(Preview) Fable 5.1 and Anthropic’s Data Retention Pivot, AI Civilizations and Related Matters, Q&A on Meta, Shopify, 3-D Printing

Server rack remodel updates

0:00Hello, and welcome to a free preview of Sharp Tech.

0:09Hello, and welcome back to another episode of Sharp Tech. I'm Andrew Sharp, and on the other line, Ben Thompson. Ben, a sincere how you doing? What's up? Oh, you know, what is it when you've been in the dark? You made it out of the server room. I wasn't sure I'd ever see you again. You emerge out into the sunlight, or you're in a desert, you're thirsty, you stumble upon an oasis. I spent two days, Andrew, without internet.

0:40Isn't an oasis fake? I don't know if an oasis is a good thing. No, that's a mirage. Oh, okay. An oasis is what you want. There are actual oasis in the desert. And that's what we're providing here. As we will get to in this episode, I may not know how to say words, but I do know what words mean. Good. So that's how we could be a good partner here. Honestly. Yes. But yeah, so before I moved to this house, I had some AV company wired up a rack and a bunch of TV stuff and all this sort of thing.

1:11Over time, as one does, if one is a computer geek and a nerd and all these sorts of things, one wants to use the rack for actual computer things. And the problem is that these companies do these racks assuming no homeowner is ever going to touch them. They just sort of want stuff to work. Right. So they do shortcuts. It's like you have – so the good thing is I knew if I was modeling a house, I wanted Ethernet jacks everywhere. And so there are like literally like I think like 70 or 80 like Ethernet runs all over my house.

1:42Like every room has multiple – it's actually awesome because it unlocks a bunch of stuff I kind of am thinking about want to do in the long run. And anything runs over Ethernet, whether you run video over it, you can run sound over it. Like it's just – it's a universal transport mechanism. It's kind of amazing this technology that like dates back so long. It's still under – you see those pictures of data centers? Like those are all Ethernet cables. Like it is one of the – it is unlike the Mount Rushmore of great tech products, the Ethernet cable. Right next to EasyPass, sure.

2:13Right. So what you actually want to do for something like this is you have a thing called a patch panel or a patch interface where all those run into – are terminated at sort of like the front of the server. And then all your computer gear is all next to it. And then you run those short little cables like you're a sound engineer or like a synthesizer from the 70s, like connecting different things together. That's how it should work. And then you have a flexibility and you can change stuff and you don't make a huge mess trying to pull apart all these different cables.

2:45It's all very clear, organized. So anyhow, I've been messing around with a lot of stuff. I've been talking about it's the year of – like hardware is back, baby. I keep getting more things. And my rack was a disaster. You sent me a picture of it last night and it gave me a panic attack looking at what you had before this little remodeling of the server room that has been taking place. Right. So, yes. But what needs to be done is you have to take everything out, take all the wires out, take all the equipment out, put it back in.

3:16You have to have all these patch panels. I think I have like six patch panels. You have to then reach in. You have to identify every single wire from all over the house, plug it into the right one. It is like this mammoth, mammoth job. Anyhow, we are almost there. One more day tomorrow. The audio video side is not done. But the internet side is done. I have internet. It's back. As part of this project, I wrote last night using the Starlink. I pulled out the Starlink, which was a part of that. Yeah. And so I did have internet.

3:46But in the future, I now have a secondary network in the house such that the whole house will have internet. Resilience is important as always. I mean, it is, I will just say from my standpoint, one of the funniest parts about working with you is I consider myself pretty well versed in like tech and business at this point in my life based on the last three or four years working together. At least fluent in the structural dynamics, the trends that really matter, hard tech and how tech actually works, I will save that for another lifetime.

4:19So I'm going back and forth with you about this project. And I asked, like, what are you actually doing? And you hit me back with pictures of all the server racks and everything else that were a complete mess. And this like color coded chart that I believe is the goal and what you're trying to create in this server room. And the chart you sent me is completely inscrutable. You sent it to me as if it would explain everything. No, it is. What are we doing? So unfortunately, the chart very quickly reality hits.

4:51And so it does fall out of date. So like we printed a big one out and it's like it's hanging in the room so you can reference it. It's already like, but I do want to update it once everything is final because I mean, it's actually great. I found more runs in my house. I didn't really even realize I had stuff. Some of us got moved around, but we added a bunch of new runs. Yeah. And I want that thing is beautiful. I want to frame it and hang it in the server room. It's beautiful and it's beautiful. It look, I enjoyed the colors on the color coded chart. In any event, we could update people on your progress next week.

Anthropic releases Fable 5.1

5:23For now, the news this week, Fable 5.1 arrived. Anthropics stressed that it is the exact same model as Mythos 5.1. So why don't we start there? Why did they stress that Fable and Mythos are the same model? I don't know. I just thought it was kind of funny. Like the very first line, they're like, no, these are really the same model. I think probably the way it was released, I'm guessing there was concern it was like or, you know, just from a branding perspective. Like, I think there's a bit, well, no, I think they did get the payoff from their scaremongering that they wanted to get.

5:57They sell like Mythos as this unbelievable, scary thing that's going to take over the world. And, oh, U.S. government, please stop us. We did it again. Look, I've been sitting here for the last three months being like, well, I guess Mythos is the internal model. I wonder how powerful that is. Fable is what they're allowed to release to the public. So, yes, it's probably a useful clarification from Anthropic PR. No, it's like marketing. It's like, no, no, no, we're giving you access to the model that's so dangerous the world's going to end. I mean, this whole release was very funny to me for lots of reasons, but this was one of them.

6:33Just first line, like we have to say, no, they're the same model. I can't believe only 11% of our customers are using it or whatever the number is because you have access to the model. Anthropic drinks, Coke, just like everybody else. You know what I mean? Well, no, it's not that. It's like they spun up the Project Glasswing. The U.S. government shut them down. Wouldn't you like that model for your business? Imagine what the software you could do with that. It's like, no, you have the same one. You really do.

7:03Indeed. Well, the company said that Fable 5.1 is more skilled at carrying out lengthy, complicated computer programming tasks, such as software projects or code reviews that require it to work across all the code for an application. The new version is also better at helping with scientific work, like coming up with experiment designs, simulating what could happen during an experiment, and reading information in dense diagrams and tables. By the way, you know who understood my server rack diagram?

7:34Who's that? Oh, Fable and Codex. Yes, yes. Great. This is great. You could write it there. They'd say, oh, you have a missing connection here, or there's an inconsistency there, or things like that. This was so beautiful, by the way. It was very satisfying. When I had the workers here who were doing a lot of – it was funny. I knew I had to hire people to help me do this. And so they were working on it. But it actually – almost every one of those cables I plugged in, in part because this was such a fever dream. I had the mental model in my head of how the whole thing works, different zones, and all these different bits and pieces.

8:05But it was very gratifying where they would come and ask me questions, and I would explain it, and they'd be like, oh, yeah, you have it right there on the map. I'm like, that's right, I do. I thought about all this. But after the third time that happened, I'm like, okay, I'm just doing it now. And so I – There we go. All day long. Yes. So we were recording at night because I'm like, sorry, Andrew. You're going to have to wait. Yeah. Eight hours later, here we are. I'm just glad that you're still standing after dealing with that all day long. And I'm glad that your new best friends, Codex and Fable, were fans of the chart and knew exactly what they were looking at.

8:41Yes, dense diagrams and tables. So I'll finish reading for you, Anthropics. That companies are increasingly trying – I feel bad. No one wants an entire episode about my rack redesign. Companies are increasingly trying to justify and rein in the cost of AI usage at a time when open-weight, lower-cost models from China are rapidly improving. Andrew, are you supposed to ask me how is Anthropics responding to that challenge? How is Anthropics responding to that challenge? You're the China expert. How is Anthropics responding to that challenge? Well, I am very curious because I'm aware that they charge massive margins. Lots of people have complained about the token costs over the last several months.

9:15People are getting more price sensitive and the competing models are getting more performant than they've ever been before. So what's Anthropics' plan to counter competitors? See, this is where I – whenever I'm frustrated with Fable and Fable just sort of like just says stuff, and I'm like, where did you get that from? Can you explain that? And then I'll come back and be like, yeah, I just kind of said that. I'm not sure why. Here's my question to you. What's the evidence that this is a problem and they need to respond to it?

9:46From what I can see, they don't have enough compute and they clearly still don't have enough compute. Their servers are down all the time. They were down a bunch like last week. In the face of them charging these very high prices, like there is an aspect – and this was kind of something that I was thinking about in the context of NVIDIA's earnings where all these like – it's really interesting to think about the long run of NVIDIA. But I kind of wrote in my update, I'm like, I'm talking about these structural things, and I'm kind of bored by it because on one hand, it's really interesting who's going to capture value in the long run in a world where supply catches up to demand.

10:26Like in that world, someone's going to get squeezed. And as long as we're not there, it's all just theory, right? Like everyone's making high margins. The Frontier models are making high margins on these – NVIDIA's margins are exploding. NVIDIA's making crazy margins. The memory guys are making huge margins. TSMC is making big margins. Everyone is making big margins. It is exciting to be in a world where supply is constrained because of these crazy earnings and all the stuff that happens.

10:57It's also kind of frustrating as an analyst because there's no forcing function to see who actually is capturing value because there's just not enough supply. So the interesting questions are abstract and not particularly relevant for the next three or four years. From what I can see, Anthropik's problem is supply. They don't have enough compute. If they had enough compute such that they could – like right now, for Anthropik to lower – meaningfully lower their prices is – where Fable 5.1, they lowered the cash prices but not the actual generation of token prices.

11:39They're still insanely expensive. And in that world, why should they take compute to serve lower prices, especially when they're a company that is spending tons of money, they have this huge R&D budget. They need the revenue. They need the cash flow. You can make some sort of fanciful, oh, for strategic reasons, they should charge less than they can to take the market. The reality is – no, the market is large enough at their price level to completely saturate their supply.

12:10They need money to fund their operations going forward given that the fact that people are lower down on the price curve doesn't tell us anything. This goes back to my who's afraid of Chinese models piece. If the leading-edge providers had infinite capacity, they would serve their models not just at the same price but at, I would argue, a lower price than these Chinese model providers do because they almost certainly have far more efficiencies in their infrastructure.

12:42So like we just – this whole thing is imposing business questions on an environment where there just isn't the structure to answer them. This whole story is pretty simple as far as Anthropic is concerned. Well, had you asked me where did I come up with that concern, one area of concern – Oh, sorry. I jumped in. One area of concern. No. Also, another thing that – yeah, that failed. Just come in. It has its own thing. Whatever. Well, I've seen a couple different charts over the last month or so indicating that Anthropic's growth has leveled off and leveled off in very relative terms because Anthropic's growth was completely effing insane.

13:22For the last 18 months or so, do you think that's a function of OpenAI entering the market and succeeding or some other story? I don't know. I mean, I think that unless we have an understanding of what their capacity issues are, like did their growth slow down just because they don't have enough capacity to grow anymore? That's NVIDIA's problem. NVIDIA grew 100% last year, and they forecasted this year we're going to grow 70%. And I think that NVIDIA, Jensen Huang, who's very deliberate on his earnings calls, like he has a theme every single time.

13:56And I think that the message here was – it's the message that I wrote back in 2020 – I think it was 2023 where I basically said I'm – it's funny. I feel compelled to cover NVIDIA earnings, and I've been saying for now three years they're actually very boring events. Right. Everyone gets so hyped up about them, but the problem is in a world where demand exceeds supply, that means your revenue is based on how much supply you have.

14:26You are selling every single chip you can make. And so when Jensen Huang comes out and says we're going to grow 70% next year, what he's saying is we are going to acquire sufficient supply to deliver 70% growth. Yeah. That's why in your NVIDIA coverage, I believe it was a week ago, you said their real moat is their supply chains at this point, and that is true. And as far as Anthropic is concerned, I guess it is sort of a mystery. It's been interesting watching OpenAI succeed as quickly as they have with their coding efforts and their sort of enterprise efforts, but I'm certainly not concerned trolling about Anthropic.

15:04Well, I do think this – I do think this – I don't think this is on the rundown, but there is – OpenAI is going to exercise the clause in their contract to no longer serve Cursor, which I think is understandable. Like, it's a direct competitor. Like, it's flowing through their apps. Like, I get it. Like, it's just at the end of the day, you are in a competition. Anthropic comes out right away on top and says, oh, we love Cursor. We would never do that. We support our developers. And it's like, what is the number one answer to their capacity problem?

15:35What's that? SpaceX giving – selling them capacity. So, yeah, I don't think they're going to sell out Cursor. Or SpaceX's, like, great hope when either side can pull the plug with a month's notice. Wow. I didn't catch that, but that now makes the Anthropic response that much more delightful.

Data retention policy rollbacks

15:56So, the other piece of the news this week with Fable is that Anthropic is rolling back its data retention policy. CNBC, it will change a controversial data retention policy for its business customers after, quote, a lot of feedback, and the company said it plans to implement a new solution called Enterprise Frontier Safeguards instead. What do you think of that news? Did the popular uprising succeed? I think so. I thought – I mean, as I put it on Stratechery, you know, the release is top of TechBeam, biggest news.

16:31Everyone's talking about it. You know, TechBeam lists, like, social media commentary. There's, like, 75, like, posts from Twitter about it. And you scroll down, down, down, down, down, and there's the news of this on the page. And number one, clearly no one cares. Number two, there's, like, no X comments and, like, eight LinkedIn comments. And I'm like, this is my world. Yeah. You just have to accept it, you know? You're a son of LinkedIn, not Twitter. Apparently. But this is by far the most interesting part of the release to me.

17:04You go back to when Fable was released, and I was on it, like, immediately. Like, the first day I wrote about it, I wrote another article about it the next week. This is nuts. Right. The idea that, the concept that they are not going to any longer honor zero data retention policies is insane. And it's not just insane in the context of AI. It's insane in the entire history of enterprise computing. The foundational divide between consumer and enterprise is enterprises are obsessed with owning their data, and they would never, ever, in a million years tolerate their vendors taking their data for themselves.

17:47Right. Never. Would never, ever happen. Consumers? Consumers don't really care. Consumers don't care. And that's why you get ads, and you get free apps. It's the, and this is why it's so hard for consumer companies to serve the enterprise, and vice versa, to be honest, but particularly consumer going to enterprise. Like, you know, when I was talking about Dropbox in the context of a company that wanted to be a consumer company but had to become an enterprise company, part of that effort was building this entire infrastructure to manage access to data.

18:22Right. You know, who gets what, who gets what, who gets all these sorts of things. Yeah. Yeah. And it's like the core of enterprise software is, is who gets to, who gets access to what data. And the idea that Anthropic was going to just unilaterally change that was incredible. Right. And like we talked about, you know. Under the auspices of safety, because that was the justification for it a couple months ago. That's right. And what happens are the auspices of safety. It's a, you're, you start rolling down that slope, and we need to hold this data in case an incident happens.

18:55We need to scan it for abnormalities. Oh, do you know how we stopped this? If we trained on this data, we could better understand what's going on. Suddenly they have access to the greatest asset in the world. The greatest asset that these companies have, certainly. Well, that's part of what made it so remarkable is you're stepping back. Like, Anthropic is potentially coming for any number of these businesses and these industries. Absolutely. The idea that you're handing over your data. It was insane.

19:26Yeah. It was, like, I could, like, so people, you know, I referenced the before, no one's buying Fable, partly because it's expensive, partly because these data policies are anathema. Anathema? Anathema? Did I say that right? You nailed it. This rundown has me nervous. I saw the question at the end. Anathema to any enterprise. And so even, doesn't matter, it was so fascinating because Anthropica is, like, weighing down the gallant. We're going to make a model that's so good.

19:56Not only is the U.S. government going to come shut us down, but you're going to give up on not having zero data retention because you need to have this model. And it failed. And here's the thing. If you want to say that, oh, no, Anthropic always was going to build this new system, which is basically store your data wherever you want it. You just have to hold it for at least 30 years.

20:26And I think you have to give Anthropic Sims sort of automated access to search for whatever it is. Even that is probably a bridge too far. But at least the data is, like, on your servers, not Anthropics, where, oh, we changed our policy again. Now we're training on it. Like, that's at least not possible. Not only is that going to happen, but number two, it's not launching for a few months, which means this wasn't something they were working on. It just wasn't ready at Fable. And until the time it is available, Fable now operates with zero data retention.

21:03So what happened to the safety concerns? That's the issue. Yeah. There was no imminent concern justifying all of this. They absolutely made a play for all of enterprise data. And the fact that I called it at the time, they poke everything in safety. They believe it internally with this insane understanding of AI, which we're going to get to in a little bit. And it so perfectly serves their business needs. It's the audacity of coming for all of enterprise's data so that you can make your model better, take over their business, inspire Satya Nadella to have a blogging career, have Alex Karp have this unhinged rant on CNBC, which was quite delightful.

21:46I couldn't sit through all 10 minutes of the Alex Karp rant, but I heard good things. It is what it is. The fact that they will make it zero data retention until this new program is available is the tell. This was not about safety. If it was about safety, you would not do it. You would just, you would say, we're going to continue with our policy until this is available. But no, the policy is gone. Right. Well, do you think that OpenAI's entry into the enterprise space as aggressively as they've been moving that direction over the last couple of months, do you think that makes it easier for companies to say, look, we're just not going to use Anthropic if that's the policy?

22:25Yeah, absolutely. Absolutely. That's why competition is amazing. And by the way, why is Anthropic doing this? Because they need to IPO. Because they need to make money. That's why the price is the way they do. This goes back to one of my longest standing takes about tech in general. And one of my longest standing criticisms of these AI companies, particularly OpenAI. Yeah. The best route to alignment from a societal perspective is the for-profit enterprise. The for-profit enterprise, you understand why they do things.

22:55You understand they have fundamental constraints on their behavior. The market is a check. The market is a check. And it is the only check that scales to the size and power of these companies. Even Facebook, where Mark Zuckerberg has total control, has to occasionally pull it back and say, okay, we're not going to go that far. We're going to step back. We're going to do like... And the fact that the idea that we're going to start a non-profit for this, that is what concerns me.

23:26Not having a check. The check should not be the good intentions of the people in charge. That always ends badly. The check has to be external. And it's the external check. The only one that works is the market. God bless cop capitalism. God bless free enterprise. God bless Anthropics IPO. This is evidence. Yeah, the IPO is definitely coming because they do enterprise contracts. And stepping back and watching the way this has played out over the last couple months, I don't think it plays out the same way if OpenAI isn't going hard after the enterprise space.

24:05And also, they, too, were checked by the market and forced to go that direction and find a bunch of revenue in enterprise that they weren't finding elsewhere as they did 10 different things. We're not doing Sora or whatever it might be. Yeah. No, absolutely. The market pushed OpenAI in the right direction as well. All right. Another topic of conversation this week. Over the weekend, Dworkesh Patel published a post on the OpenAI hugging face hacking incident. All right. And that is the end of the free preview.

24:38If you'd like to hear more from Ben and I, there are links to subscribe in the show notes, or you can also go to sharptech.fm. Either option will get you access to a personalized feed that has all the shows we do every week, plus lots more great content from Stratechery and the Stratechery Plus bundle. Check it out. And if you've got feedback, please email us at email at sharptech.fm. And if you've got feedback from Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery and the Stratechery.