Anthropic co-founder on cutting access to Windsurf • Hacker News Board

Re: Anthropic co-founder on cutting access to Windsurf

#61

Post by pbh101 » Fri, Jun 06, 2025, 2:45 AM UTC

brookst wrote: post #55

princealiiiii wrote: post #8

Any app built on top of these model providers could become a competitor to these providers. Since the model providers are currently in the lowest-margin part of the business, it is likely they will try to expand in to the app layer and start pulling the rug from under these businesses built on top of them. Amazon had a similar tactic, where it would use other sellers on its marketplace to validate market demand for p…

I 100% agree with you except your framing makes it sound like the model providers are doing something wrong. If I spend a ton of money money making the most amazing ceramic dinner plates ever and sell them to distributors for $10 each, and one distributor strikes gold in a market selling them at $100/plate, despite adding no value beyond distribution… hell yeah I’m cutting them off and selling direct. I don’t really…

I don’t think it is at all clear that Windsurf adds zero value. Why do you think this is a helpful analogy?

Re: Anthropic co-founder on cutting access to Windsurf

#62

Post by brookst » Fri, Jun 06, 2025, 2:45 AM UTC

selcuka wrote: post #51

artdigital wrote: post #41

Earlier quoted context omitted.

> This feels like a cheap trick to drive users towards Claude Code How did you come to this conclusion? It’s very much like he remarked: OpenAI acquired Windsurf, OpenAI is Anthropics direct competitor. It doesn’t make strategic sense to sell Claude to OpenAI. OpenAI could train against Claude weights, or OpenAI can cut out Anthropic at any moment to push their own models. The partnership isn’t long lasting so it doe…

> OpenAI could train against Claude weights OpenAI can always buy a Claude API subscription with a credit card if they want to train something. This change only prevents the Windsurf product from offering Claude APIs to their customers.

Other than, you know, terms and contracts.

Re: Anthropic co-founder on cutting access to Windsurf

#63

Post by pbh101 » Fri, Jun 06, 2025, 2:45 AM UTC

SoftTalker wrote: post #60

selcuka wrote: post #47

Earlier quoted context omitted.

True. Otherwise Anthropic would cut access to other code assistants too as they all compete with Claude Code.

They might still. Why not? Illustrates a risk of building a product with these AI coding tools. If your developers don't know how to build applications without using AI, then you're at the mercy of the AI companies. You might come to work one day and find that accidentally or deliberately or as the result of a merger or acquisition that the tools you use are suddenly gone.

This is true of any SaaS vendor

Re: Anthropic co-founder on cutting access to Windsurf

#64

Post by brookst » Fri, Jun 06, 2025, 2:49 AM UTC

wincy wrote: post #34

Wow excited for when Anthropic buys Cursor in a couple months and I get locked out of using OpenAI models with it. This is depressing. I just want stuff to work.

Friendly advice: you will be happier if you defer reacting to things until they are actually real.

We can all imagine all sorts or terrible futures. Many of us do. But there is no upside in being prematurely outraged.

Re: Anthropic co-founder on cutting access to Windsurf

#65

Post by osigurdson » Fri, Jun 06, 2025, 2:51 AM UTC

pbh101 wrote: post #61

brookst wrote: post #55

Earlier quoted context omitted.

I 100% agree with you except your framing makes it sound like the model providers are doing something wrong. If I spend a ton of money money making the most amazing ceramic dinner plates ever and sell them to distributors for $10 each, and one distributor strikes gold in a market selling them at $100/plate, despite adding no value beyond distribution… hell yeah I’m cutting them off and selling direct. I don’t really…

I don’t think it is at all clear that Windsurf adds zero value. Why do you think this is a helpful analogy?

The analogy is a bit like this. Imagine that there are 100 ceramic dinner plates for $6 each. Now someone comes in and buys them from you for $5 each - undercutting your margin. Then a 3rd company comes in and literally eats your lunch on your own ceramic dinner plates. The moral of the story is any story involving ceramic dinner plates is a good one, regardless of the utility of any analogy.

Re: Anthropic co-founder on cutting access to Windsurf

#66

Post by ummonk » Fri, Jun 06, 2025, 2:51 AM UTC

what wrote: post #53

jsnell wrote: post #30

Earlier quoted context omitted.

Yes, many people believe that, but it doesn't seem to be an evidence-based belief. I've written about this in some detail[0][1] before. But since just linking to one's own writing is a bit gauche and doesn't make for a good discussion, I'll summarize :) 1. There is no point in providing paid APIs at negative margins, since there's no platform power in having a larger paid API share (paid access can't be used for trai…

There are more factors to cost than just the raw compute to provide inference. They can’t just fire everyone and continue to operate while paying just the compute cost. They also can’t stop training new models. The actual cost is much more than the compute for inference.

That's all the more reason to run at a positive margin though - why shovel money into taking a loss on inference when you need to spend money on R&D?

Re: Anthropic co-founder on cutting access to Windsurf

#67

Post by huxley » Fri, Jun 06, 2025, 2:54 AM UTC

artdigital wrote: post #41

HyprMusic wrote: post #37

This feels like a cheap trick to drive users towards Claude Code. It's likely no coincidence that this happened at the same time they announced subscription access to Claude Code. The Windsurf team repeatedly stated that they're running at a loss so all this seems to have achieved is giving OpenAI an excuse to cut their 3rd party costs and drive more Windsurf users towards their own models.

> This feels like a cheap trick to drive users towards Claude Code How did you come to this conclusion? It’s very much like he remarked: OpenAI acquired Windsurf, OpenAI is Anthropics direct competitor. It doesn’t make strategic sense to sell Claude to OpenAI. OpenAI could train against Claude weights, or OpenAI can cut out Anthropic at any moment to push their own models. The partnership isn’t long lasting so it doe…

Totally irrelevant, Anthropic isn’t cutting off OpenAI, it is cutting off Windsurf users.

Re: Anthropic co-founder on cutting access to Windsurf

#68

Post by jsnell » Fri, Jun 06, 2025, 2:54 AM UTC

what wrote: post #53

jsnell wrote: post #30

Earlier quoted context omitted.

Yes, many people believe that, but it doesn't seem to be an evidence-based belief. I've written about this in some detail[0][1] before. But since just linking to one's own writing is a bit gauche and doesn't make for a good discussion, I'll summarize :) 1. There is no point in providing paid APIs at negative margins, since there's no platform power in having a larger paid API share (paid access can't be used for trai…

There are more factors to cost than just the raw compute to provide inference. They can’t just fire everyone and continue to operate while paying just the compute cost. They also can’t stop training new models. The actual cost is much more than the compute for inference.

Yes, there are some additional operating costs, but they're really marginal compared to the cost of the compute. Your suggestion was personnel: Anthropic is reportedly on a run-rate of $3B with O(1k) employees, most of whom aren't directly doing ops. Likewise they also have to pay for non-compute infra, but it is a rounding error.

Training is a fixed cost, not a variable cost. My initial comment was on the unit economics, so fixed costs don't matter. But including the full training costs doesn't actually change the math that much as far as I can tell for any of the popular models. E.g. the alleged leaked OpenAI financials for 2024 projected $4B spent on inference, $3B on training. And the inference workloads are currently growing insanely fast, meaning the training gets amortized over a larger volume of inference (e.g. Google showed a graph of their inference volume at Google I/O -- 50x growth in a year, now at 480T tokens / month[0])

[0] https://blog.google/technology/ai/io-2025-keynote/

Re: Anthropic co-founder on cutting access to Windsurf

#69

Post by paxys » Fri, Jun 06, 2025, 2:59 AM UTC

wincy wrote: post #34

Wow excited for when Anthropic buys Cursor in a couple months and I get locked out of using OpenAI models with it. This is depressing. I just want stuff to work.

If LLMs were a sustainable business then Anthropic would have no problem selling Claude to a competitor. Heck they'd brag about it ("see OpenAI uses our models to code instead of their own!"). You see this in the industry all the time. Large tech companies compete with each other and sue each other and have deep business relationships at the same time.

What this change really says is that Anthropic doesn't want to burn VC money to help a competitor. And that is the reality of "I just want stuff to work". It won't just work because there's no stable business underneath it. Until that business can be found things will keep changing and get shittier.

Re: Anthropic co-founder on cutting access to Windsurf

#70

Post by lmeyerov » Fri, Jun 06, 2025, 3:00 AM UTC

jsnell wrote: post #30

raincole wrote: post #19

Earlier quoted context omitted.

> The model providers are not in the low margin part of the business. Many people believe that model providers are running at negative margin. (I don't know how true it is.)

Yes, many people believe that, but it doesn't seem to be an evidence-based belief. I've written about this in some detail[0][1] before. But since just linking to one's own writing is a bit gauche and doesn't make for a good discussion, I'll summarize :) 1. There is no point in providing paid APIs at negative margins, since there's no platform power in having a larger paid API share (paid access can't be used for trai…

I think you miss 2 big aspects:

1. High volume providers get efficiencies that low volume do not. It comes from both more workload giving more optimization opportunities, and staffing to do better engineering to begin with. The result is break even for lower volume firms is profitable for higher volume, and as high volume is magnitudes more scale, this quickly pays for many people. By being the high-volume API, this game can be played. If they choose not to bother, it is likely because strategic views on opportunity cost, not inability.

That's not even the interesting analysis, which is what the real stock value is, or whatever corp structure scheme they're doing nowadays:

2. Growth for growths sake. Uber was exactly this kind of growth-at-all-costs play, going more into debt with every customer and fundraise. My understanding is they were able to tame costs and find side businesses (delivery, ...), with the threat becoming more about category shift of self-driving. By having the channel, they could be the one to monetize as that got figured out better.

Whether tokens or something else becomes what is charged for at the profit layers (with breakeven tokens as cost of business), or subsidization ends and competitive pricing dominates, being the user interface to chat and the API interface to devs gives them channel. Historically, it is a lot of hubris to believe channel is worthless, and especially in an era of fast cloning.