U.S. Bans New Chinese Humanoids, Zuck’s Op-Ed, eBay's $56M Lawsuit | Diet TBPN

The US government has banned Chinese humanoid robots. The US is banning new imports of foreign-made humanoid robots, quadruped robots, robot dogs, you might know them by. Very controversial based on— we got to get General Catalyst to weigh in on this because aren't they anti-robot dog? Wasn't that the whole thing of the video? Maybe I'm misremembering.

That's right.

So they must be celebrating.

Some platform VCs are very pro.

Yes.

Robot dogs.

Everyone should be pro-robot dog. Anyway, it's— there's nothing wrong with a robot dog unless it's a geopolitical discussion, and then there's another layer to the debate. So the Federal Communications Commission, FCC, today announced some restrictions, and these apply to new versions of those products. FCC Chair Brendan Carr said the move is intended to, quote, secure America's critical supply chains. China, meanwhile, has condemned the decision as protectionism. Is that because they're protecting the, like, the small children from getting smacked in the face by the humanoid robots? It is protection.

I know, I know that video. I think that you're referring to—

yeah, brutal.

Absolutely brutal. We should try to pull it up. Try to find the humanoid kicking a child, please. Sounds wrong, but I just want to get at the truth. Yes, I just want the truth.

The move comes as Chinese companies have established a commanding lead in humanoid robot robotics. Of the roughly 15,000 humanoid robots shipped worldwide in 2025, China's Unitree and, uh, AGI Bot— AGI Bot, I'm not sure how you pronounce that— each delivered.

Okay, we have the video of the robot kicking the child.

Oh, brutal. There's another one we're spinning around and gives him sort of like a spinning back fist. But that is crazy.

And that is—

seeing that, you're kind of like, yeah, I'm up for some protection.

Gross. Negligence.

If it's going to, if it's going to protect children, I think it's maybe in the best interest. No, that's not why this is happening, of course. So 5,000 units were delivered from, from each Unitree and Ajobot. By comparison, U.S. companies including Tesla and Figure shipped only a few hundred robots each, if that. China is also estimated to control about 85% of the global humanoid robotics market and also a ton of the deep supply chain, obviously. According to the Associated Press, the restrictions are expected to add new tension to US-China relations ahead of a planned September meeting between Chinese President Xi Jinping and US President Donald Trump. The FCC argues the measures are necessary to reduce potential risks posed by connected foreign-made devices that could become part of critical American infrastructure. What's Interesting here is that we're seeing a ban on a Chinese technology that is so nascent that you don't have a coalition of companies that are saying, whoa, whoa, whoa, I've deployed, my business is built on top of Unitree. I know there's geopolitical considerations, but I can't make the cars that I make without them. I can't run my business without them. And so this is going to hurt me. No one's really stepping up in the same way that they are with Chinese LLMs, which a lot of companies are like, no, I'm actually gonna cut my inference cost by 80% if I can connect to that.

And you have all of the various cloud businesses that are already making money.

Making money from them.

These models.

Exactly.

Yeah, I'm, I'm very different. I saw a bunch of people that were sort of saying that they didn't like this, that it was, you know, protectionist.

Yeah.

And that there must have been some American humanoid or robotics companies that were pushing, lobbying for this in DC. Maybe there were, but I personally don't want 100 million humanoid robots in this country running around that are not made by a US company or at least a Western ally.

Yeah. Where— since there's been the DJI ban, there's been various surveillance tools that are banned, but there's still so many different pieces of technology that come from all over the world. I wonder where people would like to draw the line generally on technology. Like, is it okay if the Wi-Fi router is made in China?

We drew the line on Huawei.

Yeah, on Huawei, on the 5G towers specifically, but most, you can still get a router that, you know, a Wi-Fi router for your house that comes from China. And you have to imagine that many of the actuators, the motors, the batteries, CATL, CATL, that company, ships batteries for all sorts of stuff. The, the Google Waymo approach, I think, is probably the correct, like, finding the line. So Waymo is sourcing, like, almost all the parts for the new Waymos from China, but they are building them, I believe, in the United States, and all of the software is from Google or Alphabet or Waymo. And so you have to imagine that they're able to tear all these pieces down and sort of maintain their own cybersecurity posture, which is really what this is all about. But it is interesting. Do you think this actually accelerates the American humanoid market?

I mean, these companies have already raised a lot of money. Elon's putting billions into it. Figure's managed to raise billions. I think that the first thing that we need is profitable use cases for humanoids. And so I think there's a lot of work that the industry needs to do to make these, to make these products like genuinely valuable in the home, in workplaces, in factory settings, all these things. I think that's a bigger problem. I think if you, it does feel inevitable at this point just because there's so much money going into it, so many smart people dedicating all their time to it. It feels inevitable. Maybe it speeds it up a little bit. It just felt like we were going to end up in a situation where they would just continue to dump humanoids onto our, you know, you could buy, you probably still can buy Unitree robot on Walmart.

I was asking Tyler about this because you used to be able to buy Unitree G1, I believe, one of their robots on Walmart, which was sort of crazy. And the price was accessible. I mean, it was still tens of thousands of dollars, but it was something that a research lab could buy pretty easily. I wonder if there will be other ways for companies to get access to these products, at least for benchmarking, because they will be competing internationally, like even in a world where Tesla gets to the frontier of humanoid robots with Optimus and then they're going up against a Unitree for a deployment in Germany like BYD. BYD and Tesla compete internationally even though Teslas can still be sold in China but not vice versa. But internationally there needs to be a lot of like, what are the capabilities so you can actually sell against your competitor?

Yeah. So far everything Unitree I'm seeing on Walmart is out of stock. So I don't know if they just marked it out of stock when this went live or they had just sold out.

Those would probably spike if they hit Cars and Bids or Bring a Trailer. I think you throw a Unitree G1 last edition, final edition, banned, can't buy anymore.

Who's making Bots and Bids?

Bots and Bids, that might be it. I was listening to Elon Musk on The Economist. He sat down with The Economist for a long-form interview. There was a clip that went sort of viral for some some sharp jabs back and forth about the media and how the public perceives the media. But the much more interesting substance was about his vision—

You're describing the crash-out?

Yes, it was described as a crash-out. But that seemed like a very small section of what was actually interesting about the more than an hour they talked. One was, I mean, he talks about humanoid robots and he paints a very clear vision for the Optimus program. And it made me really hope that there's a lot of focus there because that feels like an industry where like he obviously has a very strong philosophy around xAI and Groq and you saw it in the SpaceX S-1 where the market for enterprise software was like $22 trillion or something like that. And so there is some value there. It might make sense from a business perspective, but it was, less inspiring than rockets, less inspiring than self-driving electric cars. Because you, as you go back, you think about SpaceX and you think, if he hadn't worked on that, probably wouldn't have a rocket company in America. And same thing with the electric cars, like, yeah, GM had the EV1, there were a couple other companies that were sort of like testing things, but we'd probably be 10 years behind if he hadn't sort of like sheer force of will that into existence. And it feels the same for humanoid robots where if he can really go and push that extremely hard and marshal all the capital and do all the difficult things to build the supply chain, make it happen, make it work, you can get to a place where there is a national champion in America. But if Grok stopped training and there were no more xAI LLMs, like there would still be a massive competition between OpenAI, Anthropic, Meta, Google, like there's plenty of really well-funded hyperscalers and leading labs that are going to provide that service of like ask a question and get an answer or fire off a prompt and get some code back. That the American humanoid robotics industry feels much more fragile in the sense that if you don't have this massive hammer of capital and will, coming to bear, you could just wind up with it never happening, which I worry about.

Yeah, and America needs to figure out how to make billions of robots itself.

I do wonder how this is going to play into all the other discussions that are going to happen at that US visit by Chinese leader Xi Jinping. He's scheduled to meet with US President Donald Trump in September. I'm interested to see, are they going to be treating this—

Can you see if we can possibly get him on the show? What if she was just like, yeah, I'm meeting with the president, but I'm also going on a 50-podcast tour while I'm here.

Get him on. Yeah, there's a bunch of good shows he could do. He should do—

Modern Wisdom. I want to understand his mindset.

Hot Ones would be great. I'd love to see him on Hot Ones.

Joe Rogan, of course.

Rogan is a no-brainer.

That should be kind of the last show.

But Nelk would be like the obvious choice.

Yeah.

A lot of world leaders have been on there.

That's true.

Controversial world leaders.

That's true. Why not Xi Jinping? Dean Ball's going back and forth with Will Manitas. Dean says, do we think the Silicon Valley leaders who were so up in arms about using Chinese LLMs last week will be similarly outraged about the US government's decision to ban import of all advanced robotics from all foreign countries? Is Manitas going to write an essay about how the US robot companies need to be nationalized in exchange for this policy favor. Will says, "You're welcome to tag me if you want a take." What's your take? He does ask him for it. And Will says, "Dean seems to not understand that the FCC is issuing a published rule under statutory authority with a waiver process and judicial review is different than a thinly veiled suggestion that regulators carry out an extralegal whisper campaign to protect the interests of his employer." The robotics companies need not be nationalized until they also ask for public infrastructure, preemption, federal backstop of their debt, and a swept market just so they can call themselves national institutions. When they do, send them my essay. And Dean Ball says, so wait, if the US straight up banned Chinese LLMs using explicit statutory authorities, e.g., supply chain risk, that wouldn't necessitate lab nationalization, but doing it through soft law would? What? Your instance. That my analytic prediction was a thinly veiled suggestion continues to be bad faith. I do wonder if we'll just see more— you know, Figure shows up in those secondary demand tables. I wonder if this is bullish for that. I don't know. 1X just seems to be on a good path. Sunday Robotics just seems to be on a good path. We've talked to a lot of other companies where the regulatory and competitive landscape has always— it hasn't seemed like it's a driver of financings, right? It's not like, oh, you're waiting for approval or you're waiting for some rule to break. It's like, yeah, at some point in the future, America is going to want an indigenous robotic supply chain.

Yeah, the compelling teams have raised anywhere from $100 million to billions of dollars. And now they need to effectively deploy that and get commercial use cases starting to ramp. And then I expect them to be able to raise an order of magnitude more capital each, right? I don't, again, I don't think capital is really a constraint here, but I think they've raised their sort of like, they've raised their like 5 rounds, let's say.

Yeah, yeah. Mark Zuckerberg is in the Wall Street Journal, paywalled. You gotta pay to read what Mark has to say, but he penned an op-ed and he says the US should accelerate AI development, not restrict it. He says the AI future is for everyone. There's some interesting quotes in here. He's taking shots. He's taking shots at doom, fear-based marketing. He says, it is surprising that the discourse from many of those who are developing artificial intelligence is so filled with doom. I don't understand why anyone who believes that AI will eliminate most jobs and much of humanity's relevance would rush to build that future. The only— the notion that AI is so dangerous that the only safe path is an extreme concentration of power seems dangerous. Historically, hoping that an absolute power will benevolently provide for humanity, if sufficiently enlightened—

Is he talking about machines of love and grace?

I think so, something like that. Hasn't led to safer, positive outcomes. Humanity has witnessed many transformative advances. Each time there is fear that people will be left behind, but each time humanity has come out with more people sharing greater prosperity, health, and freedom. We believe this will be true with AI as well, and the abundance of the future can be shared by everyone. He's abundance-pilled. And so I think it's good to sort of reset, you know, where he is. He's been, he's been, he's talked about AI a bunch, but it's never been like right on the cusp. But he's just like throwing his hat firmly in this camp of, He's obviously playing in AI but has a particularly positive vision. And he's sort of throwing down the gauntlet. If you think AI is so bad, then stop working on it and let me work on it.

Let me cook.

I don't know. But not too much chatter around this. It did stand in opposition. He's saying the United States should accelerate AI development. At the same time, there's this open letter from a bunch of lab employees saying, hey, maybe we should do a mutual pause. Slow down. And I think even some Meta researchers signed that as well. But there's like a wide group from OpenAI, Anthropic, a bunch of other firms, Google, that all signed. John Palmer, of course, ever the TBPN promoter, says, TBPN must publish an open letter against distillation after their model was distilled by every competitor in the industry. And I said, I'm advocating for a complete and total podcast ban. Multifederal year— multiyear federal approval process for new RSS feeds to get issued. He says, "Better hurry up and write that open letter before Chinese technology network launches." I think a Chinese TBP would actually be electric. I think it would be amazing.

I would love to watch.

I would love that. You know, just wildly different. They just don't do the going direct thing over there the same way. You know, there was like the Deepseek investor meeting, 4 hours, that was like leaked. That's not the same as actually sitting down with someone that has a particular point of view.

Sitting down with a stack of 200 posts and just chit-chatting.

I don't know, maybe it does exist. But when Dwarkesh went to China, it was such a funny observation. He had a bunch of observations, but the one that really stuck out to me was he was like, I asked everyone and it is clear China does not have a Joe Rogan. Like they don't have someone who is like a larger than life yapper political commentator, just like a guy who talks about everything, including politics. And there's political reasons for that, but it's also maybe cultural. And it's just interesting because the demand for that would be electric. I feel like you have to imagine that like Joe Rogan clips make their way around the world and are sort of like a barometer for where the American psyche is at in a certain way, like a certain group of Americans. Feel that way. There's this other post that's sort of interesting to see where the Anthropic back and forth on the open letter goes. Obviously Anthropic clarified their stance on open source, not advocating for a complete and total ban, but they're very worried about distillation. They're very worried about cybersecurity and they're very worried about authoritarian uses of AI. All good points. Chetan over at Benchmark says Anthropic very publicly claiming there needs to be a crackdown on distillation is puzzling. It's a $1 trillion company with incredible resources and the distillation attacks at the scale they describe would seem easy to detect. I don't know how easy they are to detect. I think the people, you know, harness up a whole bunch of different VPNs and stuff, but at the same time it's like, it's an AI company, put AI on the job, maybe you should be able to detect it. So we'll take his point.

He's asking AI to be able to find a needle in the biggest haystack of all time.

Yeah.

The world's largest haystack.

The final benchmark. But he says the only cost would be giving up the associated API revenue. So you should have a very aggressive, um, a very aggressive internal policy around if you're suspected of a distillation attack, you're kicked off the API. Of course, in a very short amount of time, you can fire off a lot of prompts and get a lot of tokens, and you can do that a lot around across, across a lot of different accounts. Uh, what do you think about this?

Yeah, I mean, detecting distillation seems like a very hard problem. I think if it was trivial, they would have just done it. And also, it seems like it's not that big of a percentage of their revenue. I'd be surprised if distillation attacks accounted for a large amount of—

Yeah, I don't think it's meaningful.

Also, so it does feel like it's hard because a lot of it doesn't actually go through the direct API necessarily. It goes through— a company that is vending a service and then someone is prompting that service and so you're further obfuscating it. Like there are a number of different sort of wrappers that effectively wrap various frontier APIs and if you're getting hit in a bunch of different— you're blending all of that together, you're looking at open source stuff. There's also just people that will— like this just happened with— was it Matt Schumer? He He basically uploaded to GitHub a game that he made with Opus 5 along with the prompt. And so you can sort of distill on that, like it's only one data point or, you know, a small bag of tokens, but there's a lot of people that will share a prompt and the actual output. And so you can train on that and there's no way to detect that you found that. What's interesting here is that the knock-on effect of Chetan's point is that if Anthropic takes Chetan's advice, you wind up giving out the model less frequently and being more closed about it, which is something 3 months ago everyone in the ecosystem was like, this is annoying. I'm a small startup. I can't get into the Mythos tier. I can't get access to Project Glasswing. I'm not an enterprise-level customer. But I still— so I just don't get frontier intelligence. And I just get steamrolled by Microsoft, who has access to it, or Apple, who has access to it, or whatever. And so this, the whole idea of like if an amazing intelligence, an amazing model is made, should it be widely available? That's actually at odds with distillation because you have to hunt those bad actors down and you play this whack-a-mole. So I don't know, it is tricky. We'll see where it goes. We'll see where it goes. Anyway.

Apparently over on Facebook Marketplace, people are adding a sort of gold aura to their listings. What is this? Would this make you more likely to buy?

John.

Oh, it does look like a rare drop.

Yes, for only $25, picking up this table with that level of aura, certainly tempting.

It does make me like it more. I think it actually works. I've seen people do this where they add, they'll add like a woman to it as like the salesperson basically with AI, and that draws more attention on Facebook Marketplace. I don't know if I've ever bought anything on Facebook Marketplace, but Maybe we should try and have Tyler do a trade a paperclip into a car challenge. Have you ever seen these?

Yeah.

Ryan Trahan and—

Tyler would be exceptional at that kind of thing.

You think you could trade up from something on your desk?

I think this would definitely help it.

Yeah.

Giving it some kind of—

So basically, you have to look for the most unoptimized listings that are undervalued. And then once you buy them, give them the ultimate glow up. On, on, on with AI images, put them back up for a higher price, increase your margins, start printing. This is a very Gary Vee-esque hustle for sure.

Pretty incredible story. eBay pays couple $56 million for its role in cockroach harassment campaign.

That is a crazy headline.

A couple received threatening deliveries, including funeral wreaths. After criticizing the e-commerce retailer in a blog. So this couple, David and Ina Steiner, were harassed, it sounds like, for years for criticizing eBay. Who knew that criticizing eBay could get you in this much trouble? Of course, it's eBay that ended up getting in trouble.

But so they had an online e-commerce blog called E-Commerce Bytes, and they criticized eBay. On that. And then back in 2019. So this has been going on for like 7 years. There's a whole new executive team in, in place now. But there were actually criminal charges in 2020. 7 employees of the company were sentenced to prison terms ranging between 1 and 5 years.

And so what were they doing? eBay leadership apparently mailed a series of threatening deliveries to the Steiners' home in 2019. That's crazy. Including a box of live cockroaches, a bloody pig's head mask. I don't even know what that is.

Head mask?

Is this a real pig head that's been turned into a mask, or is it hopefully just a, uh, sort of a Hollywood remake of that? Uh, and a funeral wreath. The Steiners also accuse executives of online harassment, including online posts inviting strangers to their house for sex. Absolutely insane. Who, like, in what world is a company, is a public company like eBay just getting some criticism, which all companies do, and just takes it this personally that you're gonna try to ruin the lives of two people that are already in your industry?

Yeah, was the impact of e-commerce bites blogs showing up in your quarterly reports? So like, was it actually—

eBay pleaded guilty in 2023 in a criminal case over harassment and paid a $3 million penalty following criminal charges in 2027. Employees of the companies were sentenced to prison terms ranging from 1 to 5 years.

It's absolutely brutal.

According to court documents, eBay executives grew increasingly upset by the Steiners' critical coverage of the company in early 2019, leading then-executive Devon Wenig to text another executive that it was time to crush the Steiners and take her down, referring to Ina. That summer, harassment against the Steiners escalated from threatening messages to delivering items to their home, which also included a book about surviving spousal bereavement.

It's like sort of a death threat, I guess.

Yeah.

Right.

What's also sending adult magazines to neighbors' homes in the husband's name. So just like completely trying to destroy this couple's life and, and marriage. eBay employees also traveled to Massachusetts to vandalize the Steiners' home and stalk them around their hometown. All the employees involved in the harassment were fired in September 2019, eBay said. The lawsuit was actually started in 2021. Under the settlement— so we have a, we have a good ending here. I think it's a fair ending. Under the settlement, the couple will receive $48 million in compensation from eBay and former employees. eBay will also pay $6 million to charitable organizations, and Wenig will pay $1 million to a free speech charity in Ina Steiner's name.

Ultimate Streisand effect. I never would have heard about eCommerce Bytes or their criticism of eBay if they hadn't done this. What a mess.

eCommerce Bytes.

I'm looking up exactly what they said because eBay is, in my opinion, just the most set-it-and-forget-it company that's just had the same service for 20 years. Well, but now we're going to get to the bottom of it. Big news out of the Fed. The Fed holds rate steady, but 3 officials voted for a rate increase. The rate has been unchanged and in this July meeting. The split shows pressure building inside the central bank to act on inflation 2 months after Chairman Kevin Walsh took over. And the Dow skids 1,100 points as investors worry about Fed next steps. They're thinking that there might be a rate hike in the— in reaction to inflation, of course, somewhat driven by energy prices, oil prices. Oil is spiking again. And so lots of tricky news for the global economy. But fortunately, the economy chugs on and the economy is strong.

So I believe in it.

I believe in it too. Never, never blackpill. Dwarkesh Patel certainly not blackpilling. Gavin Baker quote tweeted Dwarkesh's post that predicted lab revenues in the trillions very shortly. Gavin Baker says, I see a series of fairly wild thoughts from Dwarkesh. I thought I was bullish, but even I am not assuming that the price to rent compute continues to go up. I will say that I am in Silicon Valley this week and Dwarkesh is capturing the zeitgeist. Net-net, public markets would probably be trading differently if they saw the OpenAI, Anthropic, Groq, Cursor, and open source numbers over the last 6 weeks because everything is growing. Even in— there's so much horse race around Oh, open source is going to come and grow. And it's like, well, sometimes open source comes and grows and everything else grows too. And that's what we've seen throughout the whole Deepseek saga when the lab revenues grew a ton, even though open source was big.

Yeah, there's a comment here that I think is on the nose.

Please.

J Capital says timeline is very interesting. All private AI companies have their best 6 weeks while retail AI/semmis have their worst 6 weeks. Very, very interesting disconnect there.

Yeah, I mean, one of them, SK Hynix, $64 billion in quarterly profit. It was a disappointment for chip investors and SK Hynix fell 10% despite record results powered by the AI boom. The profit increase was incredible. I think they, they, they grew profit over 500%, something like that. And still the stock traded down because expectations are so high. Everything's priced to perfection. In other news, HBO Max is launching HBO Max Shorts for those who don't have the attention span to watch an episode of TV. It's a mobile feature designed for users to watch iconic and memorial scenes from the HBO library in a vertical and scrollable feed. This is already happening on Instagram and TikTok. You see a lot of movie clips, just great moments, and it actually is great content. So I can understand why they're trying to own a piece of that.

You've got to get the Jeffrey Katzenberg apology form circulating.

Oh, for sure. Yeah, for sure. But yeah, High Yield Harry captures it pretty well. "Watching society become too cooked even to watch a TV episode." And this is from Idiocracy, Mike Judge's, one of his best works. The creator of HBO's Silicon Valley, actually, who will now be available in short form because you will be able to watch short clips from HBO Silicon Valley on HBO Go. Max Shorts, and sign up for our newsletter at tbpn.com. We'll see you tomorrow.

Love you.

Goodbye. That's the outro. Goodbye.