This Week in Tech Episode 1108 Transcript

Please be advised that this transcript is AI-generated and may not be word-for-word. Time codes refer to the approximate times in the ad-free version of the show.

Devindra Hardawar [00:00:00]:
You're watching TWiT, This Week in Tech. I'm Devinder Hardwar sitting for Leo Laporte. We've got Paris Marks here and Anthony Ha, and we talked about Apple's new touchscreen MacBook Pros and many, many bad AI vibes coming up on TWiT.

TWiT.tv [00:00:15]:
Podcasts you love.

TWiT.tv [00:00:17]:
From people you trust. This is TWiT.

Devindra Hardawar [00:00:21]:
You're listening to TWiT, This Week in Tech, episode 1,106. Recorded on November 1st, 2026. Bad AI vibes. It's time for TWiT, This Week in Tech. I'm Devendra Hardwar, senior editor at Engadget. I've got a great pair of folks to chat about this week's tech news with. We've got Paris Marks, author of the new book Hyperscale and podcast host at Tech Won't Save Us. Paris, how's it going?

Paris Marx [00:00:53]:
Yeah, it's going pretty well.

Devindra Hardawar [00:00:55]:
Awesome. You are in— you are literally in the middle of your book tour right now, so I'm glad we were able to, like, catch you at a good spot. Where are you sitting right now?

Paris Marx [00:01:03]:
Yeah, I'm in Montreal. I have like— I was in London. I flew in yesterday, I think, or the day before or something like that. And yeah, I have like a few days of a break. I got to go to a Halloween party last night and now I go to Boston on Tuesday and then kind of pick up and then it's a whirlwind from there, basically.

Devindra Hardawar [00:01:21]:
That sounds like fun. Like, I've seen the pings about like where you're going for your book tour. That sounds amazing. And hopefully you're having like a great reception because, hey, everybody hates AI right now, and everybody hates data centers. So perfect timing.

Paris Marx [00:01:33]:
Absolutely. The book has arrived at the ideal moment.

Devindra Hardawar [00:01:37]:
Perfect, perfect timing. And that laughter you hear is Anthony Haas, TechCrunch's weekend editor. Anthony, how's it going?

Anthony Ha [00:01:44]:
Good. Yeah, I was— this is a workday for me. So I think I hit publish on my last post 5 minutes before jumping on here. So if I sound incoherent, or either really tired or really energized, that's why.

Devindra Hardawar [00:01:56]:
That's why. And yeah, we've got a whole bunch of stuff. I mean, Anthony, you want to talk about some of the weirdness around AI? Paris, we'll certainly talk about your book as well. But I'd love your perspective on some of this week's biggest tech news. And I don't know where to start. Where to start for me is like Binance and that $6.3 billion hack, which I believe is the largest crypto hack we've seen so far. And the latest story we've heard is that Binance says that 90% of the hack, of its losses, losses are covered. It's covering the losses to, uh, to its customers.

Devindra Hardawar [00:02:31]:
That doesn't mean those funds are recovered. That means they have a crap ton of money in reserves and other places to replace that money. And, uh, they say 10%, I believe, is like in a— I think eventually they're going to try to recover.

Anthony Ha [00:02:47]:
I think what they've said is that they're trying to basically— the last— and I think actually now it's 95 to 5, or at least it was a few hours ago. But I think they're basically like, we have to They're trying to buy whatever assets were stolen. And so it's like, here are a bunch of rare, less liquid crypto assets and we have to go buy them and it's just going to take more time. But they're saying that their commitment is that no one except for Binance should lose any money on this.

Devindra Hardawar [00:03:13]:
And Binance is losing $6.5 billion on this to recover this, plus, plus the like respect and trust of its users and whoever knows how much.

Paris Marx [00:03:21]:
It still had that.

Devindra Hardawar [00:03:23]:
Apparently it had both this amount of money in its, you know, in its, in its user base at this point and the ability to just cover it entirely, which is wild. Anthony, I saw the news that the FBI says that it's likely North Korea at fault for this, right?

Anthony Ha [00:03:42]:
Yes. I mean, this is not a group that I know a lot about. I think it's called like, what's it, Trader Trader? Or maybe it's the other way around. But yeah, sort of like a known North Korean hacking group, which I don't know. I saw that and I was like, sure, that makes sense. That wasn't a huge surprise to me. And I think that it's, I guess, good to know who the culprit is. And maybe in some way there's other ways that Binance might— I doubt there's any way that Binance is getting that full $6.3 billion back.

Anthony Ha [00:04:15]:
But apparently the FBI is involved. Maybe they'll get some of it back. It's unlikely. I think this is really, really, really bad for Binance, and not just because of the financial impact, but also because they froze withdrawals for several days. They just started letting that happen again, I think, early today. And, you know, no surprise that once people could withdraw money again, they started pulling out of— pulling their money out of Binance. I think they said in the first few hours it was like $3 billion had been pulled out of the system, which they said was about 2% of their total assets under management. So they're not bankrupt yet.

Anthony Ha [00:04:55]:
But like, it's, you know, it's not surprising that folks are like, I don't know about this.

Devindra Hardawar [00:05:01]:
My question is, who are these customers? Right? Like, I feel like the crypto wave was the late 2010s. That was like peak Web3 era. We were in, you know, we were in Bored Apes and whatnot. And yet, I guess these are the people who are still trading crypto and doing a lot of business there. I am still confused about who this market actually is.

Anthony Ha [00:05:24]:
Wild generalization, but I think there's sort of 2 overlapping circles that might just be one, one circle, which is true believers and criminals.

Paris Marx [00:05:31]:
That sounds about right. Yeah. I have to say, I do kind of like— yeah, absolutely. I do kind of like sit around every year, like waiting for the story of like how much North Korean hackers have managed to steal in crypto. Every year, and it's like listen, we cannot love North Korea, but I just kind of love this story. It's like you know the rogue hermit state with like the apparently very capable hackers at like stealing billions of dollars of crypto of crypto a year to like fund its budget or whatever the hell it does with it. I don't know, but I just find it hilarious because I also can't stand these crypto companies, and I'm like you know what? You all had it coming, and if it's North Korea taking it from you. I don't know.

Paris Marx [00:06:14]:
Better in their pocket, I guess.

Anthony Ha [00:06:16]:
I wouldn't go as far as better in their pocket, but I think maybe say like, yeah, it's all bad. So the thing where you hate everyone, you just say, well, this just seems like good, clean fun.

Paris Marx [00:06:27]:
Yeah, yeah, totally.

Devindra Hardawar [00:06:28]:
It's totally fine. We still don't know who Satoshi is, right? Like, that has not been fully, fully unearthed. I think the— while— do you have a sense, Anthony? I have not kept on top.

Anthony Ha [00:06:38]:
No, but I was going to say that you may also be aware that there is going to be a Doug Liman movie investigating this question. Yeah, I think the backgrounds are going to be generated by AI and people are going to love that.

Paris Marx [00:06:50]:
Fantastic.

Devindra Hardawar [00:06:50]:
Doug Liman, what are you doing? I used to love him. I've loved a lot of his early movies. I don't know what's going on, but I think the funniest revelation would be like, oh, Satoshi was actually a North Korean conglomerate all along, right? Who has benefited the most from this whole market of fake money and, I don't know, created value out of nowhere? It's criminals in North Korea, really. This is not the first time this has happened, right?

Paris Marx [00:07:15]:
I feel like Marc Andreessen has made some money off of it too along the way. That's

Devindra Hardawar [00:07:20]:
true. They're all—

Paris Marx [00:07:22]:
Unfortunately. Yeah. Again, I'd prefer it in North Korea's pocket than Marc Andreessen's.

Anthony Ha [00:07:28]:
Coming out strong. Wow.

Paris Marx [00:07:31]:
Yeah. First appearance. Yeah. This is how I'm getting started.

Devindra Hardawar [00:07:35]:
Put that on the COVID of your book. Yeah. This guy likes North Korea more than Marc Andreessen. And you know what? Who has done more global harm at this point? Let's

Paris Marx [00:07:43]:
listen. I think you get a surprising number of people on my side on that one.

Anthony Ha [00:07:48]:
I might be coming around to this. Yes. Yeah, I think so. The debate you're seeing is not do I agree with that, is do I want there to be a clip that someone can pull of me saying that? That's—

Devindra Hardawar [00:08:00]:
Paris, I hope you told your PR person about this, and I hope they're just like, what do you do? What's going on?

Paris Marx [00:08:05]:
They're going to be thrilled. Yeah.

Devindra Hardawar [00:08:07]:
Oh, man. I have— I've historically been very crypto skeptical and it's weird. I don't know if you guys have noticed this, but I've seen some of the, some of the brightest minds in the tech world just like get really enamored with the whole crypto thing. And then that's all they can talk about after a certain point. And it's been, it's been a very strange thing. I thought we, I thought we were past this. I thought AI is the new, like, false hope, the new false god of tech. Or at least for a certain set of people.

Paris Marx [00:08:35]:
Yeah. And I find it really funny how like you had a load of people a few years ago who were telling us that like crypto was the next big thing and it was going to transform everything. We all had to get on board and, you know, like the usual thing that it's inevitable. We're hearing. Yeah, exactly. And those people, just like so many of those people, just pivoted completely to telling us that now AI is the next big thing and we all need to be using generative AI and it's going to transform everything and it's the next industrial revolution. And like, there was not a moment where they had to, like, take a pause and be like, yeah, we were pretty wrong about all that crypto stuff. You know, it was just like, no, they can, they can fully just move on from this.

Paris Marx [00:09:12]:
They can fully, you know, start evangelizing for the next big thing. You know, there's another big book out recently by none other than Kevin Roose, who was also telling us that, you know, crypto was going to be the next big thing. And he had his big guide that he put out for everyone to help get in on it, you know? And it was like, yeah, these people don't need to apologize. These people don't need to, like, not apologize, but like, they don't need to, like, have a made mea culpa for like, we believed in this stuff. It clearly wasn't accurate. Or like, you know, we fell for some stuff that, you know, we shouldn't have fell for. But now we're just going to push the next thing at you. And, you know, we don't need to, you know, kind of explain why we got it wrong last time.

Devindra Hardawar [00:09:49]:
It's the state of the world right now, right? Never apologize. Never apologize for anything, especially if, like, it makes you rich. If you, if you benefit from it, right? You never have to say like, hey, this was not bad. It worked for me, I guess.

Anthony Ha [00:10:01]:
Well, I think—

Devindra Hardawar [00:10:01]:
say that. Yeah.

Anthony Ha [00:10:03]:
I definitely tried to pull the, like, what about crypto card in, like, debating people who were, like, very gung-ho about AI. And, you know, I think their response is like, well, if you just held on to your Bitcoin or your Ethereum from, like, a few years ago, there was a crash, obviously multiple crashes. But actually, you know, there were periods in the last couple of years where that turned out to be a very good investment. I actually don't know where it is right now. I don't have any myself. But But I think that part of what they hold on to is that if you're just talking about it purely as an investment token, if you're willing to endure some pretty spectacular turbulence and crashes, it hasn't always been a bad investment. That's right. And I think a lot of that also has to do with the larger political environment, which is that the Trump administration has spent a lot of the past couple of years pardoning a bunch of crypto executives, including the CEO of Binance, and really trying to really trying to make crypto happen again.

Paris Marx [00:11:00]:
Poor old Sam, you know, the last Sam, he's still rotting away. Oh, maybe Altman will join him soon. I don't know.

Devindra Hardawar [00:11:07]:
But it's Sams all the way down. Yeah. Our president has personally also benefited from the rise of crypto and his own meme coin and everything. So again, it's— it feels like criminality all the way down. At least crypto, you like— it was easy to ignore and just be like, hey, this is this weird thing some people are doing. It's making some people money. That is a thing. I remember back in the early days, like, you could buy Teslas with Bitcoin, right? And if you had, like, if you were one of the early people that was out there mining Bitcoin in your apartment, you could probably, you know, if you— and also if you held on to that money, you could probably make a certain amount of bank.

Devindra Hardawar [00:11:46]:
Whereas AI has things that can seem, that can seem actually impressive to people, right? And Now we're in the situation where I think we— the whole world has been dominated because of a couple of executives that saw these AI demos and were like, we can, we can do this, we can, we can lead the way. And that's what happened to Satya Nadella and Jensen Huang and everybody else. But we'll dive more into all that stuff later. Do you guys have any other crypto thoughts? Crypto thoughts? I would love to get these crypto thoughts out of my head.

Paris Marx [00:12:19]:
Yeah, I would just say You know, I had not really thought about Binance in quite a while before this happened. Yeah. And then it did occur. You know, the hack occurred and there was the stories about blah, blah, blah. And I was like, oh yeah, they're still around. You know, they're still at this stuff.

Devindra Hardawar [00:12:35]:
Doing the thing.

Paris Marx [00:12:36]:
Yeah. There's like, I'm sure they're still incredibly like shady and untrustworthy, but they're still chugging away such that they have $6.3 billion not just to lose, but then, you know, to be able to kind of backfill and make up. Yeah, exactly. And, you know, I guess it's just kind of for me, like another reminder that, yeah, sure, crypto did not, you know, was not kind of the transformative technology that the, you know, so many of these companies and executives tried to make us believe, you know, that it was going to replace the financial system and all this kind of stuff that they were saying at the time. But, you know, it's still around. Of course, they're still spending loads of money on US elections to try to buy favorable regulation and politicians that will give them favorable regulation. But yeah, I wish it would just fully go away. But it was also like, okay, great.

Paris Marx [00:13:24]:
You know, they were hacked. Like, yeah, I, you know, there's not even like the smallest violin that I'm, that I'm hearing because, because of a hack of crypto these days.

Devindra Hardawar [00:13:33]:
Props to North Korea.

Paris Marx [00:13:35]:
I'm just saying everything today. I need to— I need to not talk.

Devindra Hardawar [00:13:39]:
Put it on the COVID Put it on the COVID Yeah. If you guys are interested, if listeners are interested in following this line of thought, I think I don't know if Twit has talked about Gil Doran's amazing book, The Nerd Reich, but I think that book, that book really does consolidate a lot of the bad stuff that's been happening in Silicon Valley and really, I think, focuses on that. So check out that book. And we'll—

Paris Marx [00:14:01]:
I'd also say, I'd also say props that like Gil Doran is a really nice dude. I have my book launch in Montreal on October 23rd, and he happened to be up here for an event like just before it. And he was like, I'm supposed to fly out that day. I just reached out to the people at the conference. I'm going to stick around for an extra night so I can come to your event and we can hang out. And I was like, dude, you're the best. And so, yeah, we had like a nice evening after I had my book launch. So yeah, he's a good dude.

Devindra Hardawar [00:14:27]:
Awesome. Nice guy. And also righteous book. Righteous book about how the tech world has fallen to shit. And I don't know, I feel like every time I'm on TWiT, I'm just kind of— I kind of lean on that because, folks, I know the TWiT audience. I know, I know, I know you all. I know you all love this stuff and you can both love this stuff and hold it critically. And here, here's another example of that, right? Apple has new MacBooks out, and I think for a lot of people they may be a little controversial.

Devindra Hardawar [00:14:53]:
We finally gotten the first touchscreen MacBook Pro. It is called the MacBook Pro with ProLight XDR display because I, I don't— something has gone wrong with Apple's marketing brain over there. You could have just called it the MacBook Pro Ultra. You could have, you could have added just another name to Pro. Mm-hmm. Because that's really what this thing is. I also feel like the touchscreen is not the sort of big selling point. The big selling point is like they have fully redesigned these computers.

Devindra Hardawar [00:15:23]:
The 14-inch one is 3 pounds down from like, I think, 3.7 for the 14-inch MacBook Pro and the 16-inch is down to 4 pounds. So really nice, noticeable difference in weight there. And the ProLED XDR screen is apparently very, very bright. So this is the first, also the first MacBooks with OLED. Apple has had OLED in their, um, tablets and phones for a while, but, and they've been leaning on mini LED for their, for their MacBook. So super bright, finally got OLED in Macs. And I feel like they've just been itching to get this redesign in because MacBooks actually got bigger after the, um, was it the failed Touch Bar redesign where collectively everybody was like, no. No, please, no.

Devindra Hardawar [00:16:11]:
Give, give me real ports. Give me HDMI. Give me an SD card slot. They did go super slim then, and the MacBooks got bigger and thicker as they responded to user demand. So that was like with the, what, the M2 series Pro, the M2 Pro, I believe. And yeah,

Paris Marx [00:16:26]:
I remember people celebrating that they got kind of chunky, right?

Devindra Hardawar [00:16:29]:
They got chunky. And you know what? I'm fine with it. If you give me my SD card reader, you give me an HDMI slot, like give me the things that let me be pro. I can, I can deal with, you know, a slightly, uh, you know, heavier thing. And that's really what these, these machines are. And for the privilege of having a slightly lighter MacBook Pro with OLED and a touchscreen, you're paying a lot more. Yeah. The 14-inch starts at $2,899.

Devindra Hardawar [00:16:54]:
The 16-inch model starts at $3,399. I feel like that's a, I believe that's like a $400 premium. over the existing MacBook Pros. And it joins the, uh, the stuff that's happening on the PC side too. The RTX Spark machines coming from NVIDIA and, uh, and Microsoft's partnership there. I was over at the launch event for that thing and I've reviewed one of those, the ASUS ProArt P16. It is a weird time, right? Where everything is more expensive, memory is more expensive, and across the board everyone's like, All right, but how about even more expensive computers? Let's make them even more obscenely expensive because that's what people really want. And the sad thing is, I'm sure there's an audience that will gobble this up too,

Paris Marx [00:17:41]:
right? Like any of these—

Anthony Ha [00:17:42]:
How big is the audience?

Paris Marx [00:17:43]:
Yeah, I agree. I feel like it's the same with the Duo, right? It's like there's certainly an audience for that, but like, how big is it in practice? Like, how many units are they actually moving? How many have they actually produced, right? Like I'm not sure that we really have that, that figure. You know, there's, there's obviously people— there's always with any new Apple product, it's like there are people who are going to run out and buy it. You know, they did the iPhone Air last year and I was like, I do not see the point of this product.

Devindra Hardawar [00:18:07]:
They don't even still— I've not heard Apple say the word iPhone Air in like a year, honestly.

Paris Marx [00:18:14]:
Because it was like a bridge product, right? Between like it was kind of like to set the stage for the Duo and like just to sell some phones kind of, you know, to start to introduce this kind of thing. But it does feel like, as you're saying, like it's not just the story of like the broader PC industry right now. I feel like it's the story of Apple too, right, where the prices are going up like across the board, but then they're also finding ways to introduce more expensive products and try to incentivize people to buy some of those ones as well. Like, I feel like, you know, obviously you have this new, more expensive MacBook. You know, you had the updates to the, to the, to the iPads recently. But even on the iPhone side, it's like they did update the iPhones, but you didn't even get the option to buy the cheaper iPhone yet because they're not going to update that one. If you want all the newest stuff, you need to buy the more expensive iPhones, the Pros and the Duo, because they're not even going to give you the option to get like the kind of regular models until the new year. And before it was just like the cheaper— exactly.

Paris Marx [00:19:15]:
Yeah, I think so. Yeah.

Devindra Hardawar [00:19:17]:
Yeah. Because I just upgraded my wife's phone. She had like a 15. from way back when. And it's like, okay, yeah, the only choice now is a 17 for her because she's not a pro phone person. She's a basic entry phone person. So yeah, she had to go for the 17. And that did feel like, oh, you were clearly manipulating the market in this way to just be like, okay, cheap stuff comes in spring.

Devindra Hardawar [00:19:38]:
In fall, we can highlight the more expensive stuff and the really eager beavers have no choice. Like, are you going to sit— are you really going to sit here and wait 6 months? Or you're gonna just upgrade to the new Pro. That felt particularly—I don't know—kind of kind of sad for me. My wife won't really care because there was also not a huge difference between the 18 Pros and the 17 Pros this year. I will I will put my cards on the table here. Like I'm a Duo owner. I know a lot of people who are excited about the Duo, but at least that's like that was a substantive change, you know, and also something Apple was clearly many many years late to. Uh, they are also very late to touchscreen computers, but also who— it's not that big of a deal, right? It's nice that Windows computers have had them.

Devindra Hardawar [00:20:21]:
Uh, but also MacBooks have had amazing, um, touchpads forever since like the 2006 Mac redesign, I believe. Uh, whenever the first MacBook Pro was launched, I just did a retrospective over in Gadget for that. And that was when the glass touchpad was introduced and it felt so good. It felt so good that you never missed the ability to like really swipe up at the screen. Now you can do that with these ProLED machines. And it also seems like, um, I don't know, the implementation is a little chicken shit. I don't know. It feels, it feels a little weak.

Devindra Hardawar [00:20:54]:
It feels like, okay, I can scroll the web, I can rotate 3D models. I can like, um, uh, I think scrub through timelines. I also saw somebody do, uh, developer testing. So you can test like an iPhone app as if, as if you were actually running it. I could see that being really useful. But beyond that, it's like, it's the redesign selling these machines. And it's the fact that it's just a super bright, really nice screen. It's not so much the touchscreen.

Devindra Hardawar [00:21:21]:
There were reports that Apple had some sort of like hybrid interface mode to juggle touchscreen and normal stuff. And we're not seeing any of that. So it just seems like pretty basic touchscreen stuff here. I don't quite know. Like, the thing is, the AI people and the AI researchers and the companies working in AI and who want to build on this, they tend to have deep pockets and they don't mind investing in RAM, right? So that's how I ended up with an ASUS laptop that essentially would retail for $7,800 because it had 128 gigabytes of RAM in it. And you can do that to a MacBook Pro. Right now, even without the touchscreen ones, you can plug in as much RAM as you can there. And I'm of two minds about that.

Devindra Hardawar [00:22:06]:
Like, yes, it is obscenely expensive. These folks have like, the AI industry has killed RAM pricing and storage pricing for who knows how long, but also this kind of like the North Korea thing. Also, you want to run the AI on your own. You don't want to hit the, hit the cloud anymore, and you're going to spend the money to do that. There's a bit of a, I guess, hell yeah. Yeah. about. Let's, uh, bring our computer computing back to computers, and we can, uh, jump to a Sam Altman, like, uh, story, something he mentioned recently this week.

Devindra Hardawar [00:22:37]:
But Anthony, go ahead.

Anthony Ha [00:22:38]:
That was exactly what I was going to say, is that actually, if you are an AI true believer, maybe, or at least a Sam Altman true believer, you don't think it's a good idea to have a lot of, uh, RAM on your laptop because then you are hoarding it. You are depriving the machine god of the memory he or she needs to grow and thrive. Which is essentially what Sam Altman said on this podcast, that he thinks— I forget the exact wording, but he definitely thinks it's bad that there's all this RAM on our personal computers.

Devindra Hardawar [00:23:07]:
And specifically RAM for running local models, right? It's not just like RAM for gaming. He does not want you running models, AI models, on your computers, because if you were doing that, you were not paying OpenAI and other AI companies subscription fee, right? So there's something to be said, as expensive as it is, for, hey, you have a sick-ass computer and you can, you can do a lot with that. You can do a lot of work with that without touching the cloud and relying on anybody's services. And I do feel like there was something a little, um, I don't know, old school about that. I've been in the hardware game long enough. Uh, I went to college building my own computer in 2001. So I've been like focused on pricing and how this stuff works for a very long time. And it's always been expensive.

Devindra Hardawar [00:23:53]:
The— this RAM crunch is terrible. Uh, we also had a GPU crunch for a while too. Um, hopefully it'll get better, but, uh, I don't know. I, I could see a market for this. It's just, to me, less compelling than the iPhone Duo and other things there. And oh my God, that Sam Altman quote. Does he just want to— does he want everybody to hate him? I guess he doesn't care, right? He's on top of the world, so he can say what he wants.

Paris Marx [00:24:13]:
I feel like he's like so oblivious that he just doesn't realize that, you know, this is gonna cause people to hate him. You know, like he's so disconnected from regular people. I remember, I remember like this is like 2 or 3 years ago now. I was talking to Julia Black, who's now a reporter at Vanity Fair, but at the time was a reporter at Business Insider. And she did an interview with Altman and asked him, tried to get him to consider what the impacts of the things that he was doing would be, especially with the vision that he had for regular everyday people, for single parents. And He just could not conceptualize what these people's lives would even look like, let alone what the impacts of his ambitions would be on regular people. And that's going back a few years now, and I can't imagine he's become more down to earth since then.

Devindra Hardawar [00:25:05]:
Oh, man. Speaking of AI, go ahead, Anthony.

Anthony Ha [00:25:08]:
Well, no, I think it's not just a Sam Altman problem, right? It's the fact that all of these tech billionaires and executives, like the main people they talk to is each other. And so they're just sort of all like, as much as I think, you know, some of them may be more relatable than others, some of them may have like better intentions than others. But if like your baseline is other tech CEOs, VCs, then you're, you're going to be disconnected from what normal people care about to some extent.

Devindra Hardawar [00:25:35]:
I want to talk more about that. I want to talk more about that, Anthony, because I feel like that goes to a deeper problem facing this entire industry. and a lot of these people. But actually, let's, let's take a break for a quick word with Leo Laporte. All right. Thank you so much, Leo. Anthony, to that topic, I saw this post at TechCrunch, AI execs may be too out of touch to build for the general public, by Amanda Silberling, your colleague there. I feel like this is a very important thing to just think about because in my conversations with a lot of these folks, I've not talked to Satya Nadella, but I've talked to a lot of Microsoft executives.

Devindra Hardawar [00:26:11]:
since the launch of like Bing Chat, when that was the big partnership with OpenAI and Copilot and all those rollouts. And there's a weirdness going on, just like a complete, like, worldview, I don't know, separation from normal people. Can you expand on like what you were, you were going to say earlier, Anthony?

Anthony Ha [00:26:30]:
Yeah, absolutely. I mean, I think that the sort of news hook for Amanda's story was this statement by Marissa Mayer, the former CEO of Yahoo, who has this new AI app called Dazzle. And she was talking about how, I guess, there was a school assignment where she had to come up with 3 words, 3 bullet points to describe her daughter. And she was apparently confounded by this. And so she asked Dazzle, came up with 3 adjectives. And I think understandably, a lot of people were like, why is this your use case for AI? And it also fit into this sort of long line of other examples where, you know, they're kind of making these very strange statements about how AI is useful. You know, Sam Altman has talked about the idea of like creating a customized podcast to talk to your child on the drive to school. He's talked about how he can't imagine raising kids without ChatGPT.

Anthony Ha [00:27:23]:
Mm-hmm. Mark Zuckerberg has talked about the idea that like AI could replace your romantic friends, your therapists. And so there is just this feeling of like, why are you pitching this as a desirable future? Yeah, I'll leave it at that. I have plenty of other thoughts, but I'll leave it at that for now.

Devindra Hardawar [00:27:42]:
It is. It is wild out there. Also, Marissa Mayer, who was CEO of Yahoo when we were at Yahoo together, Anthony, seems like a very Marissa app that she's launching here and sort of like a, I don't know, her statement here too. She never struck me as somebody who actually had like generally good ideas. Right. So this app Dazzle, Dazzle. I mean, what did she do?

Anthony Ha [00:28:08]:
So, okay, wait. Okay. So, I mean, I'm not an expert, but I will say that what was often— she was kind of pitched as, and kind of the main line in her resume, right, is her time at Google, and that she was the one who was just like very, very militant about the idea that you have to protect the core search experience and that it was like 10 blue links, nothing else, you know, like no distractions. And so, um,

Devindra Hardawar [00:28:33]:
I will give her props for that. That was back in the day. That was what made Google special. They sure ran away from that real quick.

Anthony Ha [00:28:38]:
Yeah. RIP Google Search.

Devindra Hardawar [00:28:40]:
RIP Google Search, which is functionally useless to me right now. I, I try to go back to search for a post that I know I wrote in 2012 or wherever, cuz I've been doing this forever. And, uh, nope, Google can't find it. DuckDuckGo can find it. Other non-AI searches can find it, but Google's index is, uh, has been shot. Because it's all AI-based now. So anyway, but I do wonder how much Marissa Mayer can lean on, I was an early Googler who really honed in on like what made Google good early on. Because this Dazzle product, it looks at your photo library, right? Yeah.

Devindra Hardawar [00:29:14]:
And it can make like, you can have like AI inferences based on that. Oh, okay. That seems kind of nice. This idea of like asking for it to define your kid in certain ways. That's just plain weird. And also shows like these are the people who are like, oh, this feels like the class of people who have nannies, the class of people who have like, who have offloaded already so much of their child rearing and their home life. Now it's like, okay, you can also offload basic elements of thinking about your child and their personality to AI. Just seems kind of sad to me.

Devindra Hardawar [00:29:51]:
We've also seen stories. I know Satya Nadella listens to Was it, uh, auto-generated AI podcasts of news?

Anthony Ha [00:29:58]:
It's that he uploads the transcripts of his favorite podcasts into Copilot and then asks Copilot about the podcast on his drive to work, supposedly.

Paris Marx [00:30:06]:
Oh my God.

Devindra Hardawar [00:30:08]:
I hate it. I hate, I hate all of this. And also Jensen Huang of NVIDIA also said like, I don't even know my address anymore. And it's like, I know you're lying. I know you're lying. What are you selling? What carnival barking are you doing over here? You don't even need to do this. want your stuff. You don't need to tell them.

Anthony Ha [00:30:24]:
Well, also, Jensen Huang has probably not driven himself in a car for many years. That is actually what that story is illustrating.

Devindra Hardawar [00:30:32]:
It could be. But it's also like— Nobody knows his address. Yeah.

Paris Marx [00:30:35]:
How— like, how— not even like how out of touch these people are. Like, you know, what kind of regular person would think it is okay to be like, yeah, you know, my kid had a homework assignment and we had to figure out 3 words to describe her, but I couldn't come up with those myself. because I don't know her well enough. And so I asked like an AI, right? I think most people would be like, that sounds ridiculous, right? Like, I can give you 3 words to describe my kid. There might be a word in there that's going to be about how annoying they are sometimes. But like, you know, I still love my kid and I know who they are, right? You know, and I can easily come up with a way to describe them in a way that apparently these guys can't. But it also stands out to me recently, you know, to kind of talk more broadly about what you're talking about, the number of these people who have either come out recently and talked about having AI psychosis. Like, I believe Elon Musk said recently that he thinks he has some degree of it.

Paris Marx [00:31:26]:
And of course, there was this big controversy this week around Proton, you know, like kind of like one of the more European alternatives to some of these companies where Andy Yen, the CEO there, was, you know, pushing the employees to use a bunch of AI and was saying that they needed to kind of like go through a bunch of performance reviews to see who was using AI. And there was probably going to be layoffs as a result. And, you know, obviously this really blew up. You know, 404 Media in particular was reporting on it. And, you know, one of the things that the employee said was that, like, Andy is very clearly, like, experiencing heavy AI psychosis right now. You know, that's, that's their, their view on it. He hasn't confirmed it. But, like, I would be wondering how many of those people are experiencing, you know, what we would generally consider to be AI psychosis.

Paris Marx [00:32:13]:
And, you know, to them, it's like it's not even something that they are, they are thinking about or that they are worried about, but they are just like so bought in on this notion that AI is everything that they constantly say it is that, you know, they just don't even question it. You know, it's almost remarkable that Elon Musk has the ability to notice that maybe there's something going wrong with, you know, the way that he's using AI and consuming it in a way that we could all see it very clearly with how he was posting, you know, weird girls that he was generating with Grok. But, you know, he seemed to think that that was quite all right for quite

Anthony Ha [00:32:49]:
I think one of the other things that speaks to is also, like, and I'm curious what both of you think about this, but to some extent, when we talk about the tech industry, obviously we're talking about a lot of different communities and different groups of people. And I do think that the tech industry in general is just much more gung-ho about AI than the general population. Absolutely. But also there's a big divide between the people running these companies and the people working at them. And that, you know, that, that, that, you know, I don't know. Again, we can debate how much of a difference this makes or whether this should influence how we think about these companies or treat these companies. But when the CEOs are saying these kind of unhinged things, it's not necessarily the case that everyone who works for them is like, yeah, this is great, we love this too. But they're also like, this sucks, but I don't know what to do about it.

Paris Marx [00:33:37]:
And also, I want to keep my job, you know, while there are— many of these companies have been doing massive layoffs for the past few years. Right. And so there are a lot of people who just want to keep, you know, in work at the moment and not be kind of lost to, you know, the layoffs and not knowing where they're going to go. You know, we see that even more extreme in the video game industry. But, you know, it is happening in the tech industry as well, right?

Devindra Hardawar [00:34:00]:
It's— I feel like it's a combination of that. There are true believers out there too. Like, we see a lot of reports. I don't know. I've lost count of the amount of things that are essentially cults in Silicon Valley and also essentially sex cults tied to AI. What's going on over there? You got— you're just a bunch of nerds and you're trying to build God. And I don't know what this is, how somehow evolved. I don't know, this is like the evolutionary, um, trajectory of everything goes to crabs.

Devindra Hardawar [00:34:27]:
The evolutionary trajectory of a lot of these, um, pseudo-intellectual fields ends up in a sex cult. Sorry, it just feels like that's how it's gone. But for a lot of these people, these like, um, you know, they're really driven. You're not an executive unless like you are a driven person who doing a lot of work. And I do think there is something intoxicating about a lot of these AI tools that just lets you be more capital P productive, right? Like you can just do more things. Oh, my emails are going out here. I am like, oh, I could just summarize all these things. I can have an agent just do all this work for me.

Devindra Hardawar [00:35:04]:
It's very intoxicating in a way that it feels like, oh, you are— time is limited, right? And you can do actually more in the day by offloading some things to AI. And I've— I feel like I felt some of this myself. I don't use ChatGPT or any of like the major things, but I do use occasionally Descript. And Descript is an online podcasting platform that has a lot of AI tools. And it's very easy to be like, oh man, wow, it just takes me like 5 minutes to like clip out a piece of this show and like build a transcript, right? And create a video that I can post to social media. And it's very easy to be like, oh yeah, I could just bang these out. I could just do these so quickly. Wow, I'm doing so much stuff now.

Devindra Hardawar [00:35:43]:
And after a certain point, I had to take a step back and be like, I don't— I have to actually go through these transcripts. I have to actually double-check all this content to make sure it's good because the feeling of doing so much is intoxicating. I think that Hank Green, the science communicator, has talked a bit about this too, where friends around him were like, I think you're leaning too much on this AI stuff, man. And there's a bit of controversy about that, but I think part of it was just like he admitted, like, he's just leaning on these tools too much. I don't even think it's necessarily evil. He's just like, oh wow, I could do so much. I'm gonna keep doing it. There's an addiction there a little bit.

Paris Marx [00:36:19]:
Yeah, absolutely. Especially when like you think about the, the kind of, you know, if you wanna say like the content creation space or, or whatever, it's like there's always that pressure to produce more, to put out more. Like that's kind of like the basis of it. And when, you know, this, this kind of like set of products comes along that, you know, purports to, and, and in many cases does allow you to do it more quickly. then, you know, people who are in that space in particular are going to jump on it. And I would say, you know, just more broadly, it's like, you know, on the one hand, I think the, the AI tools can allow you to do certain things more quickly, but then there's also like the perception bias of like, you know, I feel that I'm doing things quicker. Mm-hmm. I feel that I'm getting more done, even if, you know, it's not necessarily making you more productive at the end of the day.

Paris Marx [00:37:03]:
Or, you know, maybe there's a lot of things that are being created and, you know, the, full accuracy and things like that are not being checked over in a way that would have been done in the past because of, you know, a kind of reliance on the technology that is maybe a bit more than it deserves. And, you know, I just think of like we've seen a number of stories over the past year of like books being published where, you know, people were like, oh yeah, I relied on it for research and some help on editing and fact-checking and whatnot. And then it's like, oh well, you clearly relied on it a bit too much. because now there are a bunch of errors in there, right?

Anthony Ha [00:37:37]:
Yeah. And I mean, I think that also in the biggest use case right now for AI being code generation, I'm not a serious coder myself. And I think that probably there's some conflicting data about this. I'm sure there are plenty of programmers who tell me I'm full of it. But from what I understand, there are also studies that say that, yes, it seems like you are— yes, you are literally generating more code. That is true. But actually, it's not necessarily making you more productive because But then there's so much more work in the next steps where you have to go back and you have to check everything, you have to debug everything, that it's debatable whether it's actually making you more productive. That's even before we get into the more long-term effects of potentially a really profound deskilling and nobody understands these codebases anymore.

Anthony Ha [00:38:25]:
I think the other thing that's also an interesting question to me that I suspect that most of the generative shortcut use cases for this technology are bad. But I think there are things that probably do appeal to all of us. I was talking to Paul Ford, the great writer, about programming a few weeks ago, and he was talking about the idea of what if AI was just software and we treated it as if it was just software? I think there's a bunch of reasons why it's not just that and we should be more wary about it, but it's also interesting that that's not the sales pitch. The sales pitch is like, this is this much more transformative thing that has to, you know, that will like basically off— like cognitively offload literally every decision in your life. And it turns out that's for maybe a small percentage of the population, that's a very appealing pitch. But for most people, it's not. And it's— I think there's like some of it is like, I think there's a financial incentive, right, for these CEOs to make this pitch because they've raised so much money, they have so much debt that like this— their company, like OpenAI, is not going to live up to its valuation unless it literally transforms everything. But also, I think what you guys are saying too is that, that I don't think the CEOs fully realize how dystopian this pitch sounds to everyone else because of their social circle, because they have AI psychosis themselves.

Anthony Ha [00:39:46]:
Like, they don't realize that their sales pitch is not a good sales pitch.

Devindra Hardawar [00:39:50]:
And also, I feel like a lot of people around them, like, nobody's going to tell them no. Like you guys are saying, right? People are afraid for their jobs. There are a ton of layoffs. It's really hard to stand up and say the emperor has no clothes to a certain degree.

Paris Marx [00:40:03]:
Well, it's like it brings me back to when we had those messages during Elon Musk's Twitter acquisition, right, that came out in the lawsuit. And you could just see how everybody was saying whatever it was they thought he wanted to hear. The Jason Kalanakis ones always stand out to me. But you basically had all these people who were like the yes men in his his DMs and his text messages and whatnot.

Devindra Hardawar [00:40:26]:
The All-In podcast crowd, basically.

Paris Marx [00:40:28]:
Totally. Totally. And it was like a little bit of insight into how people approach and talk to these very powerful and very wealthy people in Silicon Valley. Because if you piss them off, you're out, right? And they don't want to be out of the group. Right. And that's just one example. Elon Musk is very powerful. very wealthy.

Paris Marx [00:40:52]:
But you have to believe, and I think it's very credible to believe, that that's the case with a lot of these guys.

Anthony Ha [00:40:59]:
I think also the case that maybe the part of their thinking maybe is that there have been other products where they were kind of ahead of the curve. Like, I think a lot about Uber, which I remember in the early days, I was like, this is never going to catch on outside of San Francisco. This is for 20-something tech Jackies. It doesn't make sense anywhere else. And, you know, I was completely wrong about that. I think we can— again, you can dig into why that turned out to be the case. And that's not necessarily just a reflection of consumer desire, but also, you know, regulatory landscape, things like that.

Devindra Hardawar [00:41:31]:
They also made themselves like a success to a certain degree.

Anthony Ha [00:41:34]:
Yes, right, exactly.

Paris Marx [00:41:36]:
They lost money for so long that this wasn't the market. Eroded the alternative.

Anthony Ha [00:41:40]:
Yeah. This wasn't like the invisible hand reaching down and saying, you are the true choice. voice of the consumer. This was— they made themselves a success. But that, you know, that there is like a precedent for, um, saying like, people aren't here yet, but we can get them here, right? Yeah. And I think that something about— and again, we'll see, because of course, obviously, there's still many, many, many, many people who use ChatGPT, who use Claude. So it's— but that does seem to be this split where before, like, they could say, well, we're kind of ahead of the curve, but we're kind of— and we have way more money than you, but we're still kind of on that same curve. And something about this political moment, about AI, is that now it's just like 2 completely different worlds.

Anthony Ha [00:42:17]:
So the bet that they're making seems like a much shakier one.

Paris Marx [00:42:22]:
It also feels like there's a broader backlash to this stuff now in a way that I feel like we haven't seen to the same degree in a long time, because it's like, yeah, people are critical and skeptical of AI and the impacts of it, and you very much see that reflected in the polling. They're very kind of even more against the data centers and physically turning out to try to stop them. But it's like we also see the growing movements against the flock cameras across the United States. And there have been concerns about surveillance for years, but it feels like it's really kind of coming to a head. You have the anger in some circles around the smart glasses. Yeah. Such that in the European Union, they're actually moving forward with legislation on that. And yeah, I don't know, it does just feel that like some kind of tipping point has been reached recently where like the public is just kind of really fed up with what the tech industry is doing.

Paris Marx [00:43:18]:
And I think that there's many different reasons for that. And like AI is just one, you know, kind of one way that that is being seen or felt. And that doesn't mean that just because people might use ChatGPT sometimes that they also love AI. Mm-hmm. But it's also just, you know, this is a product that's out there, it's easy to access, Yes. You know, people have been convinced that this is useful. And then, of course, you have the kind of broader issue, or I don't know how you would want to kind of like frame it, but where, you know, the way the UI, the way it's been designed, makes it feel as though you're having a conversation. And that is kind of like appealing to us as human beings for psychological reasons that, you know, leads us to appeal to that, I don't know, way of using it.

Paris Marx [00:43:58]:
So, yeah, I don't know. I think that there's a whole load of things kind of happening there. But it does feel like the public is like a bit, a bit more fed up with this stuff than they have been in a long time. And I think for, you know, for good reason.

Devindra Hardawar [00:44:09]:
It's— I mean, this is certainly the time of reckoning. And I do think a lot of people recognize the sort of what happened in the 2010s was not really okay. Right. Like, Anthony, you and I were on the ground level covering a lot of this stuff, like at startups and big tech companies. Paris, I'm not sure when you got started, like doing your, your tech criticism work. But like, you, you probably— you saw this. I'm sure you saw a lot of this. Absolutely.

Devindra Hardawar [00:44:32]:
And the 2010s were just a period of like unfettered growth, right? Do whatever you can. Oh, uh, local regulations holding you down? Don't worry, you're okay. Um, just like break the rules.

Paris Marx [00:44:44]:
Just plow through.

Devindra Hardawar [00:44:45]:
Yeah, yeah, pile through and you'll make it work. Uh, pay

Paris Marx [00:44:48]:
some fines later, you know. Exactly right. Pay to get some, uh, some, uh, propositions passed in California to make sure that, you know, you don't need to consider your drivers as employees.

Devindra Hardawar [00:44:58]:
Yeah, your VC backers do not care if you get in trouble. They only care about the hyperscale growth, right? And there are some places where there— I think municipalities did stand up. New York, um, really pushed back against Lyft and Uber. Yeah, there are no personal cars driving for Lyft and Uber in New York because of the taxi medallions and the whole licensing thing there too. So using those services services in New York feels very different because you're actually hiring, uh, professional drivers. But we saw what happened with social media. We saw what happened when we allowed these companies to, like, merge and for Facebook to become this, this gigantic thing. And everyone is like, um, what's his name in Network, right? We're all just pissed off.

Devindra Hardawar [00:45:44]:
Okay, what do you, what do you have now? Oh, you're creating God? Oh, that sounds great. Uh, you couldn't build a social network without causing genocides and destroying democracy. Now you're building digital All right. No, I think there's a lot of pushback in so many ways. And the privacy things too. People don't want the cameras everywhere. And having cameras on your face opens up so many issues. It's great to see the pushback, I guess.

Devindra Hardawar [00:46:08]:
And there's been a lot of like critical coverage for so long. I feel like people are finally absorbing that too, like on a visceral level, which is very, very nice to see.

Paris Marx [00:46:17]:
Yeah.

Anthony Ha [00:46:17]:
Yeah. I mean, I think particularly on the data center side, and I agree with Paris certainly that it's It's interesting because I do feel like AI and data centers have become, in a way, this sort of eye-opening moment for a lot of people. And I would include myself too, that I think for a long time I thought sort of being excited about technology and being politically progressive could be easily reconciled. And I think they can be reconciled, but partly it's about thinking about pulling apart your feelings about technology and your feelings about Silicon Valley as an economic system. Yeah. And I do want to be clear about that. There's a lot of technology I'm excited about. It's just, you know, how is— what are the incentives and how do things work in Silicon Valley that I think are not great? But I know there's a lot of bad technologies too.

Anthony Ha [00:47:00]:
But I think I did want to talk about, yeah, the data center thing in particular, because also we've seen, you know, this is actually an issue where the companies have acknowledged that they're really like getting held up, I think, in a number of ways. with the data center moratoriums. Both Amazon and Microsoft in the past few months have put out these statements that they're no longer going to ask local governments to sign NDAs in their negotiations, which is crazy that they were signing those in the first place. Yeah. And I think actually right before we started recording, I saw and posted a little story about how OpenAI, I guess they're planning this big political push for after the midterms. And one of the things they're going to try to do is say they're not data centers, They're AI compute campuses.

Devindra Hardawar [00:47:45]:
Oh, amazing.

Anthony Ha [00:47:48]:
It's gonna work. It's totally got— people are gonna love

Devindra Hardawar [00:47:50]:
it. Warm rooms for the servers and everything, common areas. The servers are just hanging out playing, uh, what do kids play? I don't know, ping pong.

Paris Marx [00:47:58]:
If we just change the name, guys, I think people are gonna love it, you know.

Devindra Hardawar [00:48:03]:
Well, you know what, Paris, this is a good time to talk about your book, your book Hyperscale, which is entirely about hyperscale data centers. Um, I've read a whole bunch of it already. I think it's, it's fascinating, and you unearthed a lot of things that I've not really thought about, just in terms of like how these companies have hidden what they're doing. And you don't hide something like that unless you know you're doing something wrong. I feel like— what is the pitch? What is the pitch for your— for Hyperscale?

Paris Marx [00:48:28]:
I feel like the main pitch is just that, you know, we have seen this real expansion of the number of massive hyperscale data centers in the past number of years, in part driven by generative AI. And I just feel that, you know, looking at the impacts of these data centers, that it's just unsustainable to be building computation at this scale, right? Especially to be doing it so quickly, to be trying to get these things built without proper community consultation, proper environmental review, proper planning procedures to make sure that the impact isn't going to be, you know, on a community or whatnot, isn't going to be so significant. But, you know, and I feel like it's kind of like kind of evidenced in the kind of COVID photo that we use in the United States and Canada. But it's like when you're looking at the scale of the planetary impact that they're having, you know, we can talk all about the local impacts that are driving people out to their, you know, city and town councils to try to pass moratoriums and oppose data center projects. But, you know, and this wouldn't— these exact stats wouldn't even be in the book because they're more recent. But when you look at the kind of amount of fossil fuels that these companies are building out in order to power these data centers, or even just planning to build out. There was one a few weeks ago, or last month sometime, that it's something like 189 gigawatts of new gas power generation is in the pipeline in the United States alone, just for onsite generation at these data centers. That's not even considering what they're getting from the grid.

Paris Marx [00:49:57]:
That's just staggering, right? Sierra Club put out a report just last week saying that they predict that in the United States, emissions from the power system are going to be 42% higher by 2030 than in 2025. Again, largely driven by the demand for these— from these data centers. Right. And to me, that's just like completely— like there's no way that you can reconcile what these companies and these executives are doing with, you know, kind of like a healthy and livable planet. at a moment where we're already seeing the natural disasters from climate change get worse and worse year on year, whether that's the droughts and the wildfires, but even more than that, right? I just think it's completely unsustainable. And Anthony, I really agree with what you were saying just before, and that's really what I try to get to in the latter part of the book. It's not that we have to hate technology, right? It's not that we need to try to get rid of digital technology and go back to an analog world, or that we should have no data centers at all. It's like, but what is the amount that makes sense? Right.

Paris Marx [00:51:07]:
And is it— if we're building out more computation, are we actually seeing like the public benefit from doing so, or is the benefit just accruing to these executives and the shareholders and the major investors and things like that? And I feel like technology has just become so disconnected from, you know, realizing a better world for most people, especially, you know, as we've been talking about in this program, when they're talking about building the AI god and not like, you know, how do we make things actually better for real people in the world? And they really don't seem to be, you know, the executives and stuff really don't seem to be able to conceptualize what that would be, you know. And yeah, that's kind of like what I'm trying to get at with the book.

Devindra Hardawar [00:51:45]:
I appreciate it. And the book is well worth a read, so highly highly recommend it to everybody listening to this show. Uh, it is wild to me that just a few years ago, a lot of these companies were touting their, like, environmental gains, right? They're doing so much to conserve power and everything. And as soon as this stuff started happening, we're just gonna quietly, uh, not talk about our environmental reports anymore. It's very, very sad, very disturbing.

Paris Marx [00:52:11]:
Um, I just— I, I point out a quote in the book because it, it just always stands out to me, right? Brad Smith, the president of Microsoft, You know, back in 2020, they're like announcing this big carbon moonshot that they have, right? You know, they are going to be— I think it's carbon neutral by 2030, and they're going to like pull all the emissions of Microsoft as a company since the 1970s, like out of the atmosphere by the night, by 2050. And, you know, this is like their big carbon moonshot. This is how they announce it. And then by— I think it's in '24, he does an interview with Bloomberg and they're like, you know, what's that? What's up with this carbon moonshot? moonshot. And he's like, well, the moon has moved 5 times farther away. I'm like, I don't think that's what happened there, my man. We're

Devindra Hardawar [00:52:52]:
looking for the guy who did this. But yeah, really, really far away. Um, this is the sort of stuff that's really, uh, I don't know, made me a little more angry over time when it comes to the tech world. Because at a time when we are supposed to be focusing on our impending climate apocalypse. We decided to create fake money, you know, we decided to start burning energy for crypto mining. And now it's like, oh, let's, let's like make that exponentially worse because now you have these companies and the, I don't know, the companies like just leading the way in AI who are giving up on all of their goals and promises to protect the Earth just to be a little better. to have like a bit of a foothold in here and all of it, it just feels like complete desperation for me. But you know what? We're going to, we're going to come back to all of this.

Devindra Hardawar [00:53:45]:
Let's take a minute to breathe and find some peace in ourselves and the tech world and—

Paris Marx [00:53:51]:
Lower the temperature a bit.

Devindra Hardawar [00:53:52]:
Lower the temperature. At least locally. Yeah, let's have a word from Mr. Leo Laporte. Thank you so much, Leo and Anthony, for— as we were prepping this episode, you created a section in our document called Overall AI Vibes. And I feel like the overall AI vibes are bad. It just seems like what a, what a week. The Anthropic IPO filing, multiple earnings reports where companies are taking major sock hits because it seems like a lot of this, the AI drive is unsustainable.

Devindra Hardawar [00:54:26]:
Let's start with Anthropic, though, because what the hell is going on there? I don't understand. I do not understand this S-1 filing. And I guess we how it's kind of a mess. Right. And at this point, it's expected to get a junk rating, most likely according to the Financial Times.

Anthony Ha [00:54:45]:
Yes. So, I mean, there's a number of remarkable things about the S-1 filing. I mean, one is just that the risk factors, which is always kind of what the journalists scroll to first in an S-1, is, I think, like 100 pages long. Half of the risks are just like, what if we cause societal collapse?

Devindra Hardawar [00:55:03]:
We need AI to break down this risk factor for me.

Paris Marx [00:55:06]:
What if our sci-fi dreams come true?

Anthony Ha [00:55:08]:
Yeah. I mean, okay, as an aside, I don't use AI at all in my own writing, but it's like this feeling of like, oh, I'm probably reading more and more AI-generated writing because I'm quoting these AI executives, these AI statements. I'm sure AI was involved in writing this S-1, and that's so dispiriting. That's not any— neither here nor there. As far as the financials, so they're— what they are trying to highlight, and I think other people have said, is quite impressive, is they went from $1 billion in revenue in the first half of 2025. They're now up to $16 billion in revenue in the first half of 2026, which is, you know, a real business. And, you know, I think, again, speaks to both the fact that there is like a real business here and also the expectations are so crazy. crazy high.

Anthony Ha [00:55:57]:
And so I think in order for Anthropic to be successful, it cannot just be making $16 billion in 6 months. It needs to be making much, much more than that, in part because— so I guess the headline that a lot of people focused on, which I think fair enough, is the net loss for that period was $80 billion. Some of that, I guess, has to do with debt and accounting.

Paris Marx [00:56:18]:
Easy to make $16 billion when you lose $80 billion, I guess.

Anthony Ha [00:56:22]:
Yeah. I guess they use this other term called operational loss, which is they're claiming that they only lost $690 million in their sort of day-to-day operations. And the rest of that $80 billion has to do with their debt and things like that. But here's the thing. They do have a lot of debt. They're building— they're very invested in getting access to all this compute and all these data centers. And so whether or not that's counted day-to-day operations, that all that, those billions of dollars seems like a real problem for them.

Devindra Hardawar [00:56:55]:
I don't, I don't know where we go from here. I guess, I guess everyone's like patiently waiting to see what happens with Anthropic. It is wild to me to see, um, Microsoft, Google taking stock hits after earnings reports because we have gone for many, many years where it's just like smooth sailing for these folks, right? You dump your AI numbers or you dump your cloud numbers out there and And the stock market loves it. And it seems like it's just like endless growth for them. And I feel like, again, that whole reckoning thing, I feel like it's starting to appear a little bit.

Paris Marx [00:57:29]:
Absolutely. I certainly think so. And like, you know, when I was thinking about the Anthropic S-2 as well and like, you know, how they're talking about their, their kind of, you know, operating loss and things like that, it brought me back to when we would always have a debate about how much money Uber had actually lost versus how much they were saying that they had lost. Because they had their adjusted EBITDA figure or whatever that they would always use. A lot of financial journalists would be like, yeah, but they're excluding a lot of things that a usual company would not be able to exclude in these metrics to try to make it a bit more rosy because they're just losing so much money. I couldn't not think of WeWork this week as well. The response when they put out—

Devindra Hardawar [00:58:12]:
Remember WeWork? Yeah, remember how we lived through that? Right now we're hearing a lot about millennial optimism from, from the younger generations. I feel like so many of these companies just felt like, oh man, WeWork was this like embodiment of like, oh, your workplace is your playground and you hang out and there's free beer. It's wild. Anyway, go ahead.

Paris Marx [00:58:32]:
No, but, but it was like, you know, obviously they put out their, their perspectives or whatnot and everything fell apart because it was like so ridiculous, right? Because it was so, you know, Newmanified, uh, you know, and just like his kind of delusions.

Devindra Hardawar [00:58:46]:
You're a wrestling company. That's what you are. Yeah.

Paris Marx [00:58:50]:
But like, his, his delusions were like so built into, you know, it was like you really needed to buy into this stuff. And, you know, seeing the degree to which, you know, the AI god stuff and, and their kind of weird visions of AI were so built into like the, the, um, you know, what, what Anthropic was positioning, uh, the company around, it felt very similar to me, right? And I feel like just the wider vibes of the past couple of weeks are very much like— it feels like, you know, investors and analysts are really starting to question the narrative around AI that we haven't seen, you know, more so than we've seen in a while. You know, we obviously saw OpenAI not get the amount of money that it was looking for as it was doing another raise. But one of the ones that really stood out to me was Torsten Slok of of Apollo, who, you know, has been more kind of critical of AI for a while and the prospects of it, but kind of came out and said that when you look at the S&P 500 right now, you could see that the big 7 tech companies, how they have benefited from the rollout of AI. But when you look at the other 493 companies in the S&P 500, you cannot see an increase in revenue per worker, you know, or an increase in profit margin from the adoption of generative AI tools. And I feel like that aligns with how we have seen And we've seen so many of these major companies that jumped on AI because it was the next big thing and whatever over the past few years really start to reverse and pull back and put limits on the amount of generative AI spending that employees can do. You know, over the past number of months, you know, even thinking about like Accenture saying that I think employees can only spend $29 a month on Claude now, whereas before they were happy to go gangbusters. Right.

Paris Marx [01:00:35]:
And we see Is it Microsoft or Amazon are both kind of telling employees not to token max anymore because they're just losing a ton of money, right? Yeah,

Anthony Ha [01:00:44]:
it's crazy to remember that. I think it was like this year that token maxing was something, or certainly less than a year ago, that the companies were bragging about their token maxing. And I think it's certainly been very encouraging to see all that. I mean, of course, if you're an American with a 401, looking at how your stock portfolio is doing, you may not feel so great about it. But I think societally, I think this is, this is very good news. There is an element of, you know, and Devendra and I certainly have a big interest also in sort of the entertainment business and how technology changes that, that it kind of reminds me of the moment when, you know, I think it was 2022, Netflix kind of had an earnings miss and suddenly the whole conversation around streaming changed and it went from growth at all costs to what's your profitability? And there is an irrational element where you're just kind of like, why was everyone on Wall Street so focused on one thing and then they switched to being completely focused on something else? Again, I don't love the way they're making decisions and what their incentives are, but it is encouraging that even by their standards, they're like, wow, this seems bad, and they're starting to say— not bringing up, I think, Arguably— no, not arguably, definitely more important arguments the parish brings up, which is just, we're going to destroy the Earth to do this, and I don't think we should do that. But they're also like, how is this going to pay off? We are spending so much money to do this. What are the steps that you're going to get that actually are going to— that it's no longer just being like, well, this is going to transform everything and we're going to be worth trillions of dollars? others in the future.

Anthony Ha [01:02:22]:
Now they're saying, okay, well, okay, how? And no one, I think, has a good answer to that.

Paris Marx [01:02:27]:
Yeah, I think it was Altman— what was it, last year or the year before? And they asked him how AI was going to make money, and he was like, well, we're just going to make the AI so powerful that the AI will figure out how it's going to make us money.

Devindra Hardawar [01:02:37]:
Amazing, amazing answers. I've definitely had some of the most baffling chats I've ever had with executives, like, after all this stuff. Like, the Microsoft people especially, like, you're supposed to be be the sane ones. You're supposed to be the boring ones who just give us the PowerPoints and the Excels. And, um, I've asked Microsoft execs, like, this stuff is wrong much of the time, like often. And they're like, we're gonna work on it. I would fire an assistant who was wrong about basic math a lot of the time or about doing research. Um, but they want us to just accept these tools.

Devindra Hardawar [01:03:09]:
And it's gotten better, but it's still kind of a mess. Like, to, to a very large degree. I want to point to this interview Steven Levy of Wired did with John Ternus at Apple. And specifically one good takeaway, John Ternus says he does not want to do AI for AI's sake. And I feel like that's one thing, you know, people have been dogging Apple for years because it seemed as if they have been behind on the whole AI trend. And they were, they were because they did not like jump whole hog into this obsession with like LLMs and like putting them in everywhere. But I also feel like just actually using Apple's products, I do think this methodical approach, like let's take a step back and see how like these tools could be useful. You know what's useful? I can hit a button and record a phone call and it gets transcribed on my phone locally.

Devindra Hardawar [01:04:02]:
And I could look at that and have notes. I can make a voice recording that gets transcribed locally. I've got notes. There are a lot of like little things that you know AI and smaller models can enable when they're on your device, and I do think like Apple's ultimately going to come out looking very smart about all this stuff. And I'm not an Apple fanboy. I run Windows as my main computer, but I do have an iPhone Duo. I do run iOS and everything else. I prefer Mac laptops.

Devindra Hardawar [01:04:31]:
It just seems like it's going to be really interesting to see where this all lands. I would prefer a slow bleed from a lot of these companies rather than a complete economic crash. Total collapse. That's the, that's the other alternative, right? Like, I don't— it's either going to happen badly or happen all at once. I don't see any way where they fully can sustain these businesses, these AI businesses in the same way.

Paris Marx [01:04:53]:
I feel like the Apple execs must be like so happy that their attempts at AI didn't pay off because like, you know, who was the Bloomberg guy, Mark Gurman? You know, who does that?

Devindra Hardawar [01:05:03]:
Gurman.

Paris Marx [01:05:03]:
Gurman. Yeah. You know, I read one of his reports recently and he was like, you know, they did want to have like the new HomePod screen thing, whatever that was, out sooner. And it was kind of sitting for ages because they couldn't get Siri sorted out. And it was like, yeah, they're probably happy that they couldn't get Siri sorted out because now all these other companies have gone so hard on AI and it's clearly not paying off for them. Meanwhile, Apple has been like, you know, sure, sure, their products have gotten more expensive. They've been getting more expensive for years, even without the kind of the memory, you know, kind of cost increases and stuff like that. But they have also found a great market for like Mac Minis and stuff like that for, you know, AI uses and people who are, who are running models and whatnot.

Paris Marx [01:05:47]:
You know, I think that, I think that they certainly have integrated some AI stuff where they have, you know, seen it to be useful. useful. I think they would have tried to integrate a lot more if they could actually kind of like, you know, get their shit together, so to speak.

Devindra Hardawar [01:06:01]:
But, you know, the Apple Watch, like 100% of your conference, like, that's a bad thing. And there's— well,

Paris Marx [01:06:06]:
and they, they tried to do like the, the news summarizations early on and they had to pull back on that because it was like making stuff up. Right. You know, like these issues have been there. But yeah, I think, I think they're going to come out looking good and not because they Not because I would say, not because they had so much foresight, but because they just couldn't get their act together.

Devindra Hardawar [01:06:25]:
Sometimes being like last and lagging behind, uh, can be helpful, but also, yeah, I do think in general they're more thoughtful about how they embrace new technologies. Like they're— the Duo is 7 years late to the foldable market, although I think Samsung only made it viable like after the third one of theirs. So like they're still late. They're at least like 5 years late to viable foldables and other companies have been making Android foldables forever. So, but yet they've still done things I think have been pretty, pretty useful. I think it's the thoughtlessness of the AI industry, like just in terms of rolling out these features, the way Microsoft pushed Copilot into everything. Nobody really wanted it. I was there when they launched those Copilot Plus PCs with the recall feature.

Devindra Hardawar [01:07:09]:
And I was asking people on the ground, this seems bad. This seems like a privacy Why, why are you not thinking about this? And then fast forward to the immediate response from everybody's like, we don't want this, nobody wants this. And they've had to pull back on all of those like major AI features. Just kind of a wild thing to see. I don't know, it feels like crazy times over here.

Anthony Ha [01:07:31]:
I've been sitting just wanting to like share probably my favorite example of AI psychosis, which is a few weeks old now, which is the— I'm sure you guys saw this already, but like the, uh, CTO of Palantir said in an interview a few weeks ago, talking about how a lot of AI scientists are now— AI engineers are now taking up smoking again because they're convinced that AI will cure cancer. Oh my God. Yes. I love that one. It's the same logic that they apply to the environmental stuff too, right? As they're saying, like, the data centers might be destroying the Earth now, but it'll solve climate change.

Devindra Hardawar [01:08:05]:
This is a simple climate change. Magic.

Anthony Ha [01:08:07]:
And then what we're seeing now with these tech companies is, you know, and there's, yeah, there's the question of just what's gonna happen with Anthropic and OpenAI when they try to go public, 'cause they have to try, they have to go public at some point, 'cause they're just gonna run out of money if they don't. But also they might just run out of money anyway, if like Wall Street is not willing to sort of accept their promise for the future. But then it's also, what does it mean for these existing tech companies? And I find sort of the most poignant example to be Google, which they said basically their AI mode searches have really eaten into their search ad revenue. Yeah. It's not just that they're spending all the money on AI stuff, but the AI stuff has destroyed— well, it hasn't yet destroyed their core business, but it's not good for their core business. It started to destroy their core business. There's this thing of, on the one hand, what they made money in before is going away, and no one has a good idea of, then how are we going to make money with this other

Devindra Hardawar [01:09:02]:
And also by letting your other thing go away, by the way, search revenue is going down. That's because editorial is dying. That's because online publishing is dying and you are also helping to kill it. Uh, that also seems like a fundamental— you need something to feed that AI. What's, what's gonna do that when all the news sites have died?

Paris Marx [01:09:22]:
Not to mention, you know, nobody's making websites or contributing, uh, you know, kind of like information and stuff like that. Uh, You know, it's like for a few years now and, you know, this won't be new to you guys, but like to make Google useful even before the AI moment, it was often like trying to look specifically for stuff on Reddit to find real advice because SEO optimization had, you know, made so many, you know, kind of like reviews of products and, and just getting advice on like doing regular day-to-day things like made it completely useless in that sense. And now just for so many more things. these search engines, uh, you know, well, Google in particular, um, is, is kind of useless. And yeah, I don't know. It's— it— I think it— I think it's just like kind of disappointing as well because there were so many promises made for like, you know, what digital technology and what all this kind of stuff was going to be, you know, even if we go back really early to like what the internet has been— was going to be. And then we have seen like the slow erosion of it for a long time. But then, then it feels like the, the kind of scale of the erosion and plus like just the, you know, kind of worsening of the services that that we did have has really accelerated over the past few years, which I think is also part of the reason that people are really frustrated with the tech industry and the technologies that they use.

Paris Marx [01:10:39]:
Right. It's like Facebook is no longer— you're seeing your friends. It's like you're seeing a bunch of AI-generated stuff and a bunch of ads and a bunch of stuff meant to make you engaged. You know, you can't trust—

Devindra Hardawar [01:10:51]:
And angry.

Paris Marx [01:10:52]:
And angry. Absolutely. You know, you can't trust the products as much. that you see on Amazon and, you know, kind of how that service works. You know, and like the list just goes on and on and on. And that's not even to mention the growing concerns or, you know, kind of the list of concerns that people have with the generative AI products themselves, right? And how those do seem to continue to escalate. And of course, how this stuff has moved into physical space. And, you know, people are angry about that, not to mention how— Mm-hmm.

Paris Marx [01:11:20]:
the flock cameras, but, you know, the smart glass cameras. And it just feels really overwhelming, right? Um, and, you know, I do think that— I do think that for most people, it's more of just like a resistance, right? Like an anger at the way that things are. Um, and, uh, I, I hope that like, well, we're going to reach the point where that anger also gets somewhat redirected to like, how, how could things be better? You know, what, what do we actually want here, right? Yeah. Um, because again, I don't think it should be like like technology is bad and we should hate technology. I think it's like technology can be bad, but technology can also be really great, you know? And I think the problem right now is just the incentives are so misaligned with delivering technology that actually delivers benefits for regular people.

Devindra Hardawar [01:12:03]:
Shout out to Cory Doctorow, who feels like the prophet of our time right now because his certification just really covers everything. But he's just— it's banger after banger with that guy. And I've been reading him forever. So it's very, it's very great to see, like, his, um, thinking and theorizing about the internet is, is very much, like, coming to pass, and it feels very relevant right now. I think part of the response from the public is just that we know that we knew things were better. Things— I remember when this was better. I remember when the search was better. I remember when Facebook was at least, like, vaguely more useful.

Devindra Hardawar [01:12:36]:
Um, don't tell me this is an upgrade. Like, don't Don't piss on my feet and tell me this is an upgrade, right? So like, um, it feels like there's a lot of that going on. I'm also maybe a little spicy because, uh, have you guys heard of Google Books? The Google, the new Google laptops? No. You've seen these? These are, uh, these are essentially potentially the successor to the Chromebooks, except they're based around Gemini AI.

Paris Marx [01:13:08]:
Oh, great.

Devindra Hardawar [01:13:09]:
Do

Paris Marx [01:13:09]:
they have RAM in them or no?

Devindra Hardawar [01:13:11]:
They have the RAM, but they don't do anything. They don't do anything. Like Chromebooks at the very least were cheap computers that did the web, did the web, did web apps and stuff. Some of the Chromebooks got really expensive. Those were bad. I think those were always a bad idea, but essentially the affordable, the mid-range, the low-end Chromebooks, the ones they sent to schools unfortunately tend to suck, but Chromebooks enabled a certain amount of computing and opened the web to like a broader market. And I think there was a certain value there. Um, these Google Books cost the same as a Windows PC and they're based on Android and they do very little.

Devindra Hardawar [01:13:51]:
You can, you can do Chromebook stuff, like you can browse the web and stuff. That's nice. They have the Google Play Store in there and it's like functionally useless, right? You can, You can launch Android apps. Some Android apps have been optimized for desktop interfaces, but it's like, okay, great, great. Uh, I'm, I'm going to have a review of the Dell XPS Google Book up on Monday morning at Engadget. And it's one of the most angry reviews I've written in a very long time because like on your website, you have a $1,000 Windows 11 XPS with the same specs. And that's a full-ass computer. You can launch software.

Devindra Hardawar [01:14:27]:
You can do all sorts of stuff. You can customize that to your heart's content. You can use that like a computer, fulfill your computing potential with that machine, or you could spend $1,000 on a Google Book and do very little. And they're pretending that that is— this is a viable platform. It's a nice computer because it's a Dell XPS. In general, those tend to be really nice premium machines. And this is less an indictment of Dell and more of indictment of like, what are you doing, Google? Like, nobody wants to buy a computer just for Gemini shit. And I'm looking at this thing.

Devindra Hardawar [01:15:03]:
They have launch partners. There are several machines coming out here. Google's trying to push them. Um, I think maybe only the tech world has really heard about them, but I'm looking at this thing and being like, what you, what are you doing here? Like, this is not a successor to the Chromebook. This feels like more trash. This feels like another, it almost feels as bad as like the Microsoft Kin, if you remember the Microsoft Kin, which were, um, you know, semi-smartphones that came from like the Danger design team, the Sidekick people, as the iPhone was launching. You look at that thing, you're like, that's going to be dead in a week. Why did we even launch this? And I look at these Chrome— these Google Books, and I'm like, I don't know who this is for or who should pay this, but this feels like a crime that we even got this far.

Devindra Hardawar [01:15:45]:
And I don't know, it's just how it feels. I am angry at a lot of these companies right now. Look at these beautiful Google Books. They have keyboards, they have nice touchpads, they have nice designs. You can't do shit on them. And it's, it's

Paris Marx [01:15:59]:
maybe once they fail, someone will buy up a bunch for cheap and like stick Linux on them or something and then actually get some use out of it.

Devindra Hardawar [01:16:07]:
Maybe, maybe like, that's the thing. Like once they, yeah, once, once they're running like normal Intel processors and Snapdragon processors, so maybe you can jailbreak it. Maybe you can reflash. I don't know. I don't know if like that's going to be a thing that's possible. But more and more in the tech world, I feel like Frank Grimes in The Simpsons, in the iconic— that iconic Simpsons episode where like these companies are just Homer Simpson. Like, look, look at what you're doing. This is— you're not doing anything useful here.

Devindra Hardawar [01:16:34]:
And I feel like I'm going to go crazy and grab some live wires as I'm like ranting about what's going on.

Anthony Ha [01:16:40]:
Well, I think also it's important to remember sometimes examples like this, because there's also this narrative being sold of, like, if you are against these data centers, against most of the use cases for generative AI, you're framed as anti-tech. And also, when I start to get steam coming out of my ears, sometimes you'll ask me, well, what is it that you want to happen instead? Are you against progress? What does the future look like? And part of What is underlying that question is this sort of narrative of inevitability, right? If the tech industry has decided that this is what the future is, that therefore that is what the future is. And I think often people forget how many other futures the tech industry has tried to sell us. And certainly many of them have been successful and profoundly transformative, often in negative ways. But we were talking about crypto before. How many different VR headsets have we seen? Actually, I was in a conversation at one point where someone was saying, why are you betting against Meta and their AI agent? They've built the largest consumer ecosystem ever. And I'm like, their name is Meta. It's not an argument.

Anthony Ha [01:17:55]:
Their company name is a bet on a future that has yet to come to pass. And of course, I'm

Paris Marx [01:18:00]:
very excited that we're doing this episode in the metaverse.

Anthony Ha [01:18:04]:
Yeah, exactly. Well, yeah, none of us have legs, so there's that at least. And technological progress is not always linear or straightforward. Maybe some of these things will come back, but there are many, many futures that they've tried to sell us that people have rejected. I think with AI, certainly the level of investment, the level of government buy-in is going to make it hard for it to sort of just evaporate. And I'm not necessarily even saying that it should evaporate, but the idea that it has to be everything, that like the main way we interact with technology is going to be in the future, in the future is going to be AI agents. Like, I'm, yeah, I'm absolutely not convinced that's, that's going to come to pass.

Paris Marx [01:18:45]:
I agree. And I feel like there's also the way how like these futures and this idea of the future also forecloses other futures, right? Because we are putting so many resources behind this vision, there's so many other things that we can't do instead, or because they're the ones who set our idea of what technology is supposed to be, how it's supposed to work, how it's supposed to roll out, that it's like other people who do not have the same kind of breadth of wealth and power cannot present something, some kind of alternative, and have that taken seriously in a way that it will with them. But I really feel that this story about the Google Book also, and the way that you, uh, kind of laid it out, Devendra, also points to how, okay, yeah, we hear Sam Altman today talk about how he doesn't wanna see RAM in like personal computers and laptops 'cause that's wasteful 'cause it should be in data centers. And we're like, yeah, he doesn't want us to be running models locally because it, it makes more sense for his company for us to be using ChatGPT or something like that. Um, but I also feel like, you know, looking back at the Chromebook example also, you know, kind of talks about kind of the broader trajectory that some of these companies have been on where they have wanted us to be relying less on things that are done locally and more on their services that are performed in the cloud that we need to be always connected in order to access. And that has been like a longer-term push that we have seen for quite some time that, you know, generative AI is kind of— and in particular, you know, the kind of services offered by companies like OpenAI and Anthropic are kind of like connected to and continuing in a way that I feel like also shouldn't be, shouldn't be like forgotten, right? Because especially, you know, especially when you think about how generative AI and so many of these companies are so dependent on the cloud infrastructure that, you know, companies like Amazon, Microsoft, and Google built up over many years. And I would argue, you know, these AI models and these AI products in the way that they exist today would not be possible if it were not for the fact that, you know, these major companies started centralizing computing power You know, a couple decades ago now, you know, and all that computational power was necessary in order to realize, not to mention all the data that they were able to collect and stuff like that, to realize the vision of generative AI today that is so much so connected to like, we need to build as big of AI models as possible to achieve the AI god and all this kind of stuff, right?

Anthony Ha [01:21:10]:
Yeah, in terms of taking things out of my our computers. I will say that going back, the one thing I would really want in a really expensive MacBook is a Blu-ray player. Bring back disk drives and Blu-ray players in our computers. But that's such a good point also about, like, as you said before, also, like, these data centers, you know, they existed before, right? And, you know, this approach, these philosophies existed before. And that AI, again, my hope is that it's not not just saying AI is bad, although I think in most cases it is, but that it's saying this is also the ultimate expression of an approach— ultimate for now expression of an approach, of a philosophy, a business strategy that has a lot of problems and has done a lot of harm to the world.

Paris Marx [01:21:57]:
Yeah. One of the things that I— sorry if I'm cutting you off, Devendra, but one of the things that I always think about when we talk about an AI bubble or some kind of collapse coming or something like that is kind of It's not just that there has been an attempt to build out all these data centres in the past few years, but also Amazon, Microsoft, and Google, and to a certain degree, the colocation data centre companies, have a business model that relies on people using more computation over time. Because if not, that business is not growing. Generative AI was fantastic for that kind of business model because it means that the demand for computation, for additional computation, for additional cloud Right. —capacity is so great that you've had this massive expansion over the past few years. If you listen to Amazon, Microsoft, and Google, they'll say, yeah, we're spending a lot on capital spending right now to build out these data centers, but once this is built out in a couple of years, we're going to be making so much money from additional cloud customers and stuff like that because of all the additional computation that we'll have and that we'll be selling. But then the question for me becomes, okay, so if this vision, if this vision of generative AI does not work out, if it collapses, if there's not the degree of demand that was expected, then you have to imagine that these companies have some other kind of ideas for the next kind of very computationally intensive technology. Crypto kind of got us started.

Paris Marx [01:23:23]:
Crypto used a lot of computation. That collapsed. We moved on to generative AI. became much more formalized and kind of taken into the, the kind of mainstream tech industry in a way that crypto was not. And so when generative AI doesn't work out, um, which is not to say generative AI is going to fully go away, but I think that, you know, the way it is used is going to change, and the idea of how it should be used is going to change once there's a reassessment of the hype. But like, okay, what comes next? And what is like the new way that they are going to expect us to be using a load more computation? And I, I'm curious to see where that is kind of going.

Devindra Hardawar [01:23:56]:
It's definitely the question. Like, I don't So if this, if the AI economy crashes, um, what's gonna happen with all these hyperscale data centers? Mm-hmm. Like there's only so much cloud gaming somebody may end up doing. And I could see like Nvidia actually, I don't know, taking over some of those because they need that for GeForce Now. I could see some simple things like that, but that's not a mainstream thing. Not everybody's gonna be doing that. Um, shout out to the like younger generations who seem to be going back to old tech. Right? There is a resurgence of people who want digital music players just like to have as a separate thing from their phone to have local music libraries.

Devindra Hardawar [01:24:33]:
Young generations have always gone back to old tech. I was there. I was there when the millennials were like loving records and everything, and that's still a thing now. But it's also wild to see the same love for like, I don't know, the stuff that was new when we were younger as well. When I was a young millennial, So I think there is a bit of a fallback there, or like a bit of a step back that we'll be doing. Local compute. To me, the future is local compute and people doing weird things. Um, have you seen these, um, these little like personalized computers people have been making? What is the word for that? Are they—

Paris Marx [01:25:09]:
Oh, Cyberdeck. Cyberdecks.

Devindra Hardawar [01:25:12]:
Cyberdecks. Cyberdecks. Yes. And I see maybe it's just, uh, the way TikTok has hit me algorithmically or something, but like, I definitely see a lot of people, like, I put a Cyberdeck in a, in a, in a purse. I put a Cyberdeck in like a random bag, but a thing I can walk around with and it looks so hacky and weird and cyberpunk and not like, not clean, right? It's not an Apple machine. Yep. Built out of unibody aluminum. It is, um, you know, a mess of wires.

Devindra Hardawar [01:25:43]:
And I think there's a desire for a bit of that because at least you have control over that sort of machine. Hopefully we'll see more of that stuff. But anyway, we'll be back to all this. Let's have another word from Leo Laporte. Thanks so much, Leo. And guys, how about a bit of espionage? That's fun. That's fun, right? Did you see this story about Jensen Huang, CEO of NVIDIA, has been the target of a Chinese spy plot Specifically because he has made an acquaintance. Alice Huang, a Chinese citizen, has, has developed a close relationship with him.

Devindra Hardawar [01:26:19]:
And apparently there are some shenanigans going on there. This was kind of a big story coming out of The New York Times, too, because like 6 bylines, Jonathan Swan and Maggie Haberman together writing a story about Nvidia. Just wild. Just big. I don't have like any major takeaways from this other than like the This feels crazy, but also, shouldn't all of the executives at all the major US AI companies are probably targets of some sort of spy plot at this point? That's the only thing I can think of here.

Paris Marx [01:26:54]:
I would imagine so. Go ahead.

Anthony Ha [01:26:58]:
And as I recall from reading the story, that the caveat here is that this is not like— there's not like a definitive legal case that's been built against this person or anything like that. It's just that I guess a lot of people felt like they wanted to leak to the New York Times the fact that this person is highly suspected. Apparently Donald Trump was briefed about this a year ago and was like, yeah, it's fine, whatever.

Devindra Hardawar [01:27:18]:
Whatever.

Anthony Ha [01:27:19]:
Yeah, which is very interesting because he's so gung-ho about quote-unquote superintelligence, but I guess this is okay. I

Devindra Hardawar [01:27:27]:
don't know where that came from, by the way, like with that whole obsession. Who was the last person he talked to that put that in there.

Paris Marx [01:27:33]:
Yeah, we made a mistake calling it AI the whole show. We should have been saying SI.

Anthony Ha [01:27:38]:
I mean, it's similar to the, um, what was it, the data center thing, right? AI compute campuses. Like, oh yeah, if we change the name, people will like it now. They're that stupid. Um, but I've heard— yeah, I'm trying to remember the details, but as I recall, like, OpenAI has also, like, been sending out, um, you know, these, like, notices to its employees to be, like, very wary about this kind of thing. And I mean, I totally believe that that's happening. I do have a little bit of wariness because I do think sometimes when people talk about it, they really fall back on these kinds of, like, you know, frankly, like misogynistic and racist tropes about like, oh, like if a hot Chinese woman is like trying to talk to you, like, oh, if you're a 5 here, you're also a 5 there. And like, I mean, some of this stuff is kind of funny, but, you know, like I'm always a little nervous about how people talk about or uneasy about how people talk about it. But I think it's definitely happening.

Paris Marx [01:28:28]:
It's also like stands out to me that, you know, China has been very much trying to advance its like chip industry. So it's very much not a surprise to me that it would be trying to get close to Nvidia in particular if this is in fact the case, you know, to try to get some like, you know, inside intelligence or, you know, information that would be helpful to their ambitions. Right. So that they are not several years behind on chip technology and can catch up. You know, we know that they've been trying to um, you know, kind of, uh, get more information on how ASML's machines work to, to, you know, kind of advance their photolithography stuff. Yeah. And it just feels kind of like par for the course, really.

Devindra Hardawar [01:29:07]:
I believe the woman in question here is tied to a company that maybe, um, was bringing GPUs into China after the trade ban. So maybe that could be part of the intrigue there. I do wonder though, like, yeah, China, China does have to like catch up on chip stuff, but also You look at a lot of things happening there or on-the-ground coverage or people just visiting and being like, it does seem like a lot of things, at least tech-wise, um, and maybe, I don't know, civil engineering-wise, some things are figured out. So I don't know what the reverse is like. What is— is America also spying to see like, how do we— is there a secret to that high-speed rail or is it just—

Paris Marx [01:29:47]:
I wish they were spying about that. Please God, figure out how to build some trains. That poor California one is like never gonna get built, you know?

Devindra Hardawar [01:29:57]:
And again, we can talk about Elon Musk there because he was like instrumental in getting that killed. Oh my God, like I, I feel like that whole story— Paris, if you want like another thing, because I know also you've covered transportation stuff, but I feel like that still needs to be dug into a little more. And also the whole Hyperloop thing, because I sat there and watched Then build that stupid Hyperloop that goes like a block in Las Vegas. And it is again one of those things where I feel like Frank Grimes. This is stupid. This is not helping anybody. You've spent a lot of money and a lot of like time and effort to do what? To go a block in a Tesla in a tunnel with a— and it's still somebody driving it too.

Paris Marx [01:30:39]:
It's not even automated.

Devindra Hardawar [01:30:40]:
Yeah. It's so funny.

Paris Marx [01:30:44]:
Are the robotaxis going to take over now, though? Right. Because they're going to be so amazing. I'm sure it's really just wild to me. Right. Like, you know, there was all this talk about the Hyperloop and you go back to like Elon Musk talking about it in the early days. Right. You know, at the very beginning of the Hyperloop proposal. And he's very clear about the fact that he just does not want to see the high-speed train system get built.

Paris Marx [01:31:04]:
Yeah. And so he's presenting this as like, you know, the alternative to distract people away from the real project that would need to get done if we're actually serious about high-speed trains. And it's just disappointing how, like, you know, he consistently is able to get away with this kind of stuff. He did it with The Boring Company and public transit in a bunch of different cities as well. You know, how someone like him can move in and just, you know, basically make transportation worse for millions of people. And, you know, nothing sticks, right? He just Yeah,

Anthony Ha [01:31:36]:
I mean, it feels like this. On the one hand, I'm like, kind of like, oh yeah, if he wants to just like set a lot of investors' money on fire, like, that's, again, it's like, you know, good clean fun, why not? But there's always a societal cost to these things, because it often is tied to sort of a disinvestment in public transit. Like, you know, like you said, he's been a big advocate for not building the high-speed rail in California. And I think it also speaks to just, I think, the I think part of the theme of this episode is like this. I think for a lot of people, sometimes there's this feeling of like, boy, like, why are we always talking about politics? I just want to talk about tech. I just want to talk about the cool new gadgets. And I absolutely relate to that. But it's not our fault.

Anthony Ha [01:32:15]:
It's not like we don't want to be talking about—

Devindra Hardawar [01:32:18]:
This man is a big GOP donor right now. Like, he did that.

Anthony Ha [01:32:20]:
Right, right. We didn't do this.

Devindra Hardawar [01:32:22]:
Influencing politics. Yeah.

Paris Marx [01:32:24]:
I remember there was a New York Times story in like the mid to late 2010s that That was literally about how all of these right-wing groups, often Koch-funded groups, were using The Boring Company and the Hyperloop and the self-driving car stuff in campaigns against public transit, kind of like ballot measures that were seeking to direct more money or kind of set up new tax initiatives that were gonna send money to public transit projects, to basically go into these fights and say, public transit is outdated, it's the way of the past, we're gonna have self-driving cars, we're gonna have Hyperloop. loops, like whatever. So you should vote against this measure, right? And like, you know, even before Elon Musk was like kind of like openly right-wing in the way that he is now, it was like still very beneficial. What he was doing was still very beneficial to what they were trying to do, you know, in cities across the United States.

Devindra Hardawar [01:33:14]:
Have you guys seen the, the 4-hour-long Alex Gibney documentary Musk?

Paris Marx [01:33:18]:
Not yet. I do plan to see it. I actually looked to see if it was showing in Montreal. But it's not at the moment. I did see You Can See Everything. So

Anthony Ha [01:33:28]:
I've been using You Can See Everything twice, and I've seen the Musk documentary zero times, but I do want to watch it. I will watch it probably.

Devindra Hardawar [01:33:35]:
It's a great at-home documentary is the thing.

Anthony Ha [01:33:38]:
Well, I mean, this is also—

Devindra Hardawar [01:33:39]:
yeah,

Anthony Ha [01:33:39]:
there's a

Paris Marx [01:33:40]:
big question right up front.

Anthony Ha [01:33:42]:
Will David Ellison allow HBO Max to stream the Musk documentary?

Devindra Hardawar [01:33:46]:
That's true. There's that also. Maybe VOD. So like maybe you could at least rent it. But, um, yeah, I haven't seen Social

Paris Marx [01:33:52]:
Reckoning. Have you seen that one?

Devindra Hardawar [01:33:54]:
Yeah, yeah, it's

Anthony Ha [01:33:55]:
fun. It's, it's corny, it's swirking. It's corny. It's fine.

Devindra Hardawar [01:33:59]:
Yeah, it doesn't accomplish anything, but it is more of a journalism movie than you'd expect. And he does have this love for old-school reporting, and I kind of— the reporter side of me at least liked that part, even though I know nothing happened after that. But the, the Gibney Musk documentary, it's so long and so involved and it doesn't actually do anything new. But I also think it's kind of essential viewing for everybody. So that makes it a great home, home rental, home streaming thing because you can break up those 4 hours and watch it in chunks and watch it like in episodes. It's not a great theatrical experience because it's not like a really engaging— it's not— you can see everything, right? Where I feel like every frame of that is so interesting. So anyway, just want to shout that out. Not much new in the Musk doc.

Paris Marx [01:34:43]:
I feel like I'm a bit nervous about it with like some of the contributors and some of the people who he has, you know, again, like not having seen it yet and just being like very familiar with the Musk story and, you know, hearing some details and asking some big questions as to like, that seems like an odd decision.

Devindra Hardawar [01:34:59]:
Yeah.

Paris Marx [01:35:00]:
But yeah, I do plan to see it and I hope to see it soon. I just haven't had the opportunity yet.

Devindra Hardawar [01:35:05]:
I think for normal people who are not us, who are not like in the tech news world all the time, it is a good encapsulation of how much he's sucks. And, uh, you know, so that's a big thing. And Elon Musk, who is a Nazi, uh, it does not shy away from the fact that he, you know, made several Nazi salutes during Trump's second inauguration. It does not shy away from his tendency towards white supremacy and everything. And I think in that respect, it is, it is at least like saying, hey, pay attention, like, this guy, the richest guy on Earth, is kind of a problem. Um, so it's worth watching. Just for that.

Paris Marx [01:35:39]:
Can I ask, again, having not seen it, do you feel like it's the kind of movie that preaches to the choir in a sense? Is it like a very kind of liberal kind of movie that is going to speak to liberals who are already concerned about Musk? Or do you feel like it can expand beyond that kind of realm and actually bring in people who might still have some good vibes or good feelings toward Elon Musk?

Devindra Hardawar [01:36:07]:
It's, it's tough to say. I mean, literally just saying the facts, it is a very matter-of-fact movie and it does not like lean on anything. But if you say the facts out loud, it looks pretty bad for him. If you show the clip of him making a Nazi salute several times, really fervently, um, really banging his chest for that one, it does not look good for him. But

Paris Marx [01:36:28]:
I don't like to throw your heart out to people, Devendra.

Devindra Hardawar [01:36:30]:
Oh man, that's what it was, right? It was It was just— it was an ancient Roman salute. It was like a symbol of respect, right?

Anthony Ha [01:36:37]:
I love to show my respect and devotion to white people.

Devindra Hardawar [01:36:41]:
Yeah. Yep. Um, I think this thing shouldn't have been 4 hours long, is the thing. Like, if you want to welcome everybody, don't be like, hey, let's sit down and watch 4 hours of the Elon Musk documentary, because that feels a little bit like preaching to the choir. But if this was like a streaming event where it's like part 1, Through 4, and you can sit and take it in. Um, Gibney is, is kind of an interesting figure. He also did a great, um, documentary about Elizabeth Holmes as well. Again, same thing, not like unearthing new stuff, but it does encapsulate the story very well.

Devindra Hardawar [01:37:16]:
I believe it's called Blood in the Machine.

Anthony Ha [01:37:19]:
I feel— no, no, that's, that's the, the Steve Jobs documentary that you and I, uh, saw. No, it's called The Inventor.

Devindra Hardawar [01:37:25]:
The Inventor. The Inventor is that one. Yeah. And, um, yeah, I've, I've talked to him a couple of times alongside Anthony back in New York. Uh, Alex Gibney has like a farm of research, like a whole floor of a building where it's just like assistants, like pulling facts and doing stuff for him to build his next movie. So like his stuff is always filled with facts, but they're not usually that visually interesting or that, you know, That inventive, I guess. So I'd love to see more of that, but it's worth watching. That's all I'm gonna say.

Devindra Hardawar [01:37:55]:
And especially for Musk. I also want to ask, I took my first complete self-driving car ride when I was in San Francisco for the Microsoft and NVIDIA event. I took a Waymo and it felt really weird, but I want to know, have you guys taken one yet or any pure self-driving car?

Paris Marx [01:38:11]:
Never have.

Anthony Ha [01:38:12]:
Okay. I've taken one, I think once or twice. twice in San Francisco. And yeah, I mean, I think that like, it was a— it was an impressive experience. It was like weird, but I—

Devindra Hardawar [01:38:26]:
It's weird. It's very weird. Keyword. Yeah.

Anthony Ha [01:38:30]:
It's, you know, it is one of those things where I think I'm not opposed to self-driving technology. And I think there are good reasons to that, why over time it may make sense to shift more and more of our rides to self-driving. It's more just like, well, okay, then, like, but like, A, let's not disinvest in public transit. I don't think of it as a replacement for public transit at all. And also, you know, like, let's actually, like, be thoughtful about how we roll this out. Like, what is the sort of regulatory framework around it so that, like, say, if a Waymo hits somebody, which has happened, that actually, like, who is, who is accountable? Like, you know, like, what are the— it can't just be this sort of thing where, like, no one is making decisions, it's just the hand of, of the algorithm, and, and we can't do anything about it.

Devindra Hardawar [01:39:19]:
Yeah, yeah, yeah.

Paris Marx [01:39:22]:
No, I'm in San Francisco next weekend, and I might try one while I'm there. I have meant to try one. Yeah. And I just haven't, like, haven't gotten around to it yet, I guess. Um, it— one of the things I feel like that I don't know, really surprised me about the self-driving story or something recently was to see that GM is planning to get back into self-driving again after it pulled out of its Cruise division. Was that— what was that, last year, the year before? Because it hit somebody and it was like, oh, you guys are doing this again. Like, yeah, I was a bit shocked on that one, to be quite honest.

Devindra Hardawar [01:39:54]:
What I think my main takeaway is that, first of all, it is a weird experience to do it yourself. Because, yep, it's all automated. It comes to you. It does the whole thing. I'm sure Twit listeners have heard this and have been trying this and doing this for years. I was surprised by how well it worked and seemed to take a good route, got me to where I was going. But the thing is, like that day, those couple days when I was in San Francisco, I saw the Waymos everywhere and it seemed like a really interesting way to just never interact with a working-class person, right? Like, literally, it is literally a bubble that is taking you from one spot to the other. And in some cases, like, yeah, you would interact with a cab driver, right? If you're getting an Uber or Lyft or something, you would interact with somebody who is doing a job.

Devindra Hardawar [01:40:38]:
I've had some great conversations with cabbies in San Francisco because often they are locals, like they are people who live there and who grew up there and are very pissed off about the state of the tech world. And I feel like at least Waymo as it is and these self-driving driving taxis as they are, it does feel like a great way to just like completely isolate yourself from hearing— again, like the tech, the AI people— hearing anybody who says they don't agree with you, right? I, it almost feels like it exists for part of that purpose, especially in San Francisco, a city that has extreme poverty alongside being the heart of Silicon Valley. Like every time I visit, I am still astounded by like just how rough is in San Francisco sometimes and how it seems like, you guys, you guys can help more, can't you? Can't you do more? I don't know. Anthony, you used to live there, so I'm sure you have like more of a sense of like what's going on. But also coming from New York and other cities, like where it feels like at least there's more of an attempt to help people who are on the streets and people who need help. I don't know. It's just a weird random thought I had. I don't think self-driving is inherently evil because like in some places it can be useful for personal vehicles too.

Devindra Hardawar [01:41:50]:
You know, I have a car, I have a 2019 Volvo that has like, it's not self-driving, but it can like stay in the lane and do a lot of the work and you hold on to the steering wheel. And that has made road trips incredibly great and very painless and very easy to like deal with. And as somebody who's like anxious about driving, it's made my life a lot easier. But I also see the world where it's like, oh, that makes people drive more too, right? Like, that's that's worse for the environment. So there's all these push and pulls, but yeah, just dropping that thought there about self-driving. And I'd love to know, Twit listeners, what do you guys think if you've been taking Waymos? I was just down in, I took my wife to a show in Atlanta. I live right outside Atlanta and you could see several Waymos like going around as well. So it feels like they are broadly getting to more places.

Devindra Hardawar [01:42:36]:
And I've heard from my reporter friends in Gadget, especially like women really like this idea Because you're taking an early morning flight taxi or something, a taxi for a flight. You don't have to deal with anybody. You don't have to deal with any creeps, any weirdos. And I totally get that aspect of it too. It definitely struck me oddly this time in San Francisco. So yeah. I think related to a lot of what we're talking about too, last week was the major like PHP exploit and everything. kind of brought the internet down.

Devindra Hardawar [01:43:11]:
There have been a lot of great conversations about how the internet may essentially need to be rebuilt from scratch. I listened to Charlie Warzel's podcast over at The Atlantic. What is it? That's Galaxy Brain, right? Yeah. Like his work in general. The Atlantic, I don't know what's going on there politically for some of their reporting, but I like Charlie Warzel and it does seem like AI is enabling a whole new thing. which is also AI-driven hacks. And it seems bad. It seems like, oh man, this thing that is supposed to be our major productivity tool is also enabling a whole new, I don't know, whole new vulnerabilities and whole new ways of making things worse.

Devindra Hardawar [01:43:50]:
Have you guys thought, have you guys seen this podcast or listened to this podcast? Or what are your thoughts about like where we are with the whole PHP hack? Okay.

Anthony Ha [01:43:57]:
I haven't listened to the podcast, but I mean, I've definitely like, yeah, read a lot of stuff about the PHP hack. I think Charlie Warzel is great. And also, you know, as the weekend editor of TechCrunch, can definitely relate to. People having strong opinions about your publication and being like, that's totally fair. Yeah. I share some of those criticisms. Ideally, judge me primarily based on my work, not on sort of your preconceptions about the publication as a whole. I think that, I mean, totally, yes.

Anthony Ha [01:44:23]:
I think like, I mean, it was really shocking to me that how I felt like a lot of security researchers who have genuinely been, generally been trying to keep keep the temperature low around AI stuff, we're kind of like, oh, shit, this is really bad.

Devindra Hardawar [01:44:35]:
This is bad.

Anthony Ha [01:44:38]:
I think one thing that surprised me that it took me several days until I realized this was that the actual PHP exploit was not discovered by AI. It was discovered by human researchers who were taking part in an OpenAI program. And then they did use AI, I think, a little bit to sort of explore some of the exploits and things like that. But it is not the case that AI just found this vulnerability on its own. Do I think it'll probably find more vulnerabilities in the future? Sure, of course. And yeah, I'm curious— I mean, I'm sure that probably the headline is partly just to be provocative because, unfortunately, I don't think we generally rebuild things from scratch. Yeah. We sort of just kind of build on top of it and do the best we can to fix the vulnerabilities that are there.

Anthony Ha [01:45:22]:
Yeah, if I were somebody who was in charge of a website right now, I would be very stressed out about both this specific vulnerability, but just what the future holds.

Paris Marx [01:45:32]:
Yeah, I feel like we're probably going to see more vulnerabilities over time, especially as more vibe-coded code ends up being in more and more of these products. I think I would say a bit beside the PHP hack point, what some of the people in my kind of community, I guess, have been saying the past couple months, you know, as we've seen more and more of these stories about, you know, the AI hacking and how that's so often framed as like, oh, the AI is so powerful and what have you. But, you know, people who are more skeptical of AI who I speak to are kind of like, in the past, wouldn't we have just called this kind of like malware? But now it gets reframed as like, oh, the AI is so powerful, not like the companies are, pushing out malware on a scale that is kind of unbelievable. It's like super malware. Yeah, exactly. But they don't get to treat it that way because they get to frame it as something else, right?

Anthony Ha [01:46:25]:
Yeah, I think the difference a lot of times is just essentially the scale that you can get because you can automate so much of it. But it's not like smarter. It's definitely— I don't think anyone— at least the people who actually know what they're talking about don't seem to be saying it's smarter in any way. It's just more— I even hesitate hesitate to use this word because it's kind of anthropomorphising, but it's essentially more persistent. It can do more, it can reach more broadly, but it certainly is not evidence of a sentient hacker or anything like that.

Devindra Hardawar [01:46:50]:
Not sentient. I do find it can do the busy work that unveils exploits, basically. Yeah.

Paris Marx [01:46:58]:
I do find the conversation about— I've seen this in Wired this week as well, about the end of the internet or rebuilding the internet or whatnot, really interesting, right? Because it also brings me back to when we would talk about like dead internet theory, um, you know, in recent years. Mm-hmm. And it was kind of like a joke, right? You know, like considering what degree of the interaction online was actually real or not. Um, and now we're getting to the point where there's just so many kind of like AI bots out there that, you know, actually a lot of the visitors to websites, uh, now, uh, in some cases it can be a majority, will actually be, you know, bots, right? Um, and, and And what does that mean for how the internet works, how websites are designed, who they are tailored for, or what they are tailored for, maybe more specifically? And I think it also picks up on what we were talking about earlier, right, about just the general frustration that I think a lot of people have about digital technology, about the platforms that we rely on, like all that kind of stuff. And it does feel like there is a growing desire, as you were talking about, Devendra, with kind of local computing and that kind of stuff. stuff, like to actually do something different. And how do we build this in a different way that does kind of bring back or recreate or create new things that are actually useful to people in a way that it feels like the current set of technologies do not deliver on, you know, if they did at all, or maybe they don't any longer. But yeah, I think it's, I think it's a productive conversation, to be quite honest, because I would hope that it starts to open us to Like, you know, how can things work differently if it's not created by these major tech companies? And also, like, one of the things that really stands out to me, which I feel like is less the case in the United States right now, but like, you know, in Canada where I am, in Europe and other parts of the world, there's a lot of discussion about getting off of US tech, right? And reducing the dependence on US technology because of the perceived and I think real vulnerability when you see how the tech companies are using their power, but also the US government is less of a reliable partner than it was in, in the past.

Paris Marx [01:49:00]:
And then I feel like that question becomes not just, you know, it can be perceived as how do we build our own versions of what we would get from Microsoft or Google or whatnot, but then there's another kind of direction to take where it's not just how do we build our own Silicon Valley, but how do we start to reconceive some of these technologies that we've become dependent on and actually build them in a way that is, you know, better oriented toward the values that we hold, you know, actually delivers more benefit for our societies. Like, you know, and I think that there's a tension right now if you look outside the United States as to which direction to go. But I think it's at least a productive conversation to, to be having, right? And hopefully leads us in a better direction.

Devindra Hardawar [01:49:41]:
I think the idea of rebuilding the internet just struck me on a personal level too, right? Like, I've been online for a very long time. And I remember when the internet was good, and I remember like what was good about it and what was like so beautiful about it too back in the day. And like, there were issues in the '90s and certainly in the 2000s, but between social media and between like all the tech companies essentially taking like taking over portions of the internet, Facebook taking over so much of it, Facebook essentially owning large parts of the internet. I do feel like there's a need to like see like what that world is without it. So maybe that's another thing. Maybe that's another thing the kids can help us get back to. Um, the days of personal websites and webrings and all that fun stuff. I was there.

Devindra Hardawar [01:50:30]:
I was there in the '90s with my anime websites and I'd be there again to show my kids how to make anime websites.

Paris Marx [01:50:37]:
Hey, I used to like making websites about the video games that I enjoyed playing and I used to make video— I used to make websites about my dog. Like, yeah, it was good fun.

Devindra Hardawar [01:50:46]:
The internet.

Anthony Ha [01:50:47]:
Yeah, I have a whole list of like failed personal blog attempts. I think one on LiveJournal, one on Blogspot, one on Weebly. Um, yeah, maybe, maybe its time has come again.

Devindra Hardawar [01:50:58]:
It's time. Well, now it's like it's all video, right? Now it's all— if you want to do this, you have to do it in short-form video on TikTok or something. But maybe we can make that I don't know, desiloed from all these companies. I would love to see it. Anyway, folks, we've got a bunch more stories to talk about. We'll be right— we'll be back right after this. Anthony, I saw you wrote up the, uh, 7 minutes of Grok's AI version of The Odyssey. This is the supposedly historically accurate version of The Odyssey that Elon Musk wanted to, to put into the world.

Devindra Hardawar [01:51:33]:
I've seen bits of it. It looks better than like the, the early clips did, but it also is still so soulless and still so like— I don't know who would watch more than a handful of clips about this thing. What are your impressions as somebody who also enjoys, you know, movies?

Anthony Ha [01:51:52]:
Uh, it sucks. It's, it's quite bad. Um, so, you know, part of the context here is, is that Elon Musk kind of had a huge tantrum about Christopher Nolan's version of The Odyssey. Mm-hmm. And I think was primarily upset about the fact that there were non-white actors and trans actors cast in the film. And he went on this whole tear about how it was just designed to win the movie some Oscars, which I think that might pay off and it might be winning a bunch of Oscars in a few months. So he said that Grok will create an entirely historically accurate version of The Odyssey by the end of the year. I mean, of course, like, on its face, that is like just such a ridiculous statement because—

Devindra Hardawar [01:52:38]:
This is the richest man on Earth having like a diaper episode about a movie.

Anthony Ha [01:52:44]:
Yes, which— and then the thing that's crazy is that's not even that unusual for him. He does that all the time.

Devindra Hardawar [01:52:49]:
He doesn't appear to work. He just sits on X all day and whines about shit.

Paris Marx [01:52:54]:
No, he works so hard. He sleeps on the factory floor.

Devindra Hardawar [01:52:57]:
Like,

Anthony Ha [01:52:59]:
didn't he— his office at the White House, he was just basically playing— I forget what the video game was, but there was some video game he was just playing all the time.

Paris Marx [01:53:07]:
But when he wasn't paying someone else to play it for him, right?

Anthony Ha [01:53:11]:
Exactly. Just to state the obvious, the Odyssey is not— didn't really happen. So you can't really make a historically accurate version.

Paris Marx [01:53:19]:
This is a bold claim for you to make. Yeah,

Anthony Ha [01:53:21]:
they spoke in ancient Greek. So, you know, yeah, this task on its face is just like laughable. So they put this out. It's not a full movie. It's 7 minutes of what's supposed to be now the Odyssey TV series. And like you, Devendra, I did not— I was expecting it to just be immediately obviously laughable. And I will admit it actually looked pretty good. And if you put it—

Devindra Hardawar [01:53:47]:
Video generation has gotten better, like, since he last previewed it. Yeah,

Anthony Ha [01:53:50]:
yeah. And, and they clearly spent a lot of time on it because I think they knew probably that there'd be people looking at it, like, getting ready to make fun of it. So it actually doesn't look as terrible as I thought. But then you just sit there and, uh, and of course, because most of us have watched an incredible version of The Odyssey just a few months ago, You sit there and you're just like, this is not well written. This is not interesting. It's basically just, if you've played a video game or watched a mid-Hollywood blockbuster from like 20 years ago, that sort of part where it's just all the exposition at the beginning, the crawl of lore of like all these names that just kind of goes on and on and on, and that you probably are like, yeah, yeah, yeah, let me get on to the actual story. That's what the segment is. Mm-hmm.

Anthony Ha [01:54:33]:
There's nothing interesting about it. There's nothing that grabbed my attention. I think I put in my article about it, when I watched the nearly 3-hour version of The Odyssey, I was never bored. I was always absorbed. I mean, I think that movie has problems, but I was never bored. Making it through the 7 minutes of this clip actually took a lot more effort than watching the 3 hours of the Nolan Odyssey.

Paris Marx [01:54:54]:
Yeah, maybe his movie efforts are gonna get a boost now that he's put a bunch of money into the new Paramount Paramount Skydance Warner Brothers.

Devindra Hardawar [01:55:04]:
That makes that even worse to me. Yeah. But, uh, nobody remembers the 2010 Clash of the Titans directed by Louis Leterrier.

Anthony Ha [01:55:12]:
I rewatched that this year.

Devindra Hardawar [01:55:13]:
Yeah. And starring Sam Worthington. You know, that's a bad movie that had a terrible 3D conversion, but there are parts of it, at least the casting. I was like, okay, this is, this is kind of baller. Like, this is at least cool. Liam Neeson as Zeus, Ralph Fiennes as Hades. Yes. Give me that.

Devindra Hardawar [01:55:28]:
And even a bad Hollywood version of, you know, of Greek epics, um, is, is still so much more watchable than whatever the hell this Grok thing is.

Paris Marx [01:55:40]:
Uh, it's kind of funny to me that Sam Worthington like made bank doing Avatar, like really lucked into that, and then like never, never did anything good or never had to do anything else that would have been good. Like besides—

Devindra Hardawar [01:55:51]:
He was busy for many years doing Avatar. Like they were locked in for a long time shooting those movies, and it's just kind of wild. hell to see.

Anthony Ha [01:56:00]:
Yes. And to be fair to the tech industry, I do think that is an example of, uh, it's not just tech executives who try to make things happen that the public just is not super interested in. And like, you know, Sam Worthington is— I think he's like a perfectly decent actor, but there was that period where you're just like, oh yeah, like they seem to think that was— he was the next big thing. And it didn't quite get there.

Devindra Hardawar [01:56:18]:
He's the one guy we got. But also that was an example of, again, Hollywood, uh, forcing technology on us, like upconverted did 3D conversions, which died off. It was gonna be the future, but they mostly died off. Like, there are still 3D releases now. Mostly though, they're like CG movies and kids movies where you can do a better job of creating real 3D rather than like upconverting from the camera.

Anthony Ha [01:56:41]:
I weirdly have one more thought about Clash of the Titans. I can't believe we're getting so deep on this. Talk about Clash of the Titans. It's also another classic example of now, like, a lot of movies from that era, right, we think of them as being so CG-driven, but actually they still had an incredible amount of practical effects. And now watching it, you know, whatever, 15, 16 years later, it's all the practical stuff that still looks good, that holds up well. And then the few moments that are basically pure CGI, you're like, yeah, this, this has not aged well at all.

Devindra Hardawar [01:57:07]:
That's— yeah, well, I just remember that being one of the worst, like, press screenings I ever attended for, like, a big-budget Hollywood movie, because, like, you could tell the 3D was garbage and they just didn't care. And that's the sort of thing you can tell the AI is garbage sometimes, and the companies producing the AI don't care because it's the next big thing and that'll be where we're all headed, right? Like, let's see, that's the ultimate thing there. Uh, other news this week that kind of annoyed me, there's a story about DoorDash acquiring Instacart for $17 billion because we don't have effective regulation in our country. Okay. That's all. That's all I have to say. Like, I guess we're just not going to— Lina Khan's FTC would never allow this. But right now we have no rules.

Devindra Hardawar [01:57:53]:
Nothing matters anymore.

Paris Marx [01:57:56]:
I wonder if there'll be any, like, pushback in other markets, you know, where I'm assuming it will have to be approved as well, like in Europe. And, you know, we often see kind of pushback in the UK to these kinds of things. But yeah, it's, it's not great, especially where, you know, I feel like there has been been more conversation. I don't know if this has been happening in the United States, but we've been having it more in Canada, um, with like the kind of surveillance pricing, um, and, you know, how things are often more expensive on like Instacart and stuff than if you would buy it in store, and how those prices can change based on who you are. Um, yeah, there's been growing concern up here about that, and, and I think it's been, been echoed down, uh, in your neck of the woods as well. Yeah. Um, but yeah, I don't know, I— you, you have to imagine that there even more pressure to do something like that if there's this big merger and now there's a ton of debt that needs to be paid as a result, as often is the case in these kinds of things, you know?

Devindra Hardawar [01:58:48]:
Yeah, these mergers make things worse, like, especially when they're so big and when they're essentially doing the same thing. In the US, we also saw the Grubhub and Seamless, like, union a while back too. And I feel like those were two smaller— they were two small companies that ended up becoming like a medium company, right?

Anthony Ha [01:59:06]:
Yes. Yeah, that's— I mean, I think Seamless was, was much smaller, and was essentially just a New York City company. Um, so I don't know that that was quite it. But, but yeah, I mean, I think that all this, that it's led to this point now where there's just a handful of companies in this industry, and they're, you know, continuing to consolidate. As I read some of the reporting around it, Instacart's logic was sort of like, this is as big and successful as we're ever going to be. So let's sell for as much money as we can now, because if we try to do this 2 years from now, it's going— we're not going to be as successful doing it. Which is, you know, like, you understand their logic. And it's also kind of— it's also sad.

Anthony Ha [01:59:44]:
Not that I necessarily think that Instacart is that great either. I think— and, you know, this is an area where I'm just a simple, you know, tech journalist. So maybe this is naive. But the other thing that just makes me crazy about these deals that Paris touched was also like how a lot of the times the way they're financed is basically there's just a ton of debt that then the, the new company has to pay off. It's like you're not actually spending any money.

Devindra Hardawar [02:00:07]:
That's how we do things.

Anthony Ha [02:00:08]:
Which is crazy that like, you know, Elon Musk went back, you know, 5 mergers ago when he first acquired X. It was just like, well, we're just going to like load up all this debt and then suddenly now I own X and then I, Elon Musk, don't have a lot of debt. Like X has a lot of debt, you know? the combined Paramount WarnerMedia is also— Warner Brothers is going to, you know, has or has an incredible amount of debt. And yeah, I assume that's probably gonna be the case here too.

Devindra Hardawar [02:00:35]:
Fun times. And these things, as they also join, they become worse. Prices could get worse as they have to, like, you know, chase after this debt and kind of resolve it too. So that's a damn shame because I use meal delivery when I have to. Sometimes it's like I I gotta feed the kids. If it was just me, I can eat what's in the closet. If it was just me and my wife, we can like survive. It's like, I gotta feed these children.

Devindra Hardawar [02:01:00]:
I've been working all day. Um, so it has come in handy for stuff like that. But also, you know what's good? Uh, call your local restaurants folks. Like actually a lot of them do offer delivery or they use like other delivery services that take less money than DoorDash and Instacart and all these others. So there are like, I think Beyond Menu is one of them that takes like very, very little, uh, of a, of a cover, of a charge from them. So that stuff is great. Um, yeah, I feel like, have you get— did you guys see the Meta humanoid robot story? They built a humanoid robot that they're not gonna sell, but they're doing it. And they sent one, they sent their like head of robotics to China for like 2 months to create this thing.

Paris Marx [02:01:44]:
Oh yeah. Yeah.

Devindra Hardawar [02:01:46]:
Okay. Meta coming to your home as a robot.

Anthony Ha [02:01:49]:
Well, and it was specifically—

Paris Marx [02:01:50]:
Just what everyone wants.

Anthony Ha [02:01:52]:
I'm totally drawing a blank right now on the name of their— the AI agent that they announced a couple weeks ago, but— Muse? Muse. Yeah, it could be like the physical manifestation of Muse, which—

Devindra Hardawar [02:02:02]:
Gotta love it.

Anthony Ha [02:02:04]:
I find crazy. I, I will admit that a lot more people have been trying to like hype up Muse to me than I expected. I assumed it was just gonna be dead in the water. But it actually seems like there are people, uh, including people I think would know better, who are like, yeah, a Facebook, uh, AI agent, that sounds great. I'm totally—

Devindra Hardawar [02:02:22]:
These people need to watch The Social Reckoning. Like, listen, we went through all of this. Nothing happened to them. Very little happened to them. But they, they did many bad things, and they were very aware of what their social network was doing. And now you're going to give them all your personal, even more personal data and access to financial things like to credit cards and bank accounts. That's awesome.

Paris Marx [02:02:44]:
I don't know if it's happening in, in the United States, but like a bunch of banks in Canada are like basically blocking AI agents from even being able to access their websites. Yeah. Because they're worried about, you know, I'm sure like security and things like that.

Anthony Ha [02:03:00]:
Yeah, I think that's happening in the US. And also with some of the major— oh, I think it was like airlines. Some of the airlines are also blocking them as well. And I don't necessarily expect that to continue to be the case indefinitely, but I think it's a totally reasonable thing to do for now and to sort of say, hey, what are the safeguards you have in place to make sure that you don't buy a bunch of plane tickets, which then we have to refund a few days from now when we realize it was a complete mistake?

Devindra Hardawar [02:03:28]:
I feel I feel like the people leaning on agents so much are a little delusional, but also this wide adoption of Muse does make me wonder if people even listen to all the warnings we're sending out there about this company and maybe don't trust this company and everything. On the other hand though, like, would you ever want a robot in your home? And I kind of would because I hate spending an hour every day cleaning my kitchen. I hate that. And I could see like some, some use for some sort of humanoid robot in your home. That will involve AI, that will involve a lot of personal data, that will involve surveillance. What do you think, Paris?

Paris Marx [02:04:04]:
I feel like I'm— I feel like I'm— I'm still skeptical of, like, the humanoid robot, right? Like, the notion that this is the form factor it needs to take to be useful. I think it's much more likely that we have more kind of, like, robots that are tailored to specific tasks. Which we have now.

Devindra Hardawar [02:04:20]:
Absolutely. Roomba is a robot. And that's what I was going to say, right?

Paris Marx [02:04:23]:
Like, yeah, I have a robot vacuum. And I got to be honest, I love it. And it keeps my floors clean and it's fantastic. But would I want a humanoid robot in my home? That I would feel a much greater degree of resistance to, I think. You know, but if you told me in 5 years that there was some kind of robot that, like, cleaned your countertops and, you know, did so efficiently and did a really good job or something, I would be like, like intrigued maybe, but I also don't see how that robot isn't like really large because of like the height aspect of it. Um, so I, yeah, I feel like I'm skeptical on that front of like how that would ever work out in a meaningful way. Um, but yeah, uh, the ones that clean my floor, they got a thumbs up from me. This

Devindra Hardawar [02:05:09]:
is the next frontier. Like, this is where they're all headed. Nvidia, every time I go to CES, like, they're pitching, uh, AI in world. And when you talk about AI in the real world, you mean robots, right? And they're all— that's kind of where they want this stuff to go, despite all the movies we've seen warning us about AI-powered robots in reality. I think there is something there, but also, how do we do this in a way that does not destroy us? You know, like, are we taking the wrong lessons away from WALL-E with this hope and this dream for robots serving us?

Anthony Ha [02:05:42]:
Yeah, I mean, you know, Paris, you were talking about this fascination with humanoid robots, which again, I can imagine being useful in some contexts, like say on like a factory floor or something. But the fascination with it—

Paris Marx [02:05:53]:
Even then though, I feel like the humanoid-like form factor isn't a good fit for the factory floor, right?

Anthony Ha [02:05:58]:
Yeah, you're probably right. Yeah, I'd actually— maybe there is just no good use case. But it seems like this part of this confluence of, on the one hand, the financial incentives, which is that we can't present as sort of the future of physical AI company unless it's a completely general-purpose robot, which partially pushes you towards the kind of humanoid form factor. And also, yeah, a lot of misread and miswatched science fiction. When they're like, okay, how can we make this a trillion-dollar company? Let me watch some dystopian movie and then pretend that was good.

Paris Marx [02:06:33]:
I also feel like it speaks to how like the generative AI moment has given a lot of energy or a lot of kind of like power or like, you know, just directed money into anything that can be like associated with AI. And this is like the physical manifestation to it, right? Like, even if I think that there might be a role for more robotics into the future, I'm skeptical that the humanoid robot is the form that it takes. but that gets the money because it's like, as you're saying, Anthony, this can do everything or what have you. And, you know, maybe some of that work will in the future lead to some more kind of robots that are made for specific tasks rather than ones that can do everything. But I also feel like, you know, obviously we've seen all the excitement around humanoid robots, and I don't know, I hear China's doing that better than, than the States. But, you know, I don't have a lot of kind of deep insight into that. But I also feel like, you know, we were talking about self-driving cars. I also feel like that's part of the reason that we've seen kind of like the revival of the self-driving car narrative the past couple years, because again, it's like the physical manifestation of AI in certain ways.

Paris Marx [02:07:44]:
And we had all the hype around self-driving cars in the 2010s, which is the same moment that we had the narratives around how like new AI and new robotics were going to destroy all these jobs. And, you know, That was kind of like one element of it. And I feel that's part of the reason that Waymo has gotten more attention in the past few years. Obviously, why Tesla has been trying to move into robotaxis, why Uber has gotten back into self-driving cars to a certain degree, why GM is back into it again, is just because there's this money and this kind of excitement around something that can be a manifestation of AI. And this is another one of those. Right.

Anthony Ha [02:08:23]:
Right. And when we talk about excitement around AI, what we're also talking about, to a large extent, is about the excitement on Wall Street about replacing human beings and firing people. And I think that also speaks to maybe part of the excitement around humanoid robots. And I think that then tying it back to sort of why there's a bit of a reckoning moment is that, A, I think that's disgusting. Even if you could replace a human being, like, you know, for something— a chatbot would be cheaper to write a TechCrunch article. For very self-interested reasons, I think that's very bad. But we're also at this moment where people are sort of saying, but is it actually cheaper? Because it's not. It just seems cheaper because it's been subsidized by all this money.

Anthony Ha [02:09:04]:
But what if you fire everyone and then it turns out to be more expensive to do everything? Which is a future that, you know, know, is one of the possible futures that they're pushing us towards.

Paris Marx [02:09:14]:
I mean, like, we've literally had companies have to rehire people, right, after firing them because

Devindra Hardawar [02:09:19]:
they've had to rehire. They've had to have people come back in to fix work. Also something you think about with Roombots.

Paris Marx [02:09:25]:
Yeah, they don't, they don't release the press releases when they rehire people though, right?

Anthony Ha [02:09:29]:
Weirdly, Ford did. Ford went on to like, on an earnings call, they did a whole talk about how they rehired their gray-beard engineers. And like, it was actually like delivering all these like good outcomes, which I found interesting both because they did that, but more that they were publicizing that as a good thing.

Paris Marx [02:09:46]:
Yeah.

Devindra Hardawar [02:09:48]:
But just like with the Roombas, right, the robot vacuum does not solve all your trouble. Then your life is about maintaining the robot vacuum, making sure the floor is clear before the robot vacuum has to work, and then replacing components down the line. So it is just another level of work that you're doing, right? You're maybe not doing the vacuuming yourself, but you're still doing something. And, you know, that's just part of like how things go. Yeah.

Paris Marx [02:10:11]:
It's like the, the work that was done on, um, in like the '50s and '60s when there was a lot of like changes to, you know, kind of, uh, the home, right? And, and, you know, like the dishwasher was introduced and all that kind of stuff. And, you know, the argument for a lot of these new appliances was kind of like, it's going to make it easier for, you know, the woman, the, the housewife to, you keep the house clean and blah, blah, blah. And actually, like, the, the kind of like demands on a woman's time did not get lessened. It was just like the standards of cleanliness and all that kind of stuff, like, evolved, uh, you know, as a result of that. Right. And there's, there's like good books on this. I'm probably butchering it with how I'm describing it. But like, you know, it's— we don't always see the kind of expected labor savings.

Paris Marx [02:10:51]:
We see, see other changes as a result of the introduction of these things. I would say, though, if there's one robot— Yeah. we should be introducing in every one of our homes in North America. This is like one of my, like, kind of like hobbyhorse, like things I push all the time is like, man, we need the Japanese and Korean toilet seats over here. We're like barbarians. Not that you have them.

Devindra Hardawar [02:11:08]:
Nobody is stopping you from getting them.

Paris Marx [02:11:10]:
Oh, I have them. Don't worry.

Devindra Hardawar [02:11:12]:
Yeah, yeah, we have.

Paris Marx [02:11:13]:
I think everyone else needs to as well. Right?

Anthony Ha [02:11:16]:
Like, do not emphasize the individual solutions to societal problems.

Paris Marx [02:11:20]:
Exactly. Exactly. Like, when is the building code going to change so that every toilet in needs the little plug behind it like they have over— that's true.

Devindra Hardawar [02:11:27]:
Yeah,

Paris Marx [02:11:28]:
you know,

Devindra Hardawar [02:11:29]:
first things I did when I moved into my home here was like, we were doing electrical work, I'm like, yeah, put that outlet right there. Yeah, I need that outlet. It does, it does make your life a lot better.

Paris Marx [02:11:39]:
100%.

Devindra Hardawar [02:11:40]:
So much better. So much better.

Paris Marx [02:11:42]:
And then you wonder like how we as a society, like, you know, obviously we have these narratives, you know, particularly right-wing narratives about how like we are the peak of civilization, right? And then, uh, you know, Europe and North America and all this kind of stuff. And then you go to Asia and you see like their bidet culture and stuff like that. And you're like, man, we are fucking disgusting. Oh, sorry. I don't know if I can curse on this,

Anthony Ha [02:12:05]:
but I

Paris Marx [02:12:05]:
already— We are so disgusting. Okay. Confirmed. But yeah, it's like, it's like, it's like, it's like, how are these not everywhere? And how can we really think that we're so advanced? advanced, uh, if we're still using, you know, paper to wipe our bums, you know, let's put it that way.

Devindra Hardawar [02:12:23]:
And also, listen, uh, many cultures, many cultures have been using water in many different ways. Uh, I don't know, India and, uh, you know, other South Asian countries, uh, the Middle East. Like, there are many different ways to clean your bums, and it is cleanliness in so many, so many different ways. I feel like as we wrap up these episodes, I love to talk to media and what you guys are into right now. Um, so I'd love to know, what are you guys watching, reading, listening to that you'd like listeners to know about? Hmm. Yeah.

Paris Marx [02:13:00]:
Do you want me to go first? Do you want to go, Anthony?

Devindra Hardawar [02:13:01]:
I'm gonna— I'll shout out, I am nearing the end because it takes me forever to finish video games these days. Um, but I am nearing the of Control Resonant, and I freaking— I love it. I love it. I love Remedy. I love everything they do. I love the Control universe. Um, and I think there was some like early growing pains for this game as they figured out how to like, um, deal with the, the, the power sets and everything. Like, I feel like the balance has changed over time in this game, but it's a very different game from the first one.

Devindra Hardawar [02:13:30]:
But I love Control. I love this universe. And Remedy is a developer I will follow to the ends of the earth, even if they make make bad games. So that's just me. And I believe if you are into Control and you have not read the book, what is the book called? There Is No Anti-Memetic Division? Yes, There Is No Anti-Memetic Division is also a wonderful book and also in a similar vein to Control and the way it tells its story. But do you guys have anything you want to shout out?

Anthony Ha [02:14:03]:
I've been watching Vision Quest on Disney+, which I—

Devindra Hardawar [02:14:07]:
I forgot that existed.

Anthony Ha [02:14:08]:
I stopped watching Marvel shows for years, but because it's the same showrunner, Terry Matalas, as season 3 of Star Trek: Picard— RIP Patrick Stewart— I checked it out. And I really, really like it. And particularly the last 2, the 2 most recent episodes that aired, I thought the show really leveled up. And I was super impressed with particularly episode 5. And it's just— I mean, a big part of it is actually because of James Spader. James Spader is just incredible. Oh yeah. In this one, they're not doing a CG robot mostly.

Anthony Ha [02:14:42]:
It's just him. And I will say that I think, yeah, the depiction of AI in the show has very little to do with AI in real life. That's fine. It's a good time. And then I also wanted to say that I recently played the Outer Worlds 2 for the first time. And, you know, that came out about a year ago, I think didn't do as well as expected. There's, you know, Obsidian is— like every studio that's part of Microsoft is having a really rough time right now. So I wanted to, you know, play, catch up on one of their games.

Anthony Ha [02:15:10]:
And it's great. It's one of my favorite RPGs.

Devindra Hardawar [02:15:13]:
The first Outer Worlds was really, really good. I've not checked out the second one.

Anthony Ha [02:15:16]:
The second one is pretty much better in every way. Yeah.

Paris Marx [02:15:20]:
Yeah, I haven't played that one. And unfortunately, I'm kind of like out of the loop on video games because I'm traveling so much for book events. I really wanted to play the new Wolverine game, but I still haven't gotten to it yet.

Devindra Hardawar [02:15:33]:
I've heard so many disappointing things about that.

Paris Marx [02:15:35]:
Really? Okay. So maybe it's for the best that I haven't gotten to it yet.

Devindra Hardawar [02:15:38]:
Take a pause on that one. I think you can wait a little bit.

Paris Marx [02:15:41]:
Okay, good. I feel less disappointed now, so that's okay. Now, I've actually been watching Buffy the Vampire Slayer. Yay! Because I had this thing where I go I just really enjoy going through some of these old shows. And I finished up X-Files recently and was like, what's the next one? And so, yeah, I started doing Buffy the Vampire Slayer.

Devindra Hardawar [02:16:01]:
So is this your first time or are you rewatching?

Paris Marx [02:16:03]:
No, first time. So I've like, I saw episodes when I was younger, but I never like watched the whole thing through. So yeah, it's my first time like actually going through it and I'm pretty excited. What a journey. Yeah. I think I really feel like, and this is something I hold really dear, But I really feel that we lost something in, like, how streaming has, like, really taken— and, like, you know, kind of like the HBO model and how that moved into streaming— Oh, definitely, yeah. Has really taken over how we make television now. Um, and I think that there was— you know, it was not perfect, and obviously it was the result of certain kind of, um, you know, uh, financial incentives and business models and whatnot for how these shows were made.

Paris Marx [02:16:41]:
But, like, you know, you do really have a closer connection to these characters when you have 20-odd episodes a season, and it goes for like 7, 8 seasons. And, you know, you just have time to like get to know them, to get in on the story. You know, the people making it have time for it to evolve and to get better and things. And which is not to say that there aren't stories where it makes total sense to have 8 or 10 episodes and maybe only 1 season and whatever. Yeah. But I also feel like there's a lot of stories that now get kind of crammed into this model that maybe aren't as good a fit for it. And I think you see that with the more recent Star Trek stuff to a large degree, right? Where it just doesn't— even when they try to do it with Strange New Worlds, it just doesn't feel the same as the older seasons and the older ways that they would do it. I feel like in terms of books, I'm reading Emily St.

Paris Marx [02:17:33]:
John Mandel's The Glass Hotel right now. Oh, cool. She actually had a new book out last month, I believe, which I'm really really looking forward to reading. Um, but she's like one of my favorite authors, and I just think she's fantastic. For people who aren't familiar, she was the author who wrote the book Station Eleven that HBO made a series like 6-ish years ago of, I think. Um, but yeah, if, if people are into speculative fiction, like, I can't recommend her enough.

Devindra Hardawar [02:17:56]:
I love her work. Have you seen that show, Paris?

Paris Marx [02:17:59]:
Station Eleven?

Devindra Hardawar [02:18:00]:
Yeah.

Paris Marx [02:18:00]:
Oh yeah, I loved it.

Devindra Hardawar [02:18:01]:
Yeah, the show, I feel like, was the first like great post-pandemic piece of art that we got that also kind of reckoned with that whole thing. So that's awesome. That's awesome. All right. We've got a couple more things to talk about, but let's take a quick break to hear a word from the folks at TWiT. All right. Okay. And let me see here what else we could do.

Devindra Hardawar [02:18:25]:
Okay. Oh, I meant to ask you guys this before we talked about the fun stuff, the media stuff, but what are you working on these days? I know you've got Paris, anything else that's cooking you want to tell the audience about?

Paris Marx [02:18:38]:
Nah, I'm very kind of consumed with book stuff at the moment, for better or worse. Um, you know, I obviously still make the podcast each week, Tech Won't Save Us. Um, but that's kind of like the extent of my, of my, you know, occasional articles and things like that as well. But that's like the extent of my, um, output at the moment. I'm, I'm already kind of like thinking about a new book idea, but obviously that won't be materialized Amazing. For the couple of years to come at this point.

Devindra Hardawar [02:19:02]:
Amazing. Because I feel like every time I've talked to somebody who's written a book, they're like, I will never do this again. And then 6 months later, they're back in hell. Yeah.

Paris Marx [02:19:10]:
Once I finished the book, that's very much how I was feeling. And then it's had time to, like, you know, kind of like sit with me for a while. And, you know, I've had time to, like, think through the things that I enjoyed. And also, like, the kind of trauma of having written the book has faded over the the few months, right, since I finished it, or the 6 months or whatever. And yeah, and so now it's kind of like, hmm, what would I like to— it's, you know, I find that the writing can be very stressful just because of like my process and whatnot. But like the part of it where you're researching and like getting to talk to people about things that you are really interested in and getting to learn more about it, like, you know, it's kind of wonderful to be able to do that. Um, and, and to get the time to be able to do it, if you can kind of carve out the amount of time, uh, to, to kind of do the research for one of these books. So yeah, that's the piece that I'm looking forward to.

Paris Marx [02:20:00]:
And then like the difficult parts of like writing it and editing it and then promoting it, uh, come after, right? And you just have to promote it.

Devindra Hardawar [02:20:07]:
The promoting is the fun part. At least you get to travel around and maybe see, maybe see things in these different cities you're visiting.

Paris Marx [02:20:13]:
Totally.

Devindra Hardawar [02:20:13]:
It's great to meet people

Paris Marx [02:20:14]:
and talk to people about it,

Devindra Hardawar [02:20:15]:
but

Paris Marx [02:20:16]:
it's all the travel is also It's,

Devindra Hardawar [02:20:19]:
it's, I'm sure, um, I would recommend to you and to everybody traveling and get some sort of display glasses. Um, you don't have to have the creepy smart glasses, but the stuff from XREAL and all those companies, you plug it into your phone or your computer or whatever, and you've got a 150-inch screen in front of you. And that was a tremendous way for me to fly while I was, um, covering that Microsoft and NVIDIA event, also just to hang out at the airport. So I recommend that to everybody. Interesting. That's my big tip. big thing there. Um, and those things— I'm

Paris Marx [02:20:49]:
a big, like, reader and, like, music listener on the plane. And I'm, I'm one of the losers who, like, puts the flight map out, uh, up the whole plane so I can see where we are.

Devindra Hardawar [02:20:58]:
Yeah, I mean, sometimes, sometimes you just kind of have to. Anthony, anything you want to shout out?

Anthony Ha [02:21:04]:
Uh, yeah, I mean, you can read most of my tech writing at techcrunch.com. Uh, I was telling Paris before we started recording that I'm currently corresponding with his book PR person about scheduling interview. So hopefully you'll see that in the coming weeks. And then I have a science fiction short story called The Tourist that is coming out in Asimov Science Fiction next year.

Devindra Hardawar [02:21:22]:
Nice.

Anthony Ha [02:21:22]:
The editors at Asimov are great in many ways. That sounds like a good press. Thank you. They're not great at saying exactly when things will come out. So I think it's gonna be January, February. I'm not 100% sure. But it's basically a story where I put all my feelings about potentially becoming a parent with a climate apocalypse All right. Congrats.

Devindra Hardawar [02:21:39]:
Looking forward to reading

Paris Marx [02:21:41]:
it.

Devindra Hardawar [02:21:42]:
It's also great to hear just like, hey, new book projects, you know, and new short story projects that people are reading. And that's something hard to think about when you're online. It seems like everything is moving towards TikTok videos. And that seems like the future and nobody wants to read articles anymore. So that's—

Paris Marx [02:22:00]:
And I would say, like, I feel like this fall, there have just been so many fantastic books that have come out. And like, I'm not just saying that because like, I had a book come out as well, but they're really like, you mentioned Gil Duran's book earlier. Naomi Klein and Esther Taylor had a new book out. Yeah, I read Esther Taylor. Paulina Borsuk's book has been reissued. Amazing. And that's just a really small selection. There are so many good books that have come out this fall, and I already know about a bunch that are coming out in the new year.

Paris Marx [02:22:29]:
Dave Karff has a new book that's coming out called Vaporware, I believe, which I'm really looking forward to. forward to. Like, it's just, I, you know, obviously we're in the AI moment at the, at the, at the, you know, right now, and, and we're worried about what that's going to mean for reading and literature and all this kind of stuff. But at the same time as that's happening, like, there's so much good stuff coming out right now.

Devindra Hardawar [02:22:47]:
That's awesome. And yeah, I've noticed all this. Didn't you talk to Paulina Borsuk, Anthony?

Anthony Ha [02:22:53]:
I did. Yeah, yeah, I interviewed her a couple weeks ago. Um, and she, she's also super fun to talk to. Um,

Paris Marx [02:22:58]:
I'm planning I'm looking forward to it.

Devindra Hardawar [02:23:00]:
Awesome. Uh, you know, uh, history there, uh, she was a renowned tech critic like way back when in the '80s and the '90s, and then her career just tanked because it seemed like there was a concerted effort to quiet her voice. So it's really cool that her book is back out there. So check that out. What's the book called again?

Anthony Ha [02:23:17]:
Uh, Cyber Selfish.

Devindra Hardawar [02:23:18]:
Cyber Selfish. There we go.

Paris Marx [02:23:19]:
Wired also has a really interesting kind of like investigation as to, uh, what actually happened. into her career, right? Yeah, trying to get some answers.

Devindra Hardawar [02:23:27]:
Really, really fascinating stuff. I'm glad to see her voice is back out there. I am doing stuff at Engadget. I'm doing a lot of coverage of upcoming laptops, so you'll hear my rant about the whole Google Books thing.

Paris Marx [02:23:40]:
Um, looking forward to that one.

Devindra Hardawar [02:23:41]:
Yeah, it's— it was fun. It was fun to write that one because I'm like, what the hell are you doing here, guys? Uh, there are some new machines coming from Apple that I may have some thoughts on. So, you know, uh, look into that next week in Engadget. We've also got some upcoming coverage on more of the, like, uh, the RTX Spark machines. So it took me forever to get a freaking, uh, Surface Laptop, uh, Ultra, and I have one testing that out, and I'll have some thoughts on that eventually too. So check that out. And also podcasts about movies and TV at The Film Cast at thefilmcast.com. And for our final segment, in memoriam, rest in peace, Patrick Stewart.

Devindra Hardawar [02:24:23]:
He is an actor and performer who I'm sure has influenced all of us. Um, I was never a Star Trek kid. I'd never watched the last Star Trek: The Next Generation, but I did see him in so many things. And for me, he was my Professor X. I was a big X-Men fan. Those early movies. movies, I thought, truly successful comic book movies after Superman. So he's been a titan.

Devindra Hardawar [02:24:51]:
I've always loved his work. But what do you— do you guys want to share some thoughts here?

Anthony Ha [02:24:56]:
Yeah, I mean, he meant a lot to me for sure. I mean, it hit me a lot harder than I was expecting.

Devindra Hardawar [02:25:02]:
It seemed like he would live forever, right? Like, yeah, he seemed ageless.

Anthony Ha [02:25:06]:
Yeah, I was actually rewatching Star Trek: First Contact after I found out. I think one of the things that's really striking about the vision— the Star Trek future in general, but particularly the way Picard lays it out in First Contact— is that actually the technological advancement is a relatively small part of what makes Star Trek a utopian future. Yeah. It's more that it's post-capitalist. It's that everyone has the freedom to pursue what's important to them. Granted, when Picard is saying this, there's an irony that maybe he's deluding himself because he wants to kill all the Borg. That's essentially the vision that part of what Picard was, was that he embodied this idea that we could become better people. The technology was part of that, certainly having the replicator and being able to create, feed yourself anytime you to feed yourself is important, but it was far from the most important part.

Anthony Ha [02:26:03]:
The other thing I'll mention is just that I've read his memoir, Making It So, and I highly recommend it. It's like this really great account of being a working Shakespearean actor in the '60s and '70s. And part of the argument he makes there is that this is what it looks like when the government invests heavily in the arts. And then apparently that all got ripped out under Margaret Thatcher. Thatcher. And so it's kind of a shadow of its former self.

Paris Marx [02:26:29]:
Yeah, you can see that a lot with music in Britain as well, right? How you had so many of these artists, working-class bands, make it big because they could just do their art and not have to worry about making money for a long time. And then you've seen many less of that kind of artist after those programs were decimated, right? But I feel like, I feel like for me, Devindra, my story with Patrick Stewart is more similar to yours where, you know, Professor X was like, you know, where I first encountered him. I was less of a Star Trek person when I was younger and actually became more of a Star Trek person as I've, you know, gotten older, right? And, you know, watch a lot of Next Generation and the films and stuff like that. But it's also just been really nice to like, the past few days, you know, since he died to see all of these really nice interviews and things that he did circulating online, speaking to the quality of character that he had. There are many that I could point to, but there was one that I found really funny where he was doing an interview on one of the late-night programs, picking up on what Anthony was saying. He was like, I first became politically active after the war. And I've been a left-wing man ever since. And he said, it's funny to be in these dinners in Beverly Hills and to tell everyone at the table that you're a socialist and see their faces.

Paris Marx [02:27:56]:
I found that really funny. And to think that even after all those years that he still kind of held those convictions really strongly. And you could see that in the way that he treated people and the way that he acted. I think it really spoke to the kind of man that he was.

Devindra Hardawar [02:28:11]:
He's a real one. He was a real one. And speaking of, like, millennial optimism, which is a term I just love to see, because I guess, yeah, I guess there was something kind of weird and special. It was the Obama-era optimism, right, in America, at least. But I was in Brooklyn when Patrick Stewart was in Brooklyn. And it was wild that he lived in, like, he was near Gowanus, friends of ours. would see him at their local wine store occasionally. He was there with a very young wife, but also essentially living the like millennial Brooklyn life and also having his, um, palling around time with, uh, with, uh, what's his name? Ian McKellen.

Devindra Hardawar [02:28:51]:
Ian McKellen. As well, their bromance and everything. Just, just a treasure. And it's so sad to see him go. And I hope people take something away from his legacy, but yeah. Yeah, Patrick Stewart, you will be missed. And thank you guys so much for an excellent episode of TWiT. Any parting thoughts as we leave this episode? I think we had a really good chat.

Devindra Hardawar [02:29:16]:
What is your main takeaway that you want people, what do you want listeners to take away from you?

Paris Marx [02:29:22]:
I feel like we're finally seeing, you know, I think some real kind of, what holes in the AI bubble. I think that there's really something starting to happen here. And obviously we'll see how that shakes out. As you said earlier, Devendra, it'd probably be better if it's kind of a slow decline and not just a total collapse. Yes. That would be ideal. But it definitely feels like there's a reassessment happening in this moment on data centers, but on generative AI and kind of the return Yeah, I think that's a really good point. And I think that's a big part of the problem.

Paris Marx [02:29:53]:
Yeah. Is that there's a lot of confusion about the return on investment and kind of the market opportunity that it actually presents versus, you know, what the companies have presented. You know, and obviously I think that the public sentiment has helped to really get people to look more clearly at what is being proposed here, whether that's the data center opposition, but also the public anger on AI more generally. You know, sometimes it's anger, sometimes it's just frustrations and whatnot. Um, but yeah, I would hope, I would hope that, you know, as we see this play out, that the story is not just, uh, it's collapsing and it didn't work out. Um, but what is the next thing that the tech industry is going to sell us and how are they going to try to like reconceive our future next? But, you know, how can we try to take back some kind of collective power over determining our own future in a way that I feel like we have outsourced to the tech industry for the past 2 or 3 decades now? And I don't think that has served us particularly, particularly well. So yeah, I think, you know, I think there's a lot of reasons to be very worried and very down on kind of the state of everything at the moment. Of course, you know, the midterms are just in 2 days.

Paris Marx [02:31:01]:
But I, you know, I also think that there are a lot of opportunities for hope right now as well with how things are going.

Anthony Ha [02:31:08]:
Yeah, I would basically just repeat what Paris said, but maybe to elaborate on it slightly also is like, Part of the inevitability narrative around AI was this idea that what you think, if you don't like this, doesn't matter. And I think what we've seen over the past year or so is that it totally matters. It matters a huge deal. And I found that to be tremendously inspiring, even if there are moments where sometimes I sort of roll my eyes at a certain anti-AI post or whatever. But fundamentally, that's been a hugely inspiring thing. And my hope is that My hope is that, you know, as both that we sort of keep an eye out for like, okay, what's the next thing they try to like sell us on? But also, yeah, what is the better thing that we can all build together? Like, what is a better version of this tech industry? I don't have any answers, but, but I'd like to find out.

Devindra Hardawar [02:31:55]:
Yeah, I feel like the better thing is human-centered, and hopefully we'll keep that in mind moving forward. Remember folks, nothing is inevitable. This has been This Week in Tech.