort of out of the barn here in a way that makes me nervous, but I am happy to see at least one company doing something about it.

Yeah. Look, I think this is very much a secondary element to the story here, but I am also interested in how OpenAI is going to try to determine the ages of its users. It is apparently not going to do what lots of other internet services do, which is just put a little box on the sign-up page that says, “Hey, tell us how old you are or when your birthday is,” and then trust that people are not going to lie their way around that.

It is instead going to try to infer a user’s age from the things that they’re chatting with ChatGPT about. So if you’re asking about your eighth-grade math homework, it might flag you as being underage and show you the kind of nerfed, youth version of ChatGPT instead. I did see some speculation that maybe teenagers are going to try to get around this by talking about back pain and mortgage interest rates and things like that so the system thinks they’re middle-aged.

Say, ChatGPT, my wife—she’s busting my chops. You gotta help me.

But I think we’ll have to see how that goes.

9. How People Use Chatbots

All right. Well, here’s one more, Kevin. Both OpenAI and Anthropic put out new reports this week on how people are actually using their chatbots.

Yeah, I thought this was really interesting. This is the kind of data that’s been relatively sparse. We just don’t know a lot about how people are actually using these chatbots, and the companies that make the chatbots are the ones, of course, with access to the best data about that.

But we got reports this week from both OpenAI and Anthropic, who talked about the various ways that people are using their models. We should also say that these are not exactly apples to apples. OpenAI’s report just looked at the consumer version of ChatGPT, so not counting all of the business and enterprise users who are using it through the API. Anthropic’s report did include both the consumer Claude version as well as the more enterprise-focused coding tools.

But let’s talk about a few things from this. One thing that stuck out to me was that with ChatGPT, there is a closing gender gap—

Hmm.

—in usage. So a year or two ago, it was the case that many more men than women were using ChatGPT, but now, according to OpenAI, more than half of ChatGPT’s users appear to be women, which is surprising to me.

That’s so interesting. Now, did they share any speculation on what has managed to close that gap?

No. The researchers who conducted this study did not give any hypotheses for why this gender gap has narrowed, but they did say that by July 2025, the share of ChatGPT users with typically feminine names had risen to more than half of all ChatGPT users: 52%.

They also said that this has become a fast-growing tool in low- and middle-income countries. Maybe there’s some relationship there. They said that by May 2025, ChatGPT was growing much faster in the lowest-income countries than in higher-income countries.

So there seems to be an expansion of use of ChatGPT both among women and people outside the richest Western countries.

It’s interesting to see. I would say that if there remained this huge gender gap that we were seeing a year or two ago, to me, that would be a really negative signal about AI, right? Because if this technology is as useful and general-purpose as executives are always telling us that it is, there should actually be relative gender parity in terms of who’s using it.

Yeah.

What else was in this report?

Another interesting thing from the OpenAI version of this report was that something like 73% of ChatGPT messages are not work-related.

Hmm.

This was surprising to me.

Hmm.

I had thought that maybe the usage of ChatGPT for personal and work-related tasks would be split pretty evenly. I certainly would guess that my own usage of ChatGPT is about half work and half non-work.

Hmm.

But apparently, the non-work-related use cases of ChatGPT are growing really quickly, and they comprise now the majority—about three-quarters—of all ChatGPT usage.

To me, that just speaks to the challenge that Google has here, right? I think that a lot of those queries are things that used to go into the Google search engine that were not about work, but were just sort of like, “Hey, how do I fix my doorknob?” Or, “Tell me about Canada, because I’m going there next week.” And all of a sudden, ChatGPT is serving a lot more of those queries.

Yeah, actually, that is the number 2 use case in this study: as a replacement for web search, what they call seeking information. What do you think the number 1 use case of ChatGPT is?

Cheating on homework.

No, they did not break that out as its own category.

Oh, okay.

It’s what the researchers called practical guidance. Basically, help me make this decision, help me think about this conversation. It’s the more companion-like use cases that we just talked about being worrisome.

Yeah, the things we used to rely on friends and family for but no longer have to.

Exactly. So that’s the OpenAI report. The Anthropic report emphasized something totally different. They found that adoption of Claude was being driven primarily by coding and what they called automation usage patterns, in businesses using Claude through the API.

Anthropic also found that their highest Claude use per capita was in Singapore and Israel. And within the U.S., the leading per-capita users of Claude were not in California or New York, as you might expect, but in Washington, D.C., and Utah.

Hmm.

The Mormons love Claude.

How about that? To me, what this suggests is that OpenAI and Anthropic are really diverging in who their customer bases are. OpenAI wants everyone. It’s like a ChatGPT in every house, on every phone.

Anthropic is increasingly going after businesses and enterprises. So my assumption is that as you fast-forward a year, you’re going to see that trend continue, where ChatGPT becomes ubiquitous and Anthropic becomes a thing that you use if you’re building something with code.

Yeah. So let’s keep tabs on this data over time. This is the kind of data that I think we need more of from all of the big companies, and I think it’s important that they do this collection and analysis in a privacy-protecting way.

Both OpenAI and Anthropic have said, you know, it’s not like we are looking line by line at users’ conversations. They’ve built these tools to anonymize and collect insights about users’ conversations without peering into them. But I think this is very valuable. We actually do need to know what people are using these things for.

I’d really love to see it for Grok, by the way. What would come out ahead? Would it be gooning toward anime companions or racism? I think it would be 1 of the 2.

Yes. There’s a strong uptick in usage from the MechaHitler-adjacent region. What’s that about?

What’s it about?