OpenAI To Cut Off AI Models For SpaceX-owned Cursor (reuters.com) 29
"OpenAI said on Friday that it plans to stop providing AI models to Cursor, the coding-tool company now owned by Elon Musk's SpaceX," reports Reuters:
"We are making this choice because we cannot be confident that SpaceX will use our technology within our terms of service, based on our experience with Elon Musk's companies violating contracts," OpenAI said in a blog post... However, Michael Truell, a co-founder of Cursor and now an executive at SpaceX, said on X that the company was speaking with the OpenAI team to resolve this issue.
Hours after OpenAI's announcement, Anthropic, which rivals both firms but has an ongoing partnership with SpaceX, said it planned to increase compute to support Claude models in Cursor, according to Tom Brown, one of the co-founders of the firm...
In its blog post, OpenAI said that X, formerly known as Twitter and now part of SpaceX, had breached the terms of its contract with OpenAI after Musk acquired the platform. "To work with a large partner like SpaceX, we typically rely on custom contracts to ensure compliance with our terms of service and that the integration provides for safety at scale," the AI company added.
The article also notes Elon Musk's response Saturday in a post on X.com: "I couldn't care less." He accused Altman and OpenAI President Greg Brockman of being "untrustworthy," while repeating his assertion that they "stole an open source nonprofit."
Hours after OpenAI's announcement, Anthropic, which rivals both firms but has an ongoing partnership with SpaceX, said it planned to increase compute to support Claude models in Cursor, according to Tom Brown, one of the co-founders of the firm...
In its blog post, OpenAI said that X, formerly known as Twitter and now part of SpaceX, had breached the terms of its contract with OpenAI after Musk acquired the platform. "To work with a large partner like SpaceX, we typically rely on custom contracts to ensure compliance with our terms of service and that the integration provides for safety at scale," the AI company added.
The article also notes Elon Musk's response Saturday in a post on X.com: "I couldn't care less." He accused Altman and OpenAI President Greg Brockman of being "untrustworthy," while repeating his assertion that they "stole an open source nonprofit."
Grab your popcorn. (Score:5, Interesting)
This is clearly a corporate "fuck you" move, and both sides of the quarrel are shitheads, but I'm here for it. Insert the "let them fight" meme GIF here.
Re: (Score:1, Interesting)
Re:Grab your popcorn. (Score:5, Insightful)
No, it is siloization happening. We've seen it before and it has been predicted for this market as well. It's just capitalism in action, as much as people want it to be salacious drama.
Re: (Score:2)
Unless they have a means to deal with the massive amounts of industrial espionage in the industry, any silo is purely temporary. I'm here for that too. I think the massive inability to keep secrets is the best defense we have against any one party or small cabal gaining dominance.
Re: (Score:2)
Yeah, it is very much a question if they will be successful in their attempts. But we all know they'll going to try as hard as they can.
Re:Grab your popcorn. (Score:5, Insightful)
The enshittification of Cursor continues.
Maybe this will force them to support more open models.
"Open" AI (Score:4, Insightful)
"Open"AI closes off AI models to a competitor, okay.
Re: (Score:3)
Ed Zitron is a moron who has predicted the past nine out of zero AI crashes, who simultaneously gets paid to hate on AI while also getting paid to promote AI (DoNotPay).
The funny thing is that he openly acknowledges that inference is very profitable (~40% margins) for all parties - he just argues that you should count training. But that argument makes no sense, because if capital tightens across the AI industry, then everyone will cut back on training and scaleup, and... then what? Where's the implosion?
Re: OpenAI is dying and SpaceX is overvalued (Score:2)
"predicted the past nine out of zero"
I may have to steal that one.
Re: (Score:2)
I may have to steal that one.
Its an old and popular economics joke. Origin:
"Wall Street indexes predicted nine out of the last five recessions! And its mistakes were beauties."
- Nobel Prize-winning economist Paul Samuelson. September 19, 1966
Re: (Score:3)
From an economic standpoint, you absolutely have to count training as a facet. If everyone cuts back on scale up, then nvidia, memory vendors, and various other players in that space take a hit. Maybe a troubled AI company can cut out its capex by pausing training, but the downstream impact still is in the broader market.
But even the capex exposure may be hard to shake, given how much debt and purchasing commitments are in play.
Re: (Score:1)
Zitron is correct that the hypserscalers (especially Oracle and Microsoft) & OpenIAI and Anthropic are overexposed to the tune of hundreds of billions of dollars. Your ad hominem is childish - please read the numbers, direct from the hyperscalers and AI companies. https://www.wheresyoured.at/pr... [wheresyoured.at]
Re: (Score:2)
The funny thing is that he openly acknowledges that inference is very profitable (~40% margins) for all parties - he just argues that you should count training.
A very profitable commodity........
But that argument makes no sense, because if capital tightens across the AI industry, then everyone will cut back on training and scaleup, and... then what? Where's the implosion? Everyone is profitable then. You may lose certain players in competition against others (their as-mentioned profitable assets being gobbled up by others), but the industry remains healthy.
Current levels of spending of investor money is unsustainable. Trillions of AI related liabilities have already accumulated on corporate balance sheets. When investors see the writing on the wall and pull back the whole house of cards WILL crumble. The only open question is when. Regarding implosion it's not the fall that gets you. It's the margin call at the bottom.
Re: (Score:1)
The funny thing is that he openly acknowledges that inference is very profitable (~40% margins) for all parties
Yeah this is a straight-up lie. Zitron goes out of his way to show - using OpenAI's own leaked financials that inference is not profitable. Nobody of repute is claiming otherwise, and the hyperscalers, OpenAI, and Anthropic have said mum about it because they know it's not profitable.
Even Torsten Slok, lead economist for Apollo Capital, has shown in black and white that inference is not profitable.
Getting spicy (Score:2)
This may be why they've decided to enter the hardware arena with Jalapeño. [openai.com]
Re: (Score:1)
The chips are interesting, but everyone else can do hardware better than they can and at larger scales. And even with these new and fancy chips the math still doesn't work - inference isn't profitable even then due to the token usage for frontier models. (Take another look and you'll notice there's not Jalepeno benchmarks for OpenAI's own models). There's nothing special about OpenAI - not their models or their hardware. If I'm Microsoft or Meta or Google, I'm happy to let OpenAI hang itself out to dry (fin
Probably an integration question (Score:3)
This is probably not really the end of GPT in cursor for interested people, but the end of GPT as one of the default options. If you subscribe to the (pay as you go) API, you get an URL you can enter into enter OpenAI-API client. Cursor just like almost every other tool are such clients and the only option OpenAI would have to prevent this is to server-side detect the Cursor system prompt, which seems unlikely (but not impossible). I think it is rather "If the user configures it manually, it's the user's fault"
You can still consider if you want to invest into corn. When it pops it's not a bubble but delicious popcorn.
a dishonest man (Score:2)
"Me? I'm dishonest, and a dishonest man you can always trust to be dishonest. Honestly, it's the honest ones you want to watch out for, because you can never predict when they're going to do something incredibly... stupid."
- Captain Jack Sparrow
Good (Score:1)
Unless controlled, reviewed and improved by somebody very competent, very experienced and very careful, AI code is crap anyways. The largest identified (!) risk is now (besides security) that software gets more and more complex and eventually becomes unmaintainable due to complexity and technological debt. Obviously, that creeps up on people over a few years. But the end-result is catastrophic.
Re: (Score:2)
More like 'neutral', Cursor is by no means the only path to vibe code and OpenAI's models are by no means the only models (in fact, generally viewed poorly in those circles). So OpenAI making it harder for Cursor users won't make a dent in the general scheme of things.
Re: (Score:2)
The "good" does not come from the unavailability. The "good" comes from the change, making it a bit more clear that you cannot rely on availability with these tools.
OpenAI? (Score:2)
clown bullshit everywhere all the time (Score:1)
this is going to be such a short period in history.
Re: (Score:2)
"within our terms of service" (Score:2)
OpenAI's First Mover Advantage (Score:1)
Heh (Score:2)
That's what happens when you pay $60B for a VS code fork.