Hard Fork
Original podcast
“At its core, what this is is like a state-sponsored industrial project.”
12,168 words · 9 speakers
Support for this podcast comes from Dell Technologies. It's time to put AI to work. Is your business ready? The challenge isn't getting started with AI. It's scaling beyond pilots while maintaining control of your data, governance, infrastructure, and costs. Dell AI Factory provides the integrated foundation you need to scale AI securely and efficiently with one trusted partner, Dell, AI on your terms. From idea to impact. Learn more at dell.com/yourwaytoAI. That's dell.com/yourwaytoAI.
What's going on with you?
Well, one of my friends, who is a Zoomer, is bringing back millennial slang, but as a bit.
Oh, boy.
And I don't like it.
Like what?
Well, now, every time we go out to a restaurant and he eats his food, he says yum.com.
And it's so upsetting to me because I remember when people, myself included, would say that unironically.
Yeah.
And now it's come back 20 years later as a bit.
I don't like this either.
I don't either.
Our culture is not a costume.
Suffice to say, this is not awesome sauce.
It is not epic bacon.
It's really not epic bacon at all.
I'm Kevin Bruce, a tech columnist at The New York Times. I'm Casey Noon from Platformer. And this is Hard Fork. This week, the intel on intel. Why the US government is taking a 10% stake in the chipmaker. Then, Waymo co-CEO Tekedra Mawakana joins us to talk about expanding to New York City, making roads safer, and what exactly the company shares with law enforcement. And finally, Kevin, it's Trump's in tech.
Buckle up.
Well, Casey, we're going to start this week by talking about the biggest tech story of the week, which is the partial nationalization of Intel, the big American chip company. This has been a big talker.
Yes, Kevin, like the Pentium computers of yore, this episode is going to have Intel inside.
Dun, dun.
And if you were born after 2000, you didn't get any of that. But you can look it up.
Yes, so Intel CEO Lip Buthan went to the White House a couple of weeks back to meet with President Trump. And at the end of this meeting, there was an agreement, a somewhat unusual agreement, that gave the United States government a 10% stake in Intel in exchange for $8.9 billion of payments to Intel from the federal government.
And Casey, we actually have a clip of President Trump talking about his meeting with Intel CEO Lip Buthan and this deal that came out of it.
Took about, I would say, less than 45 seconds. I said, you know what you should do, if you're smart? Give the United States of America 10% of your company. And he looked at me and he said, I'll do that. I said, I just made $11 billion for the United States.
So that is what President Trump is claiming, that he sort of proposed this deal, that Intel's CEO immediately agreed to it, and that this was already paying off for the US government, because now it has a 10% stake in this very valuable tech company.
Yeah, well, I mean, days before, he had said that Intel's CEO should resign immediately because he was, quote, "highly conflicted," which seemed to be some sort of vague accusation about investments that Tan had made in Chinese companies. So I think this did give the whole affair the air of a shakedown. It's like, well, OK, we'll turn on the pressure on you to quit your job, but you have to give us 10% of your company.
Yes, this is a very different sort of government intervention than we've seen in the past. You know, the Obama administration, for example, took a stake in some American car companies during the 2008 recession. Because it was worried that companies like General Motors would go out of business. And so they kind of step in and backstop those American car companies as a means of kind of keeping them stable. This is not that, right? Intel is not in existential crisis. They are doing a bunch of layoffs. They have been underperforming for quite some time. But there was not really a danger that they were going to go out of business. Instead, this is the Trump administration sort of taking a strategic investment and role
here and possibly sort of using some strong arm tactics to do that. So let's talk about this deal, why it happened, what the government is trying to accomplish by taking a 10% stake in Intel and what some of the reactions are, because this has been a pretty unusual story. So Casey, what's your understanding of why the US government is taking a stake in Intel?
Well, you know, Kevin, I believe it was last December that we did a segment about why it is that Intel was having such a hard time. For decades, Intel was a giant, very successful, very profitable company, making a lot of chips that went into all of the desktop PCs that we were buying every single year. But when that mobile revolution came along, they just kind of missed out and they've been falling behind ever since. But for a variety of geopolitical reasons that I'm sure we are going to get into, the United States government has now said, this is a big problem and we're going to put some money into solving it.
Yeah, so if I were trying to explain to someone who's coming into this totally fresh, why we are now at a point where the US government wants to partially nationalize this big American chip company, I would just start with the following facts. Chips, semiconductors, these sort of advanced AI chips that companies like Nvidia are known for, they are very important strategically, economically. You know, many billions of dollars a year are spent on these chips and on devices that contain these chips, which is basically every device. And chips have designers and fabricators. There are the companies like Nvidia and Intel and AMD that kind of come up with the designs for the chips and package them into these like finished
Processors and then there are the actual what are called fabs, the fabricators that actually manufacture those chips.
And that in itself is a fairly recent development, right, Kevin? Because there was a time when every chip manufacturer would both design and manufacture their own chips. It was the rise of TSMC, this Taiwanese company that I'm sure we're going to talk about, that came up with this idea of, hey, we will just let other people design the chips and we will make them, we'll make that our expertise. And that turned out to be one of the great business ideas of all time.
Totally. So TSMC, as you mentioned, is the Taiwan Semiconductor Manufacturing Company. That's what the initials stand for. And they-
I was just saying, they weren't that creative. It could have had a little more flair, but they were in a hurry.
Yeah, it's like one step above like amalgamated industries.
Literally.
But they have been very, very successful at sort of being the chip fabricator for the world. So many of the biggest chip designers in the world, companies like Nvidia, AMD, even some of these more specialized AI chips made by companies like Google and Amazon. Most of these, almost all of the highest powered AI chips are made by TSMC in Taiwan. And that is an issue for the United States because, among other reasons, Taiwan is next to China. And China has claimed for many years that Taiwan is part of its territory. And so we are in an AI race and trade war with China.
And if China were to, say, invade Taiwan or even to pressure TSMC to stop sending the chips that it makes to the US and its allies, we would be in real trouble.
All of a sudden, we would be unable to sort of fulfill the insane and surging demand for these high-powered AI chips and a lot of the other chips that go into the laptops and phones and other devices that we all use. So- For years now, people in and close to the American government have been sort of sounding the alarm about this dependency on TSMC and Taiwan for our biggest chip needs. And they've been saying, well, we actually need like an American chip fabricator to be making chips that will reduce our reliance on TSMC.
Right. And there have been some efforts in that regard. TSMC is building a big plant in Arizona. There have been some other efforts to do what they call some reshoring of chip manufacturing in the United States. But those are relatively nascent efforts and we have nowhere near the capacity that we would need to replace all of TSMC's operations where they just suddenly go away.
Yes, and that's in part because it's just like really hard. It's not trivial to start up a new fab. It takes like tens of billions of dollars. You need this very specialized equipment and lithography machines. And there's like a years-long waitlist for that kind of thing. So it's not as simple as just saying, we're going to start a US chip company. You actually have to like spend billions of dollars and wait, in some cases, years to make that happen.
Yeah, this was really confusing for me before I got a fuller understanding of this story because I was like, okay, it costs 20 or 30 billion dollars to build a fab. There are a lot of people that would give you 20 or 30 billion dollars if they felt like they could make 100 billion or a trillion dollars, you know, in the next 20 years. But there are a couple of things about these fabs that just make that really hard. One is individual chips are really inexpensive. So you need to have enormous volume if you're going to make a profit on that 20 or 30 billion dollars that you just invested. Another thing, Kevin, is just that because Taiwan has been doing this for so long, in fact, you could say they've been chipping away at this for a really long time.
I see you did that.
There's just this whole knowledge transfer of all the people who have been working in this industry, who are all clustered in this relatively small geographical area. So even though on its surface, it does not seem like it would be that hard to just build a dang factory somewhere, once you drill into it, there are a ton of barriers to entry for somebody who wants to compete with TSMC.
Yes, and Intel obviously understands this whole market dynamic. They understand that there are a lot of American Leaders who would like them to be producing more of their chips right here in the US. And so they started their own chip fabricator. It's called Intel Foundry. They make a bunch of chips, mostly for Intel's own processors. They are not yet making a lot of chips for other companies, the way that TSMC sort of makes chips for a bunch of different companies and slaps their labels on them. But they do have a foundry. And so part of what the Trump administration is hoping to do here is to kind of give that effort some boost, right? To sort of throw the weight and the funding of the United States government behind Intel
and say, you are our chosen national chip maker. We are going to support you. And in hopes that this will sort of give you the momentum and the capital that you need to kind of compete with TSMC and bring chip manufacturing back to America.
Right. If you were a potential customer of Intel's and you were looking at the rough few years that they've had, you might think, you know, I'm not sure I'm going to give my business to these guys. But the hope here seems to be, well, if the United States government is going to stand behind them, you now have more assurance that the company isn't simply going to go bankrupt and leave you high and dry.
Right. So that's sort of the history and general context here. There are some people who think this is a really good idea, some people who think this is a really bad idea. Like, what is your reaction to just this general plan of having the US government under President Trump take a 10% stake in Intel?
So it seems logical to me that the government would be taking an interest in how to secure a lot of chips for us in the future, to diversify away from having Taiwan be our only major source of chips. And so to that extent, all of this makes sense to me. But as I dig into the details here, Kevin, one thing that stands out is my understanding is that under the CHIPS Act that was signed into law by President Biden, Intel was already going to get this $8.9 billion. It's just that now the United States government is asking for something in return. So if that's the case, I'm wondering, how big of a game changer is this really? Because if Intel was already going to get this money, it's
Not clear why this is going to sort of vault them ahead of where they would have otherwise been, except that now they owe something to Donald Trump.
Yeah, it's a little confusing, but basically this was money, this 8.9 billion dollars that was, had already been awarded to Intel through the CHIPS Act, but it had been conditioned on meeting like certain benchmarks for Intel to actually get that money. Now it is not. Intel just gets that money automatically. So they are getting something here out of this deal, which is essentially this sort of guaranteed influx of cash from the government. But I think the bigger draw for them, the bigger reason for them to want to do this deal is now the US government is officially invested in your success, right? You have kind of a shareholder in the United States government. The Trump administration is going to want to do everything it can to look good, to have this deal sort of seem successful.
And so they're going to want to sort of prop you up and buy a bunch of chips from you and make a bunch of deals and make sure that this company succeeds.
Okay, so what is Intel gonna do with 8.9 billion dollars that is going to save them from all of these problems that they're having?
Yeah, I mean, that to me is the case against this deal. I understand the strategic rationale for wanting to reduce our reliance on Taiwanese chip manufacturing. But to me, like, Intel's problem is not access to cash, right? It is a big company, it has investors. And so I think if it wanted to raise a bunch of money to, say, build out some new fabrication capabilities, it had plenty of ways of doing that. Its problem for years now has been a lack of execution, right? It missed this shift to mobile, it missed this shift to AI. It has been sort of consistently behind its rivals when it comes to seeing what is coming around the corner and what that requires of them in their chip business.
And so I think this infusion of cash from the government doesn't really solve that. In fact, it might make it worse. I don't think that government-run or partially government-owned enterprises are typically thought of as being like very innovative, or at least more so than private market companies. And so I would have preferred that the government not prop up Intel by sort of taking a 10% stake. I would have preferred that they kind of incentivize or fund some smaller companies trying to make chips in the US and sort of build a competitive market here rather than kind of picking a winner and putting their stake behind that winner.
What is Intel's plan? We've talked earlier about how there's sort of two ways of doing chips. You can either design and manufacture all your own chips, or you can make chips for other people, or you can do, I guess, some sort of hybrid. What is it that Intel plans to do? What's its battle plan here?
So Intel has a lot of plans. You know, they want this foundry business to be very big and competitive with TSMC. They want to make a bunch of chips for a bunch of other companies other than Intel. They want to compete with companies like Nvidia and AMD on these very advanced AI chips. Like, they've got a lot of plans, but it's not clear to investors or the general public, or me frankly, that they are sort of up to the task.
You know what it kind of reminds me of is like, you know, like at the start of the year, I would be like, this is the year I'm going to cut back on sugar, you know? I'm eating too much ice cream, you know? It's time to start eating more fruits and vegetables.
Do we think Intel has a better developed plan than that?
I think slightly, but it's also like, it's not the easiest plan to enact, even if you do have a good plan, right? Like, this is not just about spinning up a couple factories to start making chips. You also have to convince all the other chip companies that are designing these chips that they want to start manufacturing with you instead of sending their designs to Taiwan, where they've been manufacturing them quite successfully for many years.
Right. It's not exactly clear what the appeal is supposed to be, aside from maybe now currying favor with the Trump administration.
Well, that's a, no, you mentioned that, and that's an interesting point here, which is that now, you know, companies like Nvidia, AMD, other chip companies, they may want to start at least moving some of their chip making to Intel because the Trump administration is now sort of officially cheerleading for this company and is going to want to use its authority and strong-arm tactics to get more business for Intel, maybe by doing some of these deals.
Yeah, it's very easy for me to imagine the Trump administration saying, hey, you know, buy some more of these Intel chips and we'll give you a break on your tariffs, you know, from importing things from China, or, you know, whatever.
Totally. But Casey, this is also like something that you have made a prediction about in the past that I want to check in with you about, which is that you have said that you think that eventually one or more of the big AI companies will become nationalized, that the government will see this as being so crucial to their strategic long-term interest that they will just decide to like take over an OpenAI or an XAI or some other AI startup.
Yeah, and this isn't just my idea. I know that this is something that gets discussed a lot inside the big AI labs. And the reason is because, of course, over there, they're all true believers. They believe that soon AI will become very powerful. It will have a lot of military applications, and it won't be too long before the United States, you know, intelligence apparatus says this is critical to our national security, and we just need to have much stronger oversight over all of this. So, yeah, I think if you sort of, you know, play it out all the way, and you can imagine, you know, AGI or some kind of superintelligence, it just seems obvious to me that the national government would want to own the thing that creates that and not just, you know, let anybody do whatever they want with it. What I did not see coming was that we would start by
nationalizing companies that are not playing that same role in our national security. Although you could argue that, you know, Intel maybe does have some significant role in our national security. What I wonder, though, Kevin, is now that we're sort of establishing this precedent, that the president can just come in and say, hey, I want 10% of your company, what other companies is he going to do that to? He is already talking about making more, quote, "deals." And I can imagine him just kind of wanting to do this because I think this is kind of how he sees the world. Hey, you want something from me? I gotta get something back. And then he can go in front of the cameras and talk about all the deals that he's making. And I think it makes him feel great. So, yes, I do not think that this is going to be the last of these deals that we see. And while I don't think OpenAI is going
to be the next company to have to give him 10% of their operation, I can't imagine it's too far down the list.
Yeah, there's been some really interesting reactions to this news and I just want to read a couple of them out because I think this is sort of scrambling some of the like conservative and liberal sort of predispositions here because at its core, what this is, is like a state-sponsored industrial project.
This is a consolidation of state power over industry. And so a lot of free market conservatives are very offended by this. They feel like this is a betrayal of their principles. So, for example, Michael Strain, an economist with the conservative-leaning American Enterprise Institute, says, "I think what we're seeing is less a strategic thoughtful shift towards state capitalism and more an opportunistic display of corporate shakedowns." Senator Rand Paul called this Intel stake a terrible idea and said, quote, "If socialism is government-owning means of production, wouldn't the government-owning part of Intel be a step towards socialism?" Senator Bernie Sanders-
I could actually answer that question, which is, no, it actually just is socialism. Go ahead.
Yes, which might explain the next reaction, which was from Senator Bernie Sanders, who supported the deal and said, quote, "If microchip companies make a profit from the generous grants they receive from the federal government, the taxpayers of America have a right to a reasonable return on that investment." So, the rare Trump administration action that is supported by Bernie Sanders and opposed by Rand Paul.
Hmm, crazy times we're living in.
Casey, if you were the President of the United States, what companies would you want to stake in?
First of all, I'd resign.
I think you'd be forced to resign before you resigned, but anyway.
Look, if I were going to try to solve the United States dependence on Taiwanese chips, I would convene a panel of experts, I would work with Congress, I would come up with a plan, I would try to get legislation passed that created a pool of money, and I would just sort of proceed in a more structured way. I wouldn't just be sort of doing these quick one-off deals that seem like they only have a tangential connection to any sort of plan.
What would you do?
I would set up a national- Reality style competition for fledgling chip manufacturing startups.
Chip Island?
I was actually gonna go with American Chip Idol. And you could televise this every week. You could have them kind of compete for billions of dollars in government funding. And you could sort of have a Shark Tank style judging panel where they would kind of inspect the wafers coming off the assembly lines and say like, oh, the lithography on this is terrible. I never want to see you again. Get out of here. And then at the end of it, one American chip manufacturing company would get the billions of dollars and 10% stake in their company.
Wow. Well.
What do you think?
I'm, um, I can't say I love it. I can't say I love it.
Okay, well, then I'll have to go with my other idea, which is that the United States should nationalize whatever company makes the Labubu dolls. They seem very popular.
When we come back, we ask the CEO of Waymo. What are you driving at?
AWS AI is how the world's leading organizations are transforming their industries, not in theory, but in production at scale right now. Like turning race day data into real time insights for Formula One fans, reinventing the journey for every United traveler, and personalizing every trip with booking.com. However big the question for your business or industry, AWS AI is how.
It's time to innovate at the speed of AI without compromising control. That's OutSystems. Take a look at what happened with one global retailer during a period of massive expansion. They used OutSystems to build AI agents to automate invoicing and translation processes, shifting 95% of their daily store operations to a new digital app, saving thousands of hours of manual work. OutSystems gives you the flexibility to deploy how you want and the governance to do it safely. Own your agentic future. Learn more at outsystems.com/nyt. That's outsystems.com/nyt.
Well, Casey, today we have a very exciting interview on the show, a conversation I've been eager to have for quite a long time.
Yeah, this is one that we have been trying to make happen for months, and I'm so glad that we finally get to talk to somebody who runs a company that you and I are both taking advantage of quite a lot, I would say.
Yes, today we are having on the show Takedra Mawakana. She is the co-CEO of Waymo. That is, of course, the former Google self-driving car project that was spun out as an independent company and has been steadily scaling and pushing into more and more cities. Over the last year, Waymo has gone from operating just in San Francisco, Phoenix, and LA to having a small fleet in Austin and Atlanta. And just this past week, we learned that Waymo is beginning to test its cars on the streets of New York City.
That's right, Kevin, but at the same time, as Waymo has expanded, there have been more and more questions about what it means that there are more autonomous vehicles on the road. Will they be taking jobs away from human drivers? And in June, five Waymos were destroyed during ICE protests, and some questions were raised about what sorts of camera footage Waymo was sharing with law enforcement.
Yes, I think this is one of the most fascinating companies in all of technology right now, and I think the problems that they face are genuinely new and interesting. And so I'm excited to talk to Tekedra about it.
Let's bring her in.
Tekedra, welcome to Hard Fork.
Thank you. Thanks so much for having me.
We're so excited to talk to you and we have a bunch of questions for you. But I want to start with a question I saw online recently, which is, why did they call it Waymo and not Google Drive?
Oh, um, so it was Google Self Driving Car Project before it spun out from Google. And I wasn't here when the decision was made to name the company Waymo, but I was really surprised when I showed up for my interview that such a decision had been made because obviously it spent like all of these years being the Google Self Driving Car Project. And then we had to make the world figure out what in the heck did Waymo mean?
Another huge win for the Google marketing department.
So Casey and I live here in the Bay Area. We're used to seeing Waymos all around. I don't even really notice them anymore. But this is not the experience that most people in America are having. Give us a sense of where Waymo is in its expansion. Which cities are you in? Where are you going next?
Yeah, so right now, we're right here in San Francisco, also Los Angeles, as well as Phoenix, where it all started. We're also in Atlanta, as well as Austin. And we've already announced that we'll be in Miami later this year and in Washington DC next year. Beyond that, you know, we're really excited. You're going to start to see our cars in a lot of cities. If you think about our business as far as scale, we're currently giving hundreds of thousands of rides every week. And in all likelihood, by the end of next year, we will be offering around a million rides per week. So if you just think about the scale of where we are now to then, and we're just really excited to start bringing this technology to more people.
So you recently announced you're doing around 250,000 paid rides a week. So you're saying like in a year or so, you think that that's going to quadruple?
Yes.
So I'm so interested in what's driving that. I was somebody who was going down to the Google Plex like 10 years ago to take rides in some of like the early prototype cars, wondering when they would finally, you know, reach the sort of scale even that they're at now. It took a while.
And I remember the engineers down there telling me things like it's really hard for these things to drive in the rain. It's really hard for them to drive in the snow. Every day they're seeing all sorts of interactions on the streets that they've never seen before. So what have you all managed to do that has kind of helped to figure out those problems and let you launch in more places?
Yeah, I mean, I think one, first and foremost, I think it's pretty well known now that we're like a super safety focused company and culture. And so we put out years ago our safety framework. And so everything has always been guided by that. But on the other hand, we have all of these really, really brilliant engineers. And in the early days of Waymo, there was a decision made to put ML into the original tech stack. And so what does that mean? The machine has learned and over time, humans have been identifying edge cases and feeding those back into the machine. And we've been addressing those over time. And so I can remember, you know, a few years ago, when it would rain hard in San Francisco, not something that would happen often, you know, we would pause service just to be sure it could be safe. And we don't do that anymore. We don't even think
about it anymore. And so it's exciting to think about same thing. Once you see us in the Northeast, you'll start to see us tackling snow. You know, that's not something that's currently in any of our cities. And then freeways is something that we've been super focused on. You know, we have a lot of employees posting their videos online. Employees have been taking rides in San Francisco as well as Phoenix and LA. But you're going to see freeways.
You mentioned the safety data. I've spent a lot of time digging through the various reports that you all have put out in some of the third party studies. But most people, I think, have concerns, especially if they've never ridden in a Waymo before about safety. It feels weird to get into a car with no human driver in it. Give us a sense of what your current up to date safety research says about how safe these cars are versus, say, a human driven car?
Yeah, I think our last report was from 71 million fully autonomous miles. And our data shows that we're five times safer than a human driver. And and this is in incidents that would normally involve either injuries or airbag deployments, which is sort of the way the data is split. But that doesn't even take into account who's at fault. And I think normally when we think about accidents, it's not just did it happen? But it's also who caused it. And if you use insurance data as a proxy for who's at fault. And one of the largest insurance companies in the world, Swiss Re, did this work on 25 million miles of autonomous data.
And they actually found that we were 10 times better than humans. And so this kind of like, whether it's five times or 10 times better than humans, I think, you know, we know that we're on this trajectory to make every community in which we drive safer.
I'll tell you, like, the way that I feel this is as somebody who lives in San Francisco, a city where a lot of people walk, I'm often a pedestrian. I'm often just walking down the street and there's Waymos cruising all over my neighborhood. And there are a couple of crosswalks I walk, you know, across every day. And when I see a human driven car, I always have that moment of being a little bit scared of is this car going to slow down for me? When I see a Waymo, I'm like leaping into the cross. Like, I have no doubt, you know? Because I know you don't want to pay me what it would cost if you accidentally hit me. So that's how.
That strange man you've seen leaping in front of cars on your security footage in the Castro? That's Casey. No, but this is an interesting question because Malcolm Gladwell recently had a very sort of interesting and funny take on this where he said that the reason that Waymo and self-driving cars in general will never work is because they're effectively too safe. That people are just going to start jumping out in front of these things. Pedestrians won't sort of give them the right of way because they know that the minute that the the car sort of senses that they're in front of them, they're going to slow down and stop. So he said, basically, there's sort of a sort of deterrence effect that is lacking here because these cars are so safe that it's just going to end up clogging up streets and people will be jaywalking all over the place. What do you make of that argument?
Yeah, you know, I think it's important to like ground this in the why, right? Like, it's so depressing and no one likes to think about it, but 40,000 people killed on the roadways, you know, whether people are.
Every year.
In America.
In the US.
And so I think if you ground it in the status quo. And you say to yourselves like, okay, let's imagine a world where, yes, some number of humans are going to create congestion by believing that these cars are part of a gamification opportunity. But for a lot of people, it's just going to make everything safer. And so that's really what we stay focused on. It's like, is the status quo acceptable? No. Can this technology improve road safety? Yes. Can we figure out also how to like delight people who otherwise have other options? Yes. And so that's really what we're focused on. And I think the riders continue to give us the feedback that like,
this is meaningfully improving their lives. You know, we launched the teen product in Phoenix this year. And it's just, you know, there's so much demand for working parents who just want a safer option for their kids. And so, you know, I think more people are going to benefit than not.
Now, tell us about this teen option, because I don't have a teenager, but my understanding is that a lot of parents want to send their kids to school or soccer practice or whatever, and they want to use a Waymo so they don't have to like get in the car and drive them all the way there. So that's allowed. You're allowed to like send your kid unaccompanied in a, in a Waymo?
In Phoenix, yes.
Got it.
Yeah, so we rolled that out there.
I was a teenager once. I had a car once. I did a lot of things in that car. Some of them maybe more legal than others.
Well, we don't have to get into it here. But what are, like, is, are there different safety precautions with teenage riders? Do the cars drive safer? Are there restrictions?
They do donuts to impress the teens.
No, like, are teens doing teenage stuff in the Waymos?
Um, no, that everyone who rides in a Waymo complies with our policies.
Okay, no, that's not true.
No, but I mean, seriously, like, getting people safely from point A to point B is what we're focused on. And the Waymo driver doesn't behave any differently with teenagers. You know, we still want to make sure you have on your seat belts, and that you keep them fastened, and you get from where you're going to where you're going. So.
You know, I, I think one reason why the, why Waymo is very popular in the Bay Area is that it feels like a very
Fancy experience, you know, you're in these like souped up Jaguars with hundreds of thousands of dollars of sensors on them. And so when you're like, you know, going to McDonald's, you feel like a king, you know, in your Waymo. I'm wondering if you guys can keep that up, you know, because at some point you're going to want to turn a profit on these things. And I'm imagining you would love to make a much cheaper version of this car. So how are you thinking about the trade offs that you make between getting to profitability and having this kind of premium experience?
Yeah, I mean, think about us as in these early days, you're right, we used to have the Chrysler Pacifica minivans when we first launched in Arizona, then we moved into these JLRs. We then have the Zeeker vehicles coming off the line. We have IONIQs that'll follow. And then think about in the future state, we're going to have tiers of products available, right? Like there's going to be people who want the date night experience. That's going to be different than the people who have the soccer team experience. And so we're thinking about these as like product offerings over time. And that's how you'll see us scale, not only in the US, but globally. There are going to be cities across Europe where we need a small vehicle form factor. So we're really thinking
about not only the experience, once you get in the car, but what is that use case that people are most likely to adopt?
I want to ask about two bottlenecks or potential bottlenecks that I see coming for Waymo that are already starting to play out. One is the supply chain. Waymo has been interested in sourcing vehicles from a Chinese company, Zeeker, for many years now. I believe you already have some of those up and running. But the Trump administration has not been a fan of bringing in Chinese electric vehicles. There are these new tariffs on top of some older tariffs that now make the tariff rate on a Chinese EV imported into America, something like 250%. How is Waymo dealing with that with its Zeeker deal? Are you just swallowing the tariffs and the additional cost of this? Have you found some waiver or exemption or way around that?
What is the current state of your Chinese supply chain?
So we are, like everyone else, staying abreast of all of the developments that are happening with the administration. We have a number of the cars in the US already, and we've been testing them, and our plans haven't changed. We still plan to scale with those vehicles. At the same time, like I mentioned earlier, we have a whole host of OEM partners. So we are also focused on how to make sure we're scaling with them in order to have the right amount of diversification on this, on this specific issue. But for all of our cars, we actually apply the Waymo driver at our facility, which is run by Magna in Arizona, Mesa, Arizona. And so all of our technology happens here in the US. And that's really
important, right? I mean, at the end of the day, what, you know, we think about a global supply chain is what pretty much every company has to deal with. But where this technology was born, where it's applied, where it's understood, is right here in the US.
The other thing I wanted to talk about is some of the opposition that Waymo's been getting from unions in some of the cities where you all want to operate, including New York City, where the TWU International President, John Samuelson, recently said, "New Yorkers be warned. Waymo will turn pedestrians into cannon fodder and will block streets for emergency responders. Waymo isn't ready for NYC's streets, and the end goal is to replace rideshare drivers, taxi drivers, and transit workers with robots." Obviously, this is not something unexpected. You all have faced opposition before from people who say you are coming into our cities and trying to replace all of our drivers with
robots. But how do you respond to that complaint? And more generally, how do you think about the political opposition that you are facing?
I think, you know, I would point to next month, actually, October, will mark five years since we removed the drivers from every Waymo vehicle in Arizona. And there continue to be a host of people with driver jobs, and we didn't take those jobs away. I think at the same time, we're creating jobs, right? Like, as I mentioned, we have, you know, people who are manufacturing and integrating the Waymo driver onto automotive vehicles. We have fleet techs who are maintaining the fleets. You know, we have depot leaders as well as depot workers.
And so there are jobs that are being created and there are jobs that will go away. I'm not, I totally agree with that point. It's just not the case that they're all going to go away because of autonomous vehicles. And I think, just like I can agree that some jobs will change, I think, you know, hopefully they can agree that the road should be safer.
Let's talk about another kind of opposition that we've seen to Waymo. In LA in June, you lost five Waymos during ICE protests. And it seemed that there was a feeling that Waymos either were or could become a part of the surveillance state. So what can you tell us about how Waymo interacts with law enforcement and whether these protesters are right to be concerned that if a Waymo drives by them, that footage could go straight to ICE?
I think we didn't view what happened in Los Angeles as in any way targeted to Waymo specifically. I mean, five Waymos were harmed, but, you know, so was a lot, a lot, a lot of other property, you know, parking structures and other people's cars, personally owned cars, et cetera. And so we didn't think that that's related. I think on the question of, you know, when and how law enforcement gets access to our data, you know, we make that publicly known. We follow the sort of legal process to receive. Footage from our vehicles. And, you know, we narrow the scope of that as needed because like at the end of the day,
we need communities to be able to trust us. Not just trust us for the safety of our drive, but it's also, you know, trust us that you understand what it is that we're doing with what we've captured.
So you're saying that you are pushing back on law enforcement requests that you think are too broad?
Overly broad, of course, of course. And, you know, not only is it burdensome, but also that's just our process.
Hmm.
I'm curious how you see the political dynamics playing out over the next few years. I mean, the last time cities in America were really faced with like a whole scale change to transportation was ride sharing, when companies like Uber and Lyft would come in and, and, you know, in some cases, kind of surreptitiously sort of just insert themselves into a city's transportation system. And they would sort of try to get really big so that by the time city council or the mayor or whoever sort of caught on, they already had this mass of customers who they could then mobilize to like defend them. And they sort of pioneered this model of like blitz scaling or just going really fast and, and hoping that you can sort of work out the regulatory issues later.
Waymo is obviously taking a different approach. You can't really do that the same way. The cars, you know, identify themselves visually, so you can't just like sneak in the way that Uber did in some cities. You have to get permission and you have to go through the procedural hoops and you have to get city council on board. So I'm just curious whether you, you sort of think that's the path forward that makes the most sense, or whether you wish you could kind of do it the way Uber did.
I think I'm kind of on neither extreme. You know, one of the reasons we're able to launch in Arizona is because Governor Ducey issued an executive order. I think the bar of what you have to have already demonstrated you can do should be very high in that case. I don't think it should be your first market. I don't think it should be your second market. I think you should have a lot of data. But I do think if you're willing to be transparent with the public and enter that market, you know, I don't think that the bar needs to be a complicated regulatory framework. So this is sort of the debate that's happening is like, how much regulation do you need in order to get to that sweet spot? I think it's going to depend.
I think there are some places that aren't going to take a heavy hand there. I don't want safety to be sort of a race to the bottom, though. I don't think that doesn't help any of us if a lot of bad things start happening with other autonomous vehicle companies. And so this is where we're trying to strike the right balance. But like move fast break things, absolutely not, you know, in any way, the way we think about it. It's more like be super deliberate and deep engagement. But where we could have a fast movement on the policy side, that would be great.
Let's talk about Tesla. They have rolled out their own robo taxi service.
Speaking of race to the bottom on safety.
They are now operating autonomous taxi services in Austin, one of the cities that Waymo is also in. They are clearly trying to compete with Waymo. They're taking a different approach, using only cameras on their vehicles instead of these more expensive LiDAR sensors. Do you see Tesla and what they're doing as direct competition? Are you worried about it? Are you making any changes to your plans based on what they're doing?
You know, we have sort of always thought there would be many, many players. I mean, remember, it was a crowded field for a long time.
And then crews got into one accident and the field got less crowded. But anyways.
A little more than that, but.
It was one accident.
I'm just saying there were more factors than that.
Well, yes, but anyway, I'm sorry for interrupting.
I thought you were saying there were more companies than that.
Oh, yeah.
So, you know, more companies trying to make the road safer, that's a positive thing. And we welcome that. You know, our approach is a full sensor suite. We think that's really important. It's been really important to get us here. And there isn't another company that's been able to remove the driver and scale to where we are, do freeways and surface streets without a human in the in the car to take over or without an extremely limited ODD.
Speaking of the sort of remote assistance, sometimes I see skeptics online who believe that there is a remote operator driving a lot of the time in these Waymo cars. Can you speak to that at all? Like, on my median ride going two or three miles in San Francisco, is the remote operator getting involved? Or is that happening in much more limited cases?
Yeah, so we have, there's no ability for anyone other than the vehicle to drive. We do not do teleoperations. We do not do remote driving. It is all onboard software system. So that is, to us, that's the definition of autonomous. Like, it is operating fully autonomously. What will happen is, you know, there's two sort of ways that data can be shared in your experience. One is rider support. You know, you push the button, someone talks to you. That's a group. They have nothing to do with the driver function. The other is, you know, kind of to the conversation we were having before, you know, there's a flooded road. That's a place where the car might stop. And then it'll ping, this is all like technology,
it'll ping saying, "The road went away." And then you can have a remote assistance dispatcher send a queue, again, coding. And then the car has to decide the safe route to take. So it's just, these are the edge cases, or the long tail of what you encounter on the road. And so they become more and more and more edge case-like, and less and less frequent over time.
That's it. I actually did not know that. I thought that there were still like remote teleoperators. So if I like encountered some blockage or something, somebody grabbed the wheel and drove it home from the Philippines.
We think like gamification of driving. You just don't, you just don't have enough context to do it safely. And so we feel very strongly that this needs to be fully autonomous, which is why sometimes the car pulls over, because it's trying to resolve the issue, but it has to resolve the issue itself.
Well, I'm a little disappointed, because I was going to ask you if Kevin and I could spend a day remote driving the Waymos.
No, you cannot.
I thought it'd be a fun team bonding experience.
You're stepping on my other pitch, which is that every Waymo in San Francisco should automatically start playing the Hard Fork podcast whenever you get in. Just an idea. Speaking of product feedback.
Oh, here we go.
To Kidra. The other day, I was trying to get a Waymo to go to dinner with my good friend Casey Newton. And I opened up my Waymo app and I pressed, you know, hail car. And it told me that it would be 52 minutes until I could get a Waymo and that it was also going to cost about twice as much as an Uber. So.
That happens to people with low rider scores. I think you're only about a 3.5 in the app.
So two separate questions here. What is happening with wait times in San Francisco? And why is it more expensive to take a Waymo than an Uber if Waymo is not paying a driver?
So first, on your experience, thank you for the feedback. We saw your tweet. And sorry that you had that experience.
Heads rolled, I assume?
Heads always rolling.
Please. Yes.
No, we take, we take rider feedback super seriously. There was a bug. That ETA was caused by a bug. So it was actually really helpful to know because you do not see wait times like that on the Waymo app. Period. Full stop. It's not, we have enough cars. It's not a, we're not supply constrained. Every car was not in use at that moment. Why is it more expensive? It's algorithmically determined, but like this is a consistent fleet experience. Like we, you know, really want this to be a premium experience for riders. And like I said earlier, over time, you'll see us kind of have these tiers of products. But right now, it's, as you guys said, it's a nice JLR, iPace.
They're clean, they're consistent. So you will see it competing with riding. It's not competing with Uber X or, you know, sort of the lowest tier of Lyft.
Yeah. If you want a low tier experience, Kevin, maybe try Uber X. Okay?
Yeah. No, I, I feel this way. I actually, I do, I am willing to pay more for a Waymo than.
Wait too long.
If I don't have to wait too long, because I do, I do prefer the experience and also it feels like you're in the future. Still, I still have that little moment. But I'm not willing to pay three times as much.
Yeah. Yeah, and I think that's fair and fine. And obviously, we have a fixed fleet. And so you're gonna see dynamic pricing happening based on our demand. But I am happy to hear you say, you know, you think of the experience as sort of value-based, which is how we think about it.
Here's the best feature to me about the Waymo is that it has never once asked me about my politics at 4:30 a.m. on my way to the airport. I don't know how you guys came up with that one, but it's incredible.
Maybe you should have a mode where it does ask you about your politics, just for people who feel nostalgic.
Yes. You just press a button. Instead of music, you can just have somebody like monologue about what's going on in Congress.
Thank you so much, Akira.
Yeah, thank you.
Great to talk with you.
This was super fun.
Thanks for tolerating our feedback.
Yeah, of course. Thank you.
When we come back, the first lady of technology has a bold new plan for AI in schools, and we barely understand it.
Support for this podcast comes from Dell Technologies. It's time to put AI to work. Is your business ready? The challenge isn't getting started with AI. It's scaling beyond pilots while maintaining control of your data, governance, infrastructure, and costs. Dell AI Factory provides the integrated foundation you need to scale AI securely and efficiently with one trusted partner, Dell. AI on your terms. From idea to impact. Learn more at dell.com/yourwaytoAI. That's dell.com/yourwaytoAI.
It's time to innovate at the speed of AI without compromising control. That's OutSystems. Take a look at what happened with one global retailer during a period of massive expansion. They used OutSystems to build AI agents to automate invoicing and translation processes, shifting 95% of their daily store operations to a new digital app, saving thousands of hours of manual work. OutSystems gives you the flexibility to deploy how you want and the governance to do it safely. Own your agentic future. Learn more at outsystems.com/NYT. That's outsystems.com/NYT.
AI spend is growing fast, but most businesses still can't see exactly where that money is going. Ramp gives businesses visibility and control across AI providers with real-time guardrails by team and model, helping ensure spend goes to the highest ROI areas. Stop guessing where your AI spend is going. Get started with Ramp. And if you're new to Ramp, you can get $250 off at ramp.com/hardfork. Go to ramp.com/hardfork. Ramp.com/hardfork.
Well, Casey, it's time for a brand-new segment that we're calling Trump's in Tech.
Everything's computer.
That's right, Kevin. We couldn't help but notice over the past few weeks that it seems like with each passing day, there's another Trump taking another prominent position in the world of technology. And we thought, well, we ought to break that down for our listeners.
So first up, Casey, we have a video that was released this week by the first lady, Melania Trump, and I think we should watch this video together.
Yes, I have not watched this because I wanted to, the first time to be with you.
Oh.
Are you ready for an AI challenge? Take part in this nationwide initiative to discover, develop, and expand AI's potential. As someone who created an AI-powered audiobook and championed online safety through the Take It Down Act, I've seen firsthand the promise of this powerful technology.
Is this a deep fake?
Now, I pass the torch of innovation to you. Just as America once led the world into the skies, we are poised to lead again, this time in the age of AI. The president's National Artificial Intelligence Challenge invites every student in America, from kindergarten to tough grade, to unleash their imagination and showcase the spirit of American innovation. Our educators will guide and empower you through this process to build a brighter, stronger future for us all.
Visit ai.gov to register today. Good luck.
Wow, a full minute, and I still have no idea what she's talking about.
Basically, my understanding from doing a little reading on this is that this new presidential AI challenge is going to be a contest for American K through 12 students where they will have to, you know, complete a project that involves AI. Maybe they're vibe coding something for their community, or they're building some tool. And winners of this competition will apparently receive cash prizes of up to $10,000 per team member. And they are also giving out participation trophies.
And cloud credits, Kevin. Don't you remember when you were eight and how much you wanted cloud credits? Well, now you can get them. And all you have to do is create an enormous amount of slop while you could be learning math.
Okay, so you are not a fan of the presidential AI challenge, I take it?
Well, look, I think these kids are going to find their way to AI just fine for the simple reason that it does their homework on their behalf. So I think in order to sort of advance the ball further, you've got to give them a pretty specific project to, you know, get their attention. And based on what the first lady just said, I truly have no idea what these children are supposed to do.
There will be a challenge, and they will be given prizes at the end of it for doing something, something AI, America. Did you know that Melania Trump, our first lady, was so into AI?
Well, apparently she got into it because she used a, an 11 labs clone of her voice to voice her audiobook. I don't know if you've had a chance to read her book yet.
Well, I listened to the whole thing. That's not true. I didn't.
Okay. Well, I haven't read it either. But apparently, once she saw how much you could slack off by using AI, she said, we have to get this to the children. And I think that's really interesting.
It's very inspiring. All right. Next up, we have another member of the Trump family making moves in the tech industry. The prediction market company, Polymarket, announced that Donald Trump Jr. is joining its advisory board. And 1789 Capital, the firm where Trump Jr. is a partner, also made an additional investment in Polymarket that Bloomberg reported was in the tens of millions of. Dollars. Casey, what do you make of this?
So there's a couple of interesting things here. One is that Donald Trump Jr. is also advisor to another prediction market company, which is called Kalshi. So imagine being a strategic advisor to Coke and then joining the board of Pepsi and you would have some kind of analog to what's going on here. The other thing that is interesting is that this is just a transparent bid to get Polymarket into the United States. So we've talked about Polymarket before on the show. You're not allowed to bet right now in the United States using real money on the sort of outcomes that Polymarket is offering up for gambling. But Polymarket wants to change that. They recently bought a derivatives exchange, which is licensed in the United States, which a lot of people see as a kind of effort to
open up here in the US. Just last month, the federal government wrapped up an investigation into Polymarket because they were worried that Americans were finding some loopholes that were allowing them to make these bets. And the Trump administration has literally just made that all go away. So what better time for Donald Trump Jr. to step in and help kick the doors open for Polymarket in the United States and make a tidy profit for Donald Trump Jr.?
Yeah, I'm not a prediction markets gambler. I've never actually wagered real money on one of them, although as we've talked about on the show, I have played around with another prediction market manifold, which uses sort of a play money. But if I were a prediction market guy, one prediction I would feel very confident buying at any price is that prediction markets will be legalized in America during the Trump administration. I think that is clearly the direction that things are heading, and I have mixed feelings about that.
Well, look, all I can think about is how many stories I had to read about what Hunter Biden was doing with his contract with Ukraine. And there hasn't seemed to have been nearly as much conversation about the fact that the president's son appears to be involved in a way to just make Polymarket hugely profitable here in the United States.
Yeah, I mean, the interesting question to me is like whether the sort of politicization of prediction markets will make them less useful. Because one reason that you might use a prediction market is to try to sort of accurately figure out what's going to happen in the world. You know, if you go on Polymarket today, you'll see things about Ukraine, global elections, will there be a Fed decision in September? Who's going to win the Bolivian presidential election? Things like that. And if all of a sudden this sort of becomes coded as like a right wing activity, maybe these markets actually just become less useful because you don't have people with differing views sort of wagering on those.
No, if you want to get an accurate sense of who is going to win the Bolivian election, you don't want Donald Trump Jr. involved. I don't know why, but that's my gut feeling.
Okay, next up, we have a story about Trump Media, the president's sort of family media conglomerate, which is launching a new company in partnership with Crypto.com. Now, we have talked many times about the Trump family's various crypto investments and entanglements on this show, but CoinDesk reported this week on a new deal involving Trump Media buying about $105 million worth of something called CRO or Kronos. That is a token issued by Crypto.com. And as part of this deal, they are going to start a new combined company,
Trump Media Group CRO Strategy, Inc., which will hold this stash of CRO tokens and will trade publicly under the ticker sign MCGA, which stands for Make CRO Great. Again, Casey, please explain this deal to me as if I'm a small child.
I feel like I had a stroke while I was listening to you describe that. We, listen, we've talked about quantum physics on this show and I had an easier time understanding it than I did trying to tear apart what is happening between Trump Media and Crypto.com. At a high level, I think the thing to know is that Trump Media's fortunes are now more closely intertwined with Crypto.com. I know Crypto.com as the company that renamed what used to be the Staples Center in Los Angeles, and it did that much maligned ad where they got Hollywood celebrities like Matt Damon to talk up the virtues of crypto shortly before crypto massively crashed.
But crypto is back now and they have a deal with the Trump administration. What does it mean? Look, to really understand what this means, you would have to buy a CRO, and I absolutely do not recommend doing that.
Yes, we can't give investment advice on this show, but I think it's safe to say that is a risky one. As part of this deal, I'm interested in how they are planning to integrate this new token into Truth Social. One of the announcements that came out of this deal was that this CRO token would be used as what's known as a utility token on the president's social network. Basically, there are these things on Truth Social called gems, which I guess are like some kind of upvote or something. I'm not a big truther, so I'm not sure exactly what gems are. But anyway, they'll be able at some point in the future to convert those into units of this CRO cryptocurrency so you could get paid for your unhinged posts.
Well, you know, several years ago, I got a lot of engagement on Truth Social for posting that Joe Biden had won the 2020 presidential election. They did not want to hear that over there. And now I'm wondering if I could turn that into some sweet, sweet CRO.
Yeah. Only one way to find out. I mean, yeah, I mean, to me, this is just of a piece with the other crypto investments that we've seen the Trump family make. Eric Trump, in particular, is sort of all in on crypto, and they have found a lot of ways, some sketchier seeming than others, to profit off of their investments in these companies.
And what's the least sketchy way that they made a profit from their investments, would you say?
I mean, none of it is how presidents have traditionally conducted business in the past.
No, George Washington did not own any crypto at all.
That's true. He was famously a no-coiner. But I think we'll just have to see where this goes and if anyone's actually going to use these Truth Social utility tokens. Finally, Casey, in our sweeping overview of the Trump family's entanglements in tech, let us end with the president himself.
What's he up to, Kevin?
Well, President Donald Trump announced that the United States was getting its first ever chief design officer. Airbnb co-founder Joe Gebbia has been named the CDO of the United States and has been given the mandate of remaking federal government websites. Gebbia posted to X that he plans to make government websites, quote, "as satisfying to use as the Apple Store." Casey, what did you make of this?
Well, it seems like all podcasters, the government is coming to the same conclusion, Kevin, which is that we gotta start using Squarespace. Build it beautiful. I'm just kidding. No, actually, I had a lot of anxiety when I read this, Kevin, because I thought, you know, while government websites have many detriments and flaws, the one thing that they have not had historically is a cleaning fee. But now that you have the co-founder of Airbnb taking over? I'm nervous. I'm real nervous.
Yeah, they're gonna charge you like what looks like $50 to renew your driver's license, but then when you get into the fine print, it's actually $400 for reasons that are outside your control.
Mm-hmm. It does seem like maybe every federal government website will now look like basically exactly the same, because that's also sort of the Airbnb MO, you know? It's, uh, it all just kind of looks same-ish, no matter where you are in the world.
Yeah, now, this was not totally unexpected. Joe Gebbia was part of the Doge push. He has become one of the more outspoken pro-Trump sort of people in Silicon Valley. And yeah, I think this is, uh, this is gonna be interesting to watch. Part of the reason, in my understanding, that government websites look so bad and work so poorly is because they are very complicated and they, people rely on them for a lot of different services and benefits and things like that. And you can't really screw them up like you could with some sort of commercial website that you can just turn around and fix.
Yeah, well, I'm sure that was the case in the past, and I'm also sure that going forward, we'll just delete them and start over and see what broke.
We're vibe coding the government now, Casey.
Do you ever get really relieved that you're not gonna need any social security for like at least another 20 years until we get this stuff straightened out? I do, I do.
Yes, let's fix the websites.
And that was Trump's in tech.
Building and scaling AI today can feel pretty chaotic, mostly because teams are juggling too many fragmented tools at once. That's where OutSystems comes in. OutSystems gives you one unified Agentic Systems Platform to manage your entire portfolio of agents, apps, and workflows. You get to innovate at the speed of AI, but with the security, governance, and control your enterprise actually needs. And the results are already there. A major manufacturer used OutSystems to build a multi-agent system that analyzes blueprints. And a global bank cut their development time by 60%. Own your Agentic future today at outsystems.com/NYt.
Hard Fork is produced by Whitney Jones and Rachel Cohn. We're edited by Jen Poyant. We're fact-checked by Caitlin Love. Today's show was engineered by Chris Wood. Original music by Alicia Baty-Tup, Marion Lozano, Rowan Nemisto, and Dan Powell. Video production by Sawyer Roque, Pat Gunther, Jake Nickel, and Chris Schott. You can watch this full episode on YouTube at youtube.com/hardfork. Special thanks to Paula Schuman, Quiwing Tam, Dahlia Hadad, and Jack from Miranda. As always, you can email us at [email protected].
You can't undo everything in life, but you can undo prediabetes. More than 2 in 5 adults in the US have prediabetes. Most don't know they have it. But with early diagnosis and action, you can delay or even prevent type 2 diabetes. Visit doihaveprediabetes.org to take the 1-minute risk test and start undoing the things you can. Brought to you by the Ad Council and the Centers for Disease Control and Prevention.
Page updated October 26, 2026 · transcript and summary are machine-generated
