Anthropic’s $900B Valuation Beats OpenAI, Claude 4.8 Drops, Former Shopify CTO on AI Risk

0:13Laura Mandaro:Welcome, everyone, to The Information's TI-TV. My name is Akash Pasricha. It is Friday, May 29th. Before we get to today's show, I want to flag some exclusive reporting The Information's Asia Bureau published this morning about ByteDance working on AI chips that could take on a similar flavor to the Grok chips, which NVIDIA is licensing. That is up on our website right now. I encourage you to check it out. Today on the show, we are unpacking Anthropik's latest Opus 4.8 model and what developers are thinking of it. We'll also break down Anthropik's big funding round that puts it ahead of OpenAI in terms of size.

0:50Laura Mandaro:That is the subject of this week's Editor's Cut. And we're also bringing on the former CTO of Shopify to talk about his new role at an AI legal and contracting company. We'll also talk about the e-commerce sector as a whole. It's going to be a fun show, so let's get right on into it. Anthropic released its newest model, Claude Opus 4.8. Anytime there's a new model released, there is a lot to break down. And so I want to bring on Kobe Blumenfeld-Gantz, the co-founder and CEO of AI-powered Medicare Broker Chapter, to share with us his team's earliest thoughts on the model. Kobe, welcome to the show.

1:26Laura Mandaro:It's great to have you here. Thanks for having me. So you guys have been playing with Opus 4.8 for a couple days now, sounds like. What's the initial reaction? It's been fun to play with. It's not that different a model in terms of the user experience and what we can see than 4.7. What we have seen a little bit of a difference on is the decrease in mistakes that it's making. And in a high consequence environment that we operate in as helping seniors with their healthcare, with their health insurance, that does actually matter. But the improvements do seem rather marginal at this point. And are you guys using it where in your text?

2:07Laura Mandaro:Are you using it for coding? Are you using it to help people find the plans themselves? Just tell us a little bit about the use case. We use a bunch of different models for a lot of different use cases, but specifically for 4.7 and 4.8, we do use it for coding. Actually, some of our team recently switched to Cursors Composer 2.5 because it's just much more cost-efficient, which is actually built on an open-source model. But broadly, we use LLMs and a lot of Opus for automating workflows, for doing a lot of background work so that our advisors and the end user, the senior, don't have to fill out as much work and really stringing together end-to-end workflows.

2:47Laura Mandaro:Now, how do you guys assess the effectiveness of these models when you're using it? There's obviously the benchmarks, but then there's the effectiveness on the ground. I mean, do you have a quantitative way of measuring the success of these models at your company, or is it kind of just anecdotal? It's somewhere in between. I think most companies don't actually need to build robust benchmarking because the reality is either the model solves the problem or it doesn't. And as long as you have good checks and good quality verification, that's actually the most important thing. and whether it's living up to some arbitrary or abstract benchmark, I think is secondary in most operational real use cases.

3:34But we are seeing that as the models come out, what actually matters the most is the context. It's the data that goes into it. Most of the models are actually quite good at reasoning today and there will continue to be improvements. But it's really how do we have the brain? How do we build the eyes? How do we give it the legs, the nose, the hearing, the ears? so that the full context can be provided to the model so it can actually give good output. And that's much more determinant of quality of model than the model itself today.

4:05Laura Mandaro:I've not heard yet the analogy of giving the model legs, ears, and eyes, but I like it and I see your point with it. Tell me, how does it compare to OpenAI's models from what you can see? Again, it's not so different. We constantly switch back and forth, frankly, based on what's most effective for a given use case. I think the reality today is most models have just gotten to a point and they will continue to improve again, but it's rather incremental. At least that's what we're seeing. So I think 4.8 is a little bit better than the latest OpenAI model, but it's not materially better. And I wouldn't be surprised if in a few months there's a new OpenAI model that is a little bit better than 4.8.

4:51Right.

4:51Laura Mandaro:Now, I want to go back to what you were saying earlier, though, about coding models. You said that you were working with open source models for coding. And my understanding of open source models is they're often not the best models, but they can be more cost effective. So sounds like you guys have taken that approach and you're actually not needing to use the best in class right now for coding. We're not dogmatic about it as a company. Our view is we want to give our team the best tools for a given purpose and the different models do have different performance areas in different contexts. So I think it depends on the type of coding you're doing.

5:32If you're building things that are relatively simple that don't require nearly as much context, then some of the open source models can be very efficient and effective. And because they aren't trying to overcomplicate things, sometimes they're actually even better than some of the latest models, which can overcomplicate some very simple tasks. But yes, if you're doing more difficult, if you're trying to build a more difficult or more complex system, we generally then do rely on the latest models. So it's using the right tool for the right purpose.

6:04Laura Mandaro:Do you see your use of open source as a proportion of your model usage, do you see that going up in the future? It's hard to say. My guess is yes. Open source is still a very small percent of what we use. We still are vast majority on OpenAI, Anthropic, and sometimes Gemini, but more and more OpenAI and Anthropic. We are just starting to use some of the open source models as they get better and better, but I wouldn't be surprised if that becomes a more common trend. And the reason I'm asking because I was having coffee with an executive the other day. We were talking about the extent to which enterprises may in fact seem more used out of using the older models because of the fact that there's just more practice with them.

6:56Laura Mandaro:you have a better sense for how to integrate them. In some cases, they're cheaper too. Kind of reminds me of the chip discussion, right? I mean, you might not need the Verirubin. You might be getting everything you need out of a Blackwell. I mean, do you see that being the future while all these Frontier models are coming out? Are we actually going to see enterprises more adopt models that came out maybe a year ago, 18 months ago, because they're good enough? I think there's a high likelihood of that, in particular, because it seems like Anthropics has kind of taken a page out of the Apple playbook where they have, whether it's delivered or not, I don't know, but they have decreased the quality of the prior models right before they launched the new model so that the new model feels better.

7:42Laura Mandaro:What do you mean by that? Explain that a little bit more. So from March to April, Anthropic said that they were, they later claimed that they were, or acknowledged that they were having a compute issue on 4.7. And so they had tried to make it seem like 4.7 was doing very well. They were actually having compute issues. They've moved more compute to 4.8. It makes 4.7 feel a lot slower and feel a lot worse. In fact, our team felt it in March and April. That's part of why we started looking at open source models because the cloud models, the entropic models were not performing as well. And so Apple is famous for making its older hardware seem much older with new software upgrades so that they incentivize you to buy the new hardware.

8:32It seems like Anthropic is doing something similar with the new models. And so as they try to push users to the newest models, which are probably going to be more and more expensive, I wouldn't be surprised if companies both go to more open source models and go to older models that aren't being arbitrarily depressed.

8:49Laura Mandaro:So that's, I have not heard that before, I mean, this idea that the Apple playbook could be relevant in models here. I mean, I wonder what implication you think that has then on who ultimately wins. And I ask this question because we saw the news that Anthropic is now valued higher than OpenAI, for the moment, at least. We don't know how these things are going to shake out. But, I mean, in this world where companies need to keep innovating themselves, and also they're facing so much competition, And, you know, I wonder, I mean, how do you think it affects who ultimately wins? I think it's kind of like watching four seasons and flowers bloom.

9:29We're certainly in the anthropic season and the anthropic flower is blooming. Rewind nine months ago, everything was open AI and that was sort of the season. I wouldn't be surprised if in three months we're in the Google season or the X AI season when those companies have an advancement. but really they all have so much money. They all have strong talent. Sure, there are differences in maybe focus and execution on the margin. But at this point, I really don't think there's going to be a material difference between the four largest players, at least over the next two years, call it. And I think that that just creates a really interesting competitive dynamic, which is probably very good for consumers of LLMs because they'll get better and better and cheaper and cheaper.

10:16I don't know what that means for those four companies. I think there's a lot of demand out there, so they'll all be fine. It's a plenty big market for all of them, but I don't think there will be just one winner. Right.

10:26Laura Mandaro:Let me ask you one last question. So your business, I mean, you are in the business of helping people find the best health insurance plan that they can. You must talk to a lot of these health insurers frequently throughout the course of your business. are health insurance companies are they willing to use the latest and greatest models do they prefer the older models that sometimes can be safer you know we we know that there could be risks that the models could be um susceptible to hacks what's their view on it most health insurance companies are pretty similar to most large uh risk-averse enterprises whether it's governments or financial institutions, you name it, they all operate fairly similarly in that they're still very much in the moving phase, early stages of how do we move to even using LLMs.

11:23And so for them, certainly a part of it is risk aversion that they don't want to use the latest model. But a large part of it is they're still just trying to manage that change to get their organizations to be comfortable using AI at all or generative AI at all. So because they move relatively slowly, especially relative to the way tech companies and startups move, they are naturally going to be on much older models, partly for security reasons and partly just because they don't move that fast. So yes, I think what you're saying is spot on.

11:52Laura Mandaro:Great. Well, Kobe, I want to thank you for coming on. That is Kobe Blumenfeld-Gantz, the co-founder and CEO of Chapter here on TIT. Anthropic, like we said, has marched ahead of OpenAI in terms of valuation with its latest funding round. Anthropic raised$65 billion at a$900 billion valuation before the investment. For more on that rivalry, I want to bring on Laura Mondero and Jason Dean for this week's edition of The Editor's Cut. Laura and Jason, welcome to the show. It's great to have you here. Great to be here. Thanks. Okay. So we just talked about Opus 4.8, the latest model. And, Laura, I want to come to you to help us unpack this valuation from Anthropic.

12:36Laura Mandaro:I mean, our last guest was saying, I mean, this is just the season we're in. It kind of depends when you fundraise. But what? Anthropic is winning now, I guess? Well, I mean, OpenAI still has been a powerhouse in fundraising. And, you know, their valuation isn't that shabby either. I think what happened this year is that Claude Code and Cowork, the virtual assistant, really took off. And investors or potential investors saw that. And, you know, we and others were writing that their valuation looked very, very cheap, particularly compared to OpenAI. And it's kind of, you know, was probably the least surprising thing that happened that, you know, investors tried to, you know, invest.

13:27I mean, I think before this funding round kicked off and it was only a couple months after their last fundraise, there was a lot of hullabaloo about whether they could invest in, investors could invest in secondary shares or shares, you know, buy shares from employees. There was a lot of demand there. And you had one of these situations where it didn't seem like Anthropic was planning a fundraise now, but there was intense investor interest. And, you know, at the same time, this revenue growth, which the company's executives had said took them by surprise and seemed a little bit disruptive, to be honest.

14:10You know, Dario was saying on stage that, you know, this exponential growth is great, but maybe, you know, it'd be good to have something that they could, you know, predict a little bit more readily. that meant that they needed more compute capacity because without it, you know, things are slow or they have to rate limit customers more and it tends to be a bad customer experience and OpenAI is ready to pick up that slack very happily.

14:37Laura Mandaro:Right. So Jason, I want to get your thoughts here and also I want to ask you this question in the context of when I talk to people, I mean, look, we know Anthropic, okay? The business community knows Anthropic, But still, you know, you ask people if they're using Claude or ChatGPT. I think most people will tell you ChatGPT. And so there is sort of like an interesting disconnect here between like what the average person uses and then the valuation of the company. Yeah, I think that's true. I mean, really, the challenge for OpenAI on that front is that they are still kind of figuring out how to monetize that brand recognition.

15:17whereas Glianthropic has figured out how to monetize its prowess in workplace applications particularly coding and that's where OpenAI is playing catch up but I totally agree with Laura that this is a long game and OpenAI is the laggard at the moment but I don't think that that is that predicts that it will be forever and And the fact that Dario and company Anthropic were caught off guard by the pace of their growth has consequences. We don't know the pricing of these compute deals that they've been signing, but it stands to reason that they're pretty expensive given that they were done out of desperation to try to catch up with their growth.

16:08whereas OpenAI has been not only raising more money, but spending it, you know, more quickly in a good way that, you know, they've secured compute, more compute earlier in a moment where they weren't necessarily as desperate to do so. So, and Akash, I mean, one thing to remember, there's a lot of noise and focus on the business applications, you know, and that's certainly, you know, when we talk to big institutional investors, they really like Anthropik because it's like, it feels like they know what they're getting, you know, a company that has been focused and is consistently delivered on that.

16:47But there is, in all business cycles, you swing from, you know, valuing the very focused enterprise to valuing the company that has a lot of different business lines for diversification, right? And so OpenAI had been following the more diversified model and still is. Our colleague Anne just wrote a pretty detailed story about the ad business. And so it does seem like they're continuing to just execute on that and try to develop it in just the same way we We saw the big media companies a decade ago like Facebook do. So I wouldn't – I mean, I think it would be unwise to just think, okay, the whole OpenAI's big future is sort of turning into an anthropic-type company.

17:40I think they laid out a different path. It is a little bit muddy because they seem to – you know, they reorg a lot and they have different business leaders leading this function. and they talk about it in sort of a changing way, but they have basically a company that has a consumer business and an enterprise business. And at a certain point, investors may value that even more highly than a very focused enterprise business. Right. I was just going to say, I agree with Laura on that. I think that is a blessing and a curse for OpenAI because it also creates an execution challenge that is in some ways much greater than Anthropics.

18:23There are very few companies, big tech companies, that have successfully built both enterprise and consumer businesses. And I'm not sure that any of those that have done so have built them simultaneously. You know, if you look at Amazon, they built consumer business first, and then the enterprise business with AWS, Google, follow the same pattern. So to try to do both at the same time and sort of shifts the focus from consumer historically now to enterprise, perhaps back to consumer, you know, is going to test their strength as a management organization. And I would just also add that the additional challenge on the consumer side, this is not a two-horse race.

19:06Google is very much a player in consumer AI. Meta is trying. I mean, we saw this week Meta wants to be as well. Meta is trying.

19:15Laura Mandaro:Yeah, they're trying their best at least to make a name. I mean, enterprise is a whole separate question, but I mean, they're trying to get people to pay now for the consumer AI. That's right. That's right. What did you think of that, Jason? I mean, you saw the, so they're testing the consumer AI products. It's a pilot program. My reaction to it, honestly, was that I know they released the Spark models and, you know, people were kind of, they kind of said, okay, well, you gave us something, thank God. but you know they're asking people to pay and I'm not really sure that there's a consensus that their AI is really like the cutting edge thing that people want to pay for what do you think I agree I think that's a huge question people are very meta has trained people over a very long period of time now to use their their various apps for free and to break people out of that you're either going to have to provide an additional functionality with these subscription services that is really, really great because the price points here are not cheap or you're going to have to deprecate somehow the existing services and sort of shift functionality into these subscription models which risks alienating your user base that the advertisers are all targeting and companies have struggled with that sometimes in the past.

20:35So I don't think that's... it's taken them a long time to get to the point where they're beginning to launch subscription. It could well take a long time before that turns into a major revenue source for them. But it's indicative of the fact that they feel both the opportunity, but also the need to justify the enormous amount of money that they are spending on their own compute.

21:04Laura Mandaro:So, Laura, I mean, a year from now, do we think cloud, I mean, the cloud compute opportunity for Meta a year from now, what do you think? Is it a focus for them a year from now? Well, I was going to sort of respond to something else and then I'll start as one of that. I think one thing on the Meta sort of consumer subscriptions is that, you know, OpenAI is trying this too. I'm sure Meta is seeing this, right? ad-supported, low-priced subscriptions,$8,$5. So, I mean, why wouldn't a consumer company try that if OpenAI is going to try that? And, you know, Meta does try consumer subscriptions from time to time, and they're so big that they can generate meaningful revenue, even if a small percentage of their customers do that.

21:55So it may not be as unlikely as it would seem. But I think there will be an interesting thing to watch in the next six months if, you know, what we've thought of as sort of a$20 consumer chatbot subscription, which is, you know, on par right now with Netflix, if that becomes a widespread consumer, you know, cost, but maybe on the cheaper version, right, like the ad-supported version. I don't know if I can predict if Meta's going to be successful becoming more of an enterprise player. I mean, that is something that they've tried before. This is their culture. They try a lot of things. This is the hacker mentality that Zuckerberg has always espoused, and you just see in business cycle after business cycle, which makes covering it fun.

22:52I mean, that's, you know, you'd hate to cover a company that didn't try anything.

22:56Laura Mandaro:Right. Jason, last thought from you. On this idea of them sort of turning into a neocloud, possibly, I mean, I think people get excited when Mark talks about it. If you look, if you break down the way he's talking about it, he's gotten asked this twice, and he basically said, like, it's on the table, which it would kind of be stupid to say otherwise. but he also says we are not doing that now because we think we have enough use for our compute. In fact, they think they need a lot more, obviously. And if we get to the point where we think we've overbuilt and we can't use it, then that would be an option.

23:31So in some ways, it would be a bearish sign. They would get new revenue if they could rent out their compute, but it would be a pretty bearish sign for their position, I think, in the future of AI if they have spent all this money in the hopes that they can leverage this massive compute for their own products, whether they be consumer or some additional enterprise offerings. And then they find out that they can't, so they have to rent the space out. I'm not sure investors will actually see that as positively if and when that time comes as something now.

24:07Laura Mandaro:But could mean more money, more revenue. I mean, you know, at the end of the day, That's where they could go with it. So anyway, great points all around. I want to thank you both for coming on. That is Laura Mandaro and Jason Dean from our San Francisco Bureau here on TITV. Shopify's former CTO, Jean-Michel Lemieux, has a new position advising AI contracts and legal company Spellbook. I want to bring him on to talk about his new role as executive individual contributor and how he's thinking about some of the technical topics in AI today. Jean-Michel, welcome to the show. it's great to have you here the cash great to be here so tell me about what spellbook does and what exactly is your role as executive individual contributor going to be yeah so i mean spellbook basically you know in some way runs the economy you know everything we want to get done around the world has to run through contracts and you know in some way you can say the throughput of the planet is bottlenecked by contracts and you know as an engineer you know you realize that that there's an api layer that has to be in place and and contracts power that so i've been an advisor and investor in spellbook for a while um you know and spellbook had a had a lead um they've been doing this since 2022 um you know so 4 500 companies around the world in 80 countries i think use it so there there's you know i see the potential i think the world needs us now what's interesting is you know as an advisor and an investor i've been getting involved in helping scott who's a ceo in the company um and at some point scott said you know do you want to come help us like things are going well um and he's like i'd love to have you know for you to have the role of cto and i said no i said i don't think that's the right role you need um and the reason why you know and i wasn't just trying to be a you know an ass i was really just trying to say listen i think you know what i'm seeing around the world right now and what i experienced so post shopify actually i started a startup from trash you know i wanted to build i wanted to go back to the gym as you'd say.

26:03And I told Scott, I said, I think the way we're going to win now in the future is it's how the company is going to be shaped. And you see it around the world. People can copy your code and they can vibe code something. I'm like, I think our IP now is going to be how we're going to shape this company for the future.

26:20Laura Mandaro:And what do you mean by shape? So how is that going to be different? Well, I can't give you all our secrets, but I think the idea now is the throughput of the organization is is as important as the product and that throughput is what i wanted to work on and i'll give you like a small example right so if you look at history of like when we've had technology shifts electricity was introduced okay so electricity is introduced 120 years ago and you've got a factory okay so the first thing you do is you you say let's turn on the lights and let's have some electricity and the lights and like that's kind of cool that's interesting now if you went to every employee in that factory and you ask them is your life different they go well it's it's a bit brighter it's kind of cool but you know they wouldn't say i've got a 5x improvement improvement right and then the second thing that happened in the factories is that man we can turn our machines and at the time every factory only had one machine okay let's put electricity into that machine and they're like they got some throughput some cost savings and the third phase is where things actually changed you know people said let's put electricity in every machine in the factory and that's where the factory rearranged itself it rearranged itself and I get a 10x improvement.