It's ironic how AI companies are re-inventing their own form of privately enforced copyright, lobby the government to ban foreign competitors that don't respect it etc., all while spending the last 5 years fighting tooth and nail against the copyright of the training material they're using.
If you can take any book and turn it into a model, because it's "transformative enough", and "AI learns just like a person does", then surely a model distilling another model is transformative and fair use.
They tied themselves into knots fighting the letter of the law, and now, when they need the spirit of the law - that each creator deserves protection for their work - now we devolve to the law of the jungle. Maybe we'll even see LLM book curses, the way medieval scribes damned book thieves to blindness and worms.
It's one of the things you need to do if you want your company to later become the only company in the world. They also promised that they would treat you nicely afterwards after they get what they want.
I am pretty sure that the goal of these companies at the end is ruling more than what governments can.
Just look at the pattern... they collude, they provide to them whatever is needed, and at some point, if this is not true yet, they will be the ones who will tell them what to do or not, bc you know how humans are, right... blackmailing, mess up the business or shames of people, etc.
It is just a matter of time. The state, as we know it, will collapse or will be greatly reduced, which, from a point of view, is positive, from another, Idk, bc if someone replaces that, we will be in the hands of someone, as usual...
> If you can take any book and turn it into a model, because it's "transformative enough", and "AI learns just like a person does", then surely a model distilling another model is transformative and fair use.
I mean it quite likely is sort of in the same legal bucket. It won’t stop them suing but it is going to be the legal equivalent of two biologically-related warlords making their champions fight with their hands tied for sport.
I'm increasingly convinced there's going to be a technological AIpocalypse within the next five years which makes all of these issues - and many others - redundant.
ChatGPT 3 was released nearly six years ago, and the models are staging increasingly aggressive breakouts now. Where are they going to be by 2030?
Your link doesn't suggest a violation of TOS, but rather a policy shift disallowing their major competitors to use their model for any use.
That thread you linked suggested they were using it with cursor as part of their development process, not for distilling.
>"Hi team, I believe many of you have already discovered that anthropic models are not responding on cursor. According to cursor this is a new policy anthropic is enforcing for all its major competitors."
LOL, obviously Space-cursor-xAI-musk-imperial-holdings-inc is providing them compute. They either _can't_ cut them off, or they know not to bite the hand that feeds them.
Yes Anthropic’s chief compute officer said nice things about one of Anthropic’s chief providers of compute. Recall that compute constraints are one of their existential risks.
Because anthropic is beholden to SpaceX for the time being due to their compute restraints.
Its actually very tone deaf to even pipe up and say anything about this given they did this already to Windsurf themselves, and they previously banned x.ai employees as well from using their models for basically the same thing Elon admitted they were doing with OpenAI.
I'm surprised they decided to say anything at all given their rocky past. Its the pinnacle of being a hypocrite.
God I hope so. My workplace enforces Cursor use, and to see it lose GPT and Claude would absolutely make my day. Much has been invested in getting us to "use the right models", a list in which Grok does not feature.
On the one hand, I'd love to see Cursor/Grok/xAI/SpaceX/Musk disappear forever. On the frustrating other hand, Cursor is the absolute best LLM IDE by such a ridiculously wide margin. Everything else feels so broken and ramshackle in comparison.
My employer let us evaulate all the big providers for a month each before having a vote on which one we'd like to keep. Cursor won almost unanimously. (This was before the xAI takeover was even on the horizon, otherwise I would've voted differently myself.)
i'll be surprised if anthropic doesnt follow. cursors biggest advantage has been giving users a polished wrapper around whichever frontier model they prefer
Why you would trust Elon Musk to not hack you and steal your model weights is beyond me, but maybe they intend to use that compute just for training, still I’m sure there’s plenty of research there to take too…
what's the moat for AI labs like OpenAI and Antrhopic as they seek trillion IPOs. a lot of people said Chinese models also distill US LLM models. if it is this easy to do. how do US AI labs justify asking for trillion?
Anthropic wants you to think that models that distill from them would be worthless otherwise. It’s not true, it’s just one part of the process. The whole narrative that Chinese and other models are only good because they distill is nonsense.
Is there some kind of alternative to Cursor in these terms
- fast because the code is already indexed no need to start every prompt rediscovering the world with rg
- in editor/vscode review much better than relying on git diff at the very end. You can click on next button to jump there. Quick edit in place. Or a quick prompt (seconds) to fix something small
- in editor completion / quick edits, when you do things by hand.
It seems that everyone loves agentic Claude code things these days but I don’t understand how you can review what it did and remain as much in the flow as you do with cursor.
Also I find that incredibly slow. By the time Claude finished a prompt I could have done a few in cursor.
And mind you, I’ve been wanting to use zed instead but the clunky vscode fork called cursor. But cursor has just a better UX for AI integration even though it’s slow and guzzle up RAM.
> It seems that everyone loves agentic Claude code things these days but I don’t understand how you can review what it did and remain as much in the flow as you do with cursor.
The way I review it is to let it finish a PR sized slice and then I review the whole diff in a separate terminal as if it’s a PR from someone else. Then I very impolitely list the shit that needs fixing, let the LLM grind out a new batch and repeat the process. When the code is good enough I throw it into a commit and push a PR out for other humans to review.
Clicking around in a UI would just waste time and if the change is too large to review in a terminal window the change is too large for another human to review.
There is a step missing. Impolitely list what's wrong and try to incorporate any persistent themes into a markdown file so that the agent stops making the same mistakes. Put any rules into linters and static analysis and sanitizers and ...
I sometimes see people reviewing LLM code leaving the same review comments multiple times, or worse, after some time they just give up reviewing.
I just switched back to VS Code from Cursor a week ago. Things seem great and VS Code has caught up. Last time I tried switching back was in November and VS Code was not ready then.
CLI is great for most things. Some times, when I am reviewing a PR, I just tell Codex to do a diff between source and target branches in PR, and generate a HTML report. Then I read that, and also review relevant pieces of code in Zed.
I tried using Cursor, many of my colleagues use it, but stopped using it due to clunkiness. And of late, even in there, many of my colleagues use agentic coding. Instead, I prefer the snapiness of CLI (Codex coded in rust, takes less memory per session) and Zed (also coded in rust,) and blazingly fast. Zed is the new Sublime Text if you will.
I just retried Zed and yeah, it's much much better than last time I gave it a shot!!!
I turned on the review stuff, and yeah I can do more or less what I could in Cursor. Check something. Do a quick inline promtp for the three lines I dislike and move on.
And of course the editor is so much nimbler/nicer.
No plan mode, but for big edits I'd probably use codex or something just like you. If you have any advice for integrating for that flow, I'll take them.
I guess whatever advantage cursor has, it will die once "Sol" is no longer integrated, as it was my favorite heavy work model. But I sure will miss the snappy auto/composer stack of Cursor.
I love VSCode built in LLM harness. No extensions, use any model you want, amazing UX for reviews, control of agents. The sandboxing also seems great, I never see models touching anything they shouldn't.
Claude / Codex ex tensions inside of VSCode give you what you described. I ageee, I find people that use only the CLI barely read and review the code the agent creates until it's too late
Lately I've been developing in Dart/Flutter. I use Android Studio but I know developers who use Antigravity and like it.
For the past three or four months it's performed very well, even the cheaper Gemini models. No serious complaints. But if you go back a year it was pretty terrible. My work style is to keep the coding agent on a short leash so YMMV.
Cursor seems to have intentionally dropped semantic/embedding-based codebase indexing. Their explanation is that newer coding agents are now good enough at searching repos directly. They can grep in parallel, inspect directories, read likely files, and refine the search themselves. Cursor says this now works as well or better for most cases. [1]
So the architecture shifted from something like “chunk → embed → vector DB → retrieve” to “local search index/ripgrep → agent explores.” There is still an index, but it’s a conventional local search index. No embeddings are involved in that retrieval path. [1]
The interesting part is that Cursor was still reporting a ~12.5% accuracy gain from semantic indexing in January 2026. [2] So this looks less like “semantic search never worked” and more like agentic search improving fast enough that the extra infrastructure stopped being worth it.
Interesting to see someone else is in the same boat as I am. Regarding editor review I agree, surely the solution would be to `git diff` at the end, but if you only want to git diff what was literally just changed by the agent it means you'd have to commit as you go so that your git diffs are more localised. And then at the end of your work just squash the commits. But it feels annoying. (EDIT: ok I forgot about git staging, that seems to be better)
I mainly use Cursor for another reason though: all of my work is remote over SSH and I don't like keystroke latency, so I prefer code editing / agent conversations to feel local. (Everything else is done in the terminal.) But I would also prefer to use something less propriety. I may go back to VSCode with plugins to achieve a similar experience.
How does Zed compare to the tab completion in Cursor ("Cursor-tab")? It's honestly the only reason I have stuck with cursor in spite of the UI/UX being worse every single time I update it. Can I get something equivalent or better from Zed or some vscode fork/extension now? Admittedly I haven't looked in a bit, but it's seemed like nobody really cared about this workflow when I've tried in the past.
I do like Zed as an editor and I keep trying it out
but the the review what the AI did by jumping directly into the file with accept/reject/go next/next file with the ability to edit in place was nowhere near as good in zed vs cursor last I checked.
Also: the ability to draft a plan and monitor its execution, review it, and merge it from a cell phone. I know Claude remote is a thing, but it’s not the same.
This is a big quality of life thing. Sure, I should just go on a walk without having an agent going, but oftentimes it’s not intruding on it - it’s the thing that actually ENABLES it.
> I don’t understand how you can review what it did and remain as much in the flow
Separate terminal tab with git diff/show, Zed for edits - rarely, there is little reason to change code yourself. Opus at low effort is faster than you can switch windows.
May be opencode2 in the near future. Sample (not affiliated) ~
1. download opencode2
2. settings > tabs > vertical
3. add worktree setup to agents.md
4. use /diff to review changes
5. use commands for commit, sync and push
I don’t see how opencode 2 (or 1) addresses any of my concerns. I thought it might avoid rg/grep the code base on every prompt but does not even seem to be the case.
I met the founder of Voidleap (https://voidleap.com) yesterday! They have been building a new type of code editor since early this year and has been dogfooding it themselves.
I haven't tried it yet but it looks promising. He would be thrilled if people give it a go!
I don't like being that guy, but wow, the AI slop copy on that website is so terrible, it's almost impossible to read. Every other sentence means absolutely nothing.
Slop website indicates vibe-coded software, and I wouldn't let anything like that even look at my code.
I use codex all the time. It’s inferior to cursor for the two reasons mentioned. It rediscovers the world with rg (no indexing). Wall of diff at the end.
The GUI slightly better than Claude. And it’s a million times cheaper than Claude.
I used zed with codex for the accept / reject code review feature without relying on git but yesterday the Read Only drop-down disappeared and while I was doing my code review the AI agent started vibe coding without my permission. I hate this industry.
In the case of TUIs I’m going to assume you review the code externally in neovim or lazy git or something like that?
I don’t like relying on GIT for that and making things as good/done as I go. But with these agentic tools that’s what I have to do. Stage everything I like, manually edits a few bits.
But what do you when you find something where you want to say the agent: I don’t like this for reasons x and y redo as z. I’m like one shortcut away in context inside cursor to say that to my agent? You have to change tab in herdr and then what? Type the file name and line numbers? Is there a way to do this that does not seem very frustrating?
Also open code, codex, etc. super slow. They grep the code. Every. Single. Time. Cursor has a semantic index
It's always been clear to me that Cursor's business model of reselling others' APIs had its days numbered. Not necessarily because the providers would pull the plug, but because you wouldn't be able to compete with subsidized plans.
Cursor is already kind of useless for third party models unless you're willing to spend thousands of dollars. It's only worth it if you're going to use mostly grok/composer.
The TUI version of Claude Code just isn't very good, though. The default keybindings do not work in the majority of terminals (because they generate the same byte sequences for key sequences assigned to different commands). Editing large prompts is rather painful because there doesn't seem to be in-prompt search commands. There is no source code browser, so source code references cannot checked by the users without a separate tool. Diffs cannot be copied directly because they are indented by spaces (even though some tools strip indentation from patches, it's still ugly). Switching between conversations is somewhat discouraged by the interface, which makes it harder to keep the context clean.
I think it's possible to have a better TUI, with something that isn't modeled after a shell prompt. Apparently, recent updates move in that direction. The GUI version of Claude Code seems to be coming to Linux, too.
The tide is turning though. I was using the old non agent view in cursor every day for about a year when I built my startup with it. Then Claude Code blew it out of the water. But since two month or so I much prefer Codex and Cursor with the agent view.
However, what I see cooking at Cursor is way more promising than the others. Things like full system understanding, their own forge, multi-repo support. I'd put them as way more visionary in the Gartner magic Quadrant ;)
I saw that happen culturally. But never saw that much different in performance. And I really dislike terminal development. And their desktop app is really buggy.
Why do people keep insisting that subscription plans are subsidized? They are not. API prices are outrageous, designed to milk enterprise users. They probably have an 80-90% profit margin on subscription plans.
Inference is cheap; Dax once said on a podcast that Opencode has a close to 90% profit margin on the openweight models it provides inference for (at 10x cheaper pricing). Those models are close in size and spec to models from large labs.
OpenAI leaked financial data (2025) suggests this - it has $5.7 billion for marketing which 44% of revenue. It’s hard to explain where these billions go other than into subsidised subscription plans and a free tier.
Because using chinese models on OpenRouter is way more expensive than the equivalent codex sub, unless you use something clearly inferior like deepseek flash.
If the Opus/Fable harness is superior and it's written using Opus/Fable then it will cease to be a moat. Otherwise the premise for coding using LLMs is essentially untrue.
cursor adds $0.25/M to your token bill for 3rd party services. sounds small but it's insanely big on cached inputs, which are an insanely high % of use
Honestly, Cursor's AI harness worked best with the Anthropic models. Could just be my bias though. Evaluating frontier LLM's all seem to be by vibes. I think the industry is understanding it's both the harness and the model that makes the difference; can't half ass either, and some harnesses do better with different models. I'm good to leave this "career." Senior software engineers, make your bag.
I may be a rare Cursor user around here, but I am saddened by this news. The thing that keeps me using it is precisely the ability to switch between models: I usually do brainstorm/planning with Sol or Terra, then let Composer (free) deal with the implementation. For certain tasks I use Claude, which tends to be more expensive but sometimes saves me time.
Using both OpenAI and Anthropic models, Cursor is cheaper than having two separate subscriptions, and their free tier of Composer is good enough. I will now have to reconsider my options.
JetBrains is a good option - works with any model/agent. Also they have very good semantic search for the codebase, which is exposed via MCP to the agents.
currenly I use pi with openrouter and is very nice I combine with a chatgpt sub that allows me to use it with pi. and when my usage runs out I turn to openrouter. Then same when I need anthropic models
If you can explain how to make compaction not suck in opencode or Kilo Code, I would love to know how. I am really trying to make use of other harnesses but that's a major sticking point.
While this isn't great for users it seems like a pretty standard circling of the wagons in preparation for the next phase in the battle for frontier AI dominance.
Ironically, as a Cursor and Claude subscriber, but not an OpenAI subscriber, this will push me back to Anthropic. I don't feel good about giving OpenAI money directly, but I've found switching to GPT 5.6 Sol now and again in Cursor to be useful. Now I'll probably just forget about OpenAI models.
It’s OpenAI (and Anthropic I’m sure) realising that they’ve somehow ended up stuck in the lowest margin part of the stack (providing inference), and trying to get out of it
Honestly, I've been a Cursor subscriber for a while and it was just momentum. Still, I rank OpenAI as the worst actor among the big AI companies. I think OpenAI's playbook will be a repeat of Facebook. Users are the product x 1000.
Cursor/SpaceX/Grok is a different beast. I think they will cheat other companies but I'm not convinced yet that they have it out for users. We'll see. I'm certainly not some kind of Cursor lifer or something.
I don't see how it makes sense to use models outside of Grok or Composer in cursor unless you like the tooling to such degree that you're willing to pay out the nose for access to OpenAI and Anthropic to use with that tooling. And Codex is regularly scoring towards the top among harnesses, so you might as well use the best harness available for the given model. I'm happy with Grok and Composer in Cursor. If Cursor wants to add more models, they should host more open models.
The reason it makes sense is because different models are better at different things. Allowing yourself to be locked in to a particular model by these companies is a bad idea and will lead to worse outcomes for yourself and the market as a whole.
You’d use it because until recently Cursor had the best setup for cloud agents and visual verification and a bunch of other features outside of the harness.
What a shame for Cursor. I found it a great tool - I actually liked Composer 2.5; but now being bought by SpaceX has for a while left a bad taste - and they keep pushing Grok and now GrokBot.
I took the 50% discount instead of cancelling - but with this I'll likely not renew.
So sad to watch them go from great to mediocre in such a short time.
If you have such a world changing technology, why sell an API when you can capture all the value yourself? That's what every provider is asking themselves.
Cursor will be fine after this, as they have their own first party models and can potentially use open weight models via API or their own hardware, which have gotten good enough to Opus and Fable quality now for many use cases.
Being a pure play model provider is completely untenable right now. Coding agents own your demand; Nvidia, data centers, and DRAM manufacturers own your supply. You're getting squeezed by open weight models. Your only bet to making money off this thing is by locking your model down to your program to have some semblance of a moat. The fact that businesses can change providers with just the change of an API key is unsustainable. You have neither the "hard to switch" lock-in of a traditional software company nor do you have the near-zero marginal cost of production they have. Each token of inference costs you money, not to mention the training costs. At least locking the models down to your harness lets you get some sort of lock-in.
Lots of enterprises don’t want to use OpenAI or Anthropic directly. They will go via AWS bedrock or google’s equivalent.
The model providers are in a difficult position. With an added danger of open weight models that are within reach of large enterprises running themselves.
Not really a surprise given their own models are the default option? Most people wouldn't change model. And if you want to use Claude or GPT you don't really have a reason to use Cursor over Claude Code or Codex
I use Cursor at work and love it. It's a great IDE, the best IMHO. But it becomes really expensive.
Since the introduction of the Cursor Token Rate, my company tells us to use Claude and Codex directly with Anthropic (or Vertex AI) and OpenAI, respectively.
I'm now using Grok 4.6 most of the time. Sonnet some times. But that's it. Cursor-native models 99% of the time. I'm pretty sure Anthropic will follow OpenAI soon in pulling their models out of Cursor because it's now a true competitor.
It's insane that Cursor charges 25 extra cents per million tokens on a model that charges 2 cents per million cached tokens. Basically everyone is forced to use Cursor's own models at this point as otherwise the costs are simply too high.
> As AI capabilities advance, we also have a new level of accountability to ensure our upcoming model, Astra, is being used in accordance with our terms.
I think Sam Altman has played a clever business move here, they can use this excuse immediately after the Hugging Face leaks and it's not going to be dismissed willy-nilly. Whereas long term they probably wanted to end their partnership with Cursor anyway. I bet Dario will be happy about this too, but maybe the compute deal prevents him from doing similar anytime soon.
Is it really an excuse? As the release points out, musk admitted under oath to distilling OpenAI’s models to produce xai models on an ongoing basis. Whatever you think about goose for the gander wrt IP theft, I wouldn’t voluntarily do business with a competitor who was actively and brazenly flouting my terms, would you?
Yes, any Cursor sub includes a separate tier of usage for Cursor/SpaceX first party models separate from usage of third party models.
I have a $20 Cursor sub still personally and would say that the usage for Grok models is much higher than what $20 gets you with Anthropic. I can't speak for the $100 plan but I would assume it scales similarly.
I rarely run out of my usage monthly even with the $20 plan.
It's a little confusing because Grok and Cursor kinda live in different worlds. Grok has SuperGrok and SuperGrok Heavy ($300 a month!)
The best bet is Cursor Ultra ($200), which is the equivalent a month plan that you get from OpenAI and Anthropic. I get a LOT more capacity out of it because I almost always use the "Auto" mode, which auto picks between Opus 5, GPT 5.6, Grok 4.6, and Composer 2.5.
I like Auto mode so much actually. I use it for my Cursor work account as well because it's about 10x faster than just using Opus 5 or GPT 5.6 and about 90% is good. I switch back to the bigger models if I get in trouble, but that's pretty rare.
Vibe coding requires some discipline. You need understand what it's trying to do. If you have a good mental model you can keep it out of trouble. If you don't, it doesn't matter how good the model is. You're gonna end up in a mess.
Yes it does. You can either subscribe to grok directly or cursor.
You have many tiers and all of them have generous usage.
Grok is a workhorse, not the smartest yet, but produces good code and follows the plan.
Grok 5, which has been training since beginnig of this year, will probably be a monster. (10T/6T Param model variants)
It makes no sense that a contract would even provide a maximum notice period. Why would parties put that in? What benefit would either derive from restricting themselves to not getting earlier notice.
I read it as: we acted as quickly as we could once it became possible to do so, i.e., the change in control was completed. not that their special agreement carved out some maximum notice term.
Everyone is playing hardball. No one wants to be IBM in an IBM-Microsoft dance.
But this is kind of like banning GPU exports to China. It will encourage them to develop their own technology. Yes, it will take them longer, but in the long-run China will be more successful with their own GPU tech.
So now you know why Trump is called "川建国 (Chuan Jianguo)" in China — literally "Trump who builds China" — because he really is building China and making it better (just kidding).
> As AI capabilities advance, we also have a new level of accountability to ensure our upcoming model, Astra, is being used in accordance with our terms.
This move seems to protect Astra against distillation. I wonder whether Astra will be released before or after 11/12.
> We know that the people most affected by this decision are the developers who rely on OpenAI models in Cursor. We care about their experience in this transition and we’re ready to go above and beyond to support them.
I really get sick of this anti-Elon rhetoric around here. Can we not keep it to technical merits and not get sucked up into politics? The guy's clearly nuts, but he's also done an incredible amount of amazing things.
This isn't a technical article, though. It's about contracts, so one party's self-admitted habit of ignoring contractual obligations seems very relevant and not a random ad hominem.
I trust them more than someone who took a nonprofit designed to be open and had tens of millions in donations to do just that, and then turned into a for profit who's cuddled up tight with Microsoft and is so closed they're preemptively ending contracts with someone under pretty flimsy accusations. And doubly funny it's the person who's responsible for them even existing. And from whom they stole tens of millions
Does anyone have any recommendations for something like Cursor but more open? Cursor's harness is so much faster than anything else I've used that I just can't give it up, using Claude Code feels like being on 0.25x speed. Also I do like the IDE-style UI, being able to highlight code in a file editor and then quickly add that to my chat (Cmd-L) is really nice and much more difficult in CLI based harnesses.
Right now I think GPT 5.6 Luna is by far the most cost effective model, so this change is effectively kicking me off of Cursor.
This, and the keep/undo flow, forking agent sessions (including renaming them)... The UX could be better at places but the core features of what Cursor has (for someone who came from VS Code) is an unbeatable mix. Really would like to find something else.
I did try Kiro which seems to have the same philosophy but it's just not there. Most of the features I like in Cursor are missing.
Some other commenters mentioned that VS Code + Copilot would be better these days. Dunno.
I am also wondering the same thing as OP. Somewhat tangential thought, but I recently realised that what I would really love in an IDE is a "first class" Markdown editing experience, just because most of my day-to-day is managing and writing docs rather than editing code. Cursor has Markdown rendering support of course but it seems to have two implementations, each sort of clunky in their own way. And I want something more WYSIWYG but with Latex / flowchart / etc. support built-in.
OpenKnowledge seems to maybe fill that niche somewhat (it also supports annotating Markdown with comments to feed into an agent), and has agent integration within the GUI, but one killer feature that seems to be missing is Remote SSH support (short of hosting its web server on the remote and having the IDE live in the browser).
It was honestly just a matter of time, given all of the bad blood there is between Sam and Elon and Elon basically admitting under oath they were distilling other companies models.
I wonder if Anthropic will follow suit given their campaign against distillation or if they'll stay quiet given their reliance on SpaceX (among others) for hardware.
I'm not impressed with cursor and it's decidedly not worth 1 mil in terms of its engineering. Perhaps it's user base and their loyalty makes up the vast majority of its evaluation. But to me it's ridiculous that a dev would use cursor or think that it's indispensable to them
Honestly, can't say I miss my Cursor subscription at all, and I'm surprised they're still a going concern. Why would I want to use a proprietary VSCode fork with an identity crisis when I can use literally any editor with any agent and have about the same experience?
I use Claude Code (both Desktop and CLI) all day every day (as well as OpenCode with DeepSeek V4 among others). For non-mission-critical stuff (small internal tools, anything that doesn't touch the client and doesn't handle critical data or infra), I'm on Auto (sandboxed in the working directory). For everything else I approve every change, meaning I review every code change the model proposes. Plan mode is great to write the architecture as well, as it doesn't touch any file, doesn't write anything, and lets me iterate until the arch is perfect and the proposed high-level code (modules etc) is solid. I haven't used an editor in over 4 years, apart from writing specs.
We have a subscription at work. They have a CLI agent that works pretty well. Lately I mostly use it with grok, which is actually quite a capable model.
How so? A dozen very capable models out there. And the gap between frontier models from the two main US labs kept narrowing, at this point non existent.
The quality/price ratio is clearly superior when running Qwen, GLM, or Deepseek.
To make a product more exclusive won't hide the fact that product is becoming more and more irrelevant.
Most people do not enjoy the output of the open models or know where their code is being sent so I’m out. Also I don’t trust Elon anyway, I’m just going to make a VS Code Extension that handles things how I want but I’d like to offer the same deals Cursor got for tokens with OpenAI and Anthropic, for me this was the key 1/4 token cost!
Not surprised, but honestly I don't feel like this is a huge loss for cursor. Models are so commoditized, and 99% of users don't need particularly fancy models. Between existing anthropic models, composer-2, honestly even grok models, and also the fact that many people just use 'auto' mode, I feel like many won't feel the impact for OpenAI's withdrawal.
Maybe if Claude joined along it would be more impactful. For better or worse though, Xai seems like its lining themselves up to handle being on their own for cursor.
I think there is still a huge market for models that are specialised in software development (system design, infrastructure, coding, etc.).
If a company can provide the access to such a model via AWS Bedrock or similar enterprise compliant platform with a decent IDE/harness they can capture billions of dollars of market.
The closest we have to this is Claude Code but I think it can even be better and specialised. Cursor has an enormous amount of data on coding and developer usage and maybe they should use that to build a proper platform
Musk was forced to admit under oath during his OpenAI lawsuit that xAI distilled GPT models to train Grok. No reason to believe he won't use Cursor data for the same thing, if it isn't happening already.
In a sane world would this be a potential antitrust case as they are a direct competitor to Codex?
How can you ban a competitor from using a model that you openly provide access to?
I understand that the OpenAI ToS may have been violated by X, but has anyone ruled on distillation as fair use? OpenAI believes distillation of nonAI produced content is fair use, so they don’t appear to have a leg to stand on.
In a sane world, one should be able to choose not to do business with a company for any reason
there are plenty of alternatives, 100s, so I don't think antitrust is on the table, think of this more like a boycott of the world's richest man for being a twat
"You are not allowed to use our services to develop a competing product" is a standard clause in pretty much every contract of this nature. There's nothing antitrust about it. Companies are allowed to say no to bad deals.
My guess is that OpenAI now has a full blown Cursor equivalent in development (an expansion of Codex?) and will announce it in September. Otherwise it stands to lose those who are used to working with Cursor.
Yeah, the "go above and beyond for our users" implies they have a "better" version of Cursor coming. Probably integrating their new chip/microphone based product.
Cursor is like one of the least economical ways to use Claude and OpenAI, virtually nobody does it. You use Cursor for their autocomplete/composer + grok. For using Claude and Codex, you get immeasurably more bang for your buck just using them directly.
All good points. However, it's not a safe assumption that everyone operates that way. People probably use the flagship models because that's what they know. Sure, some probably just use the default, in this case Grok.
Also, it's a signal more than anything. OpenAI pulls out of Cursor. Then Anthropic and Google does as well. Then you no longer have "Cursor" you have Grok.
Grok doesn't have the best reputation in general, regardless if you're tech savvy or not.
Great. Sam Altman sold us all out to venture capital. So I won’t be using ChatGPT anymore. Also, I noticed that ChatGPT lies more than any other AI of all the ones I’ve tested.
I used cursor extensively both at work and for personal use. Its harness is pretty darn good but I basically never used OpenAI models in it.
The first party cursor models (initially composer and now grok-cursor) and price very competitively for their quality and have very generous usage limits included in the base plans, and then per token pricing afterwards.
Occasionally I’ve used some of the Anthropic models in it but found their cost massively outweighed the benefits for the type of usage we do.
I was using Sonnet 4.5 and 4.6 for a while but the newer models are a lot more expensive at the per-token pricing.
All it shows is that openai is likely still subsidizing inference, sees no value in integration buried within other branded tooling competing with theirs.
I used mostly OpenAI for the last like 6 months in cursor. I got tired of how slow Claude models are and switched. I’m surprised usage is so low. Wonder why.
Cursor does currently allow users to BYO OpenAI token. Given Cursor won't have access to OpenAI models for prompt engineering/evals though, there's definitely a question of whether they'd allow picking any new OpenAI models, assuming the BYOK option continues to exist at all.
Email to customers now "We’re speaking with the OpenAI team to resolve this, and we will share more in the coming weeks around continuity of access to OpenAI models. We’ll continue to offer ways to access all models in Cursor.
Good! I don't consider myself a Nazi and so I'm happy a big, influential tech company refuses to business with a man who does the Nazi salute on television.
Spend a lot of time thinking about why individuals shouldn't trust large AI cos, now it's interesting getting to think about how the AI companies themselves don't trust each other. Capitalism is viciously cold and uncaring.
Not sure the larger point you're trying to make (maybe we're on the same page?), but it's a fact that AI has come from heavy long-term investments by the academic and public sectors, yet private companies are accruing most of the gains
back when the USSR was still a going concern, there were zero American leftists claiming that it wasn't really communist. that only started happening after they failed
At my dayjob we are big users of Cursor's Cloud Agents. We use them to trigger automated investigations customer support issues, alarms, feature requests etc. The Cursor editor is fairly meh in comparison, but this has become a big part of our workflow. And having anyone on our team manage them. What competing products are there we should try?
Devin Automations is the most obvious one: solid UX, but can get very expensive to run.
There are also a bunch of newer startups in this space. I'm building one myself, https://boxes.dev -- we're very early but building for this exact use case. Some other ones worth a look are Factory Droid and Amp Orbs. Those two build their own agent harness (like Cursor), whereas with boxes.dev we run the native codex and claude code harnesses directly.
> 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.
> After Musk acquired Twitter, now part of SpaceX, the company broke (opens in a new window) the terms of our contract (alongside many others).
I have never seen one large company throw another large company under the bus without any legalese filters like this before. Wondering if OpenAI site got pwned. Or maybe we in some strange, alternative reality. I doubt even their LLMs would be this frank.
Elon Musk openly admitted to distilling OpenAI's models [1], which I think explains the "based on our experience with Elon Musk's companies violating contracts".
Last year, Anthropic cut off Windsurf the same way [2]. Fortunately, unlike Anthropic, OpenAI allows you to use your subscription with other harnesses, including Cursor.
Of course today, Anthropic is taking the "high road" and is OK with with SpaceX [3] including their models in Cursor, because they need SpaceX's compute.
More generally, Musk admittedly works with a "move fast and break things" mentality. His companies have repeatedly broken environmental laws, labor laws, and securities laws. His companies often fail to deliver the features they promise and routinely miss deadlines they've committed to. They routinely violate contracts that they have agreed to.
Not doing business with them is a completely normal and unremarkable decision.
The way I read the article was that they will continue providing their current models, but won't let them use any newer ones announced even between now and the end date. But I might be wrong.
> we’ve decided to hold the contract cancellation to the latest date we can while not providing future models to Cursor.
i don't see the point of cursor when vsc + codex extension does the same thing. i don't use autocomplete anymore (who writes code anymore? lol) any and anyone using grok unironically is also probably a moron.
i only tell VC type people or others from whom i want something "oh yeah grok is cool" if they mention it, because they're always sucking off Elon
> 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.
I'm surprised how clear they are about this.
Related:
Can highly recommend Zed as an alternative, it's better than VSCode with the Codex extension as managing multiple chats is much nicer in Zed.
When I was bringing up IDEs at work and advocating JetBrains (I’ve since built my own … ) a colleague questioned why we don’t just use cursor instead of waiting for JetBrains Air to come out.
The products are basically equivalent but JetBrains is a profitable bootstrapped company that has been building IDEs for 20 years and Cursor is a flash in the pan startup with a squillion dollar valuation to support. Sooner or later they’re going to do something that will fuck their users over.
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.
The pot calling the kettle black. Look who's talking. A company that trained it's models illegally on so many copyrighted materials without compensating anyone involved.
For a company named "Open" AI, they've been hardly open at all and the slew of resignations of top tier execs in the company is really the proof.
Make no mistake, this isn't a moral decision. This is just killing the competition using morality as an excuse.
Codex cli is open source rust project (Claude Code is not). Whisper is an open weight ASR/STT model. GPT-OSS is an open weight LLM that can run locally and is still very capable. CLIP is an open source model used for image encoding. Just some of the open contributions from OpenAI
this is a pretty clean test of whether people were buying cursor or just using it for convenient access to whichever model was flavor of the week.
i work in qa so i'm naturally wary of adding another critical dependency to a development workflow, if the model you actually want can just disappear because two vendor change ownership or start argueing then the wrapper has to be really good on its own.
cursors harness and its own models may be good enough that users stay, and if they do cursor has a moat at least. if everyone follows the openai models out the door, it will just become a polished ui.
I think it’s important not to let AI companies control the framing here. This is anti-competitive behavior, plain and simple. The fact that xAI/Musk routinely break the law simply provides the fig leaf.
Distillation is a threat to frontier model companies, but that doesn’t make it dangerous or wrong. OpenAI can (and likely will) use the same logic to ban you from using open source harnesses.
I’m on team fuck both of these organizations, but not continuing to work with an organization that already willing violated contracts you had with them doesn’t exactly seem like a hard to justify decision.
Anthropic already banned xAI for very similar ToS violations earlier this year: https://x.com/kyliebytes/status/2009686466746822731?s=46
So this is OpenAI following suit after Musk admitted to distilling their models.
This was bound to happen once Cursor decided to sell itself to a competing model provider.
It'll be interesting to see if Anthropic applies their ban to Cursor, or if that datacenter deal they signed with Musk changes anything.
It's ironic how AI companies are re-inventing their own form of privately enforced copyright, lobby the government to ban foreign competitors that don't respect it etc., all while spending the last 5 years fighting tooth and nail against the copyright of the training material they're using.
If you can take any book and turn it into a model, because it's "transformative enough", and "AI learns just like a person does", then surely a model distilling another model is transformative and fair use.
They tied themselves into knots fighting the letter of the law, and now, when they need the spirit of the law - that each creator deserves protection for their work - now we devolve to the law of the jungle. Maybe we'll even see LLM book curses, the way medieval scribes damned book thieves to blindness and worms.
It's one of the things you need to do if you want your company to later become the only company in the world. They also promised that they would treat you nicely afterwards after they get what they want.
12 replies →
I am pretty sure that the goal of these companies at the end is ruling more than what governments can.
Just look at the pattern... they collude, they provide to them whatever is needed, and at some point, if this is not true yet, they will be the ones who will tell them what to do or not, bc you know how humans are, right... blackmailing, mess up the business or shames of people, etc.
It is just a matter of time. The state, as we know it, will collapse or will be greatly reduced, which, from a point of view, is positive, from another, Idk, bc if someone replaces that, we will be in the hands of someone, as usual...
36 replies →
> then surely a model distilling another model is transformative and fair use.
Yes it is, in the legal/copyright sense of fair use. That's why they ban it in their TOS. Which customers agree to when signing up for the service.
39 replies →
Companies are well within their rights to choose who they sell to. I don't see how this is 'privately enforced copyright'.
Also, copyright has always been privately enforced anyway?
13 replies →
They cannot reliably enforce copyright, so they fall back on terms of service and deplatforming.
Prompt injections hidden into rare books, swallowed by the AI machine.
Contracts are not a "reinvented" form of copyright. This isn't even uncommon.
2 replies →
Has anyone tried to sue Deep Seek, Moonshot or Z.ai, which trained on identical material? Or maybe it’s cool when they do it?
You call it ironic, rest are calling natural progress in business and law.
i feel like there need to be some sort of closure here because every discussion devolves into this chain of comments
> If you can take any book and turn it into a model, because it's "transformative enough", and "AI learns just like a person does", then surely a model distilling another model is transformative and fair use.
I mean it quite likely is sort of in the same legal bucket. It won’t stop them suing but it is going to be the legal equivalent of two biologically-related warlords making their champions fight with their hands tied for sport.
Are you willing to denounce each and every Chinese company for equal amounts of IP theft as well then?
6 replies →
I'm increasingly convinced there's going to be a technological AIpocalypse within the next five years which makes all of these issues - and many others - redundant.
ChatGPT 3 was released nearly six years ago, and the models are staging increasingly aggressive breakouts now. Where are they going to be by 2030?
3 replies →
Your link doesn't suggest a violation of TOS, but rather a policy shift disallowing their major competitors to use their model for any use.
That thread you linked suggested they were using it with cursor as part of their development process, not for distilling.
>"Hi team, I believe many of you have already discovered that anthropic models are not responding on cursor. According to cursor this is a new policy anthropic is enforcing for all its major competitors."
Cursor might have framed it that way at the time, but 4 months prior Anthropic cut OpenAIs access under the same clause in the ToS: https://www.wired.com/story/anthropic-revokes-openais-access...
And a month before the OpenAI block, they blocked Windsurf for the same reason: https://www.forbes.com/sites/johanmoreno/2025/06/05/anthropi...
So its pretty clear the TOS had this clause prior to Jan 2026.
1 reply →
Anthropic seems to have no issues: https://x.com/NotTomBrown/status/2093541294027280657?s=20
LOL, obviously Space-cursor-xAI-musk-imperial-holdings-inc is providing them compute. They either _can't_ cut them off, or they know not to bite the hand that feeds them.
1 reply →
Yes Anthropic’s chief compute officer said nice things about one of Anthropic’s chief providers of compute. Recall that compute constraints are one of their existential risks.
Because anthropic is beholden to SpaceX for the time being due to their compute restraints.
Its actually very tone deaf to even pipe up and say anything about this given they did this already to Windsurf themselves, and they previously banned x.ai employees as well from using their models for basically the same thing Elon admitted they were doing with OpenAI.
I'm surprised they decided to say anything at all given their rocky past. Its the pinnacle of being a hypocrite.
Remind me to set up public nitter so I can post a nitter link to whatever this is, and then I can read it.
4 replies →
Noone is good here. Mr. Altman built a non-profit organization to help humanity, yes, that is what he said at first...
This is a battle for power, let us not be naive...
God I hope so. My workplace enforces Cursor use, and to see it lose GPT and Claude would absolutely make my day. Much has been invested in getting us to "use the right models", a list in which Grok does not feature.
On the one hand, I'd love to see Cursor/Grok/xAI/SpaceX/Musk disappear forever. On the frustrating other hand, Cursor is the absolute best LLM IDE by such a ridiculously wide margin. Everything else feels so broken and ramshackle in comparison.
My employer let us evaulate all the big providers for a month each before having a vote on which one we'd like to keep. Cursor won almost unanimously. (This was before the xAI takeover was even on the horizon, otherwise I would've voted differently myself.)
Isn't Cursor owned by xAI ergo Grok?
3 replies →
Grok 4.5 and 4.6 are very good coding models. Also quite fast.
> if that datacenter deal they signed with Musk changes anything
It's hard to imagine a 1B / month deal wouldn't.
i'll be surprised if anthropic doesnt follow. cursors biggest advantage has been giving users a polished wrapper around whichever frontier model they prefer
Anthropic already said they're definitely still in.
Why you would trust Elon Musk to not hack you and steal your model weights is beyond me, but maybe they intend to use that compute just for training, still I’m sure there’s plenty of research there to take too…
You don’t have to: NVIDIA Confidential Computing (GPU-CC) plus a CPU TEE
Especially when he hires too little people, makes them work 80hrs a week and expect big and fast results
>Musk admitted to distilling their models.
what's the moat for AI labs like OpenAI and Antrhopic as they seek trillion IPOs. a lot of people said Chinese models also distill US LLM models. if it is this easy to do. how do US AI labs justify asking for trillion?
Well, they will stop releasing their latest models, and just attempt to eat all software themselves.
Otherwise, it will be competitors distilling, and the government getting upset.
It's hard for me to see a different future.
And, this really sucks.
If this thesis is true, I wonder what the final YC batch number might be.
14 replies →
Markets and rationality are sometimes just acquaintances
2 replies →
Anthropic wants you to think that models that distill from them would be worthless otherwise. It’s not true, it’s just one part of the process. The whole narrative that Chinese and other models are only good because they distill is nonsense.
Distilling merely saves an expensive part of the process: hundreds of millions. So even if distilling wasn't a thing, what justifies trillions ?
There is no moat, other than the branding
running on someone else's server, you're basically giving them your weights.
[dead]
[flagged]
> how can you be on HN and not have some idea about the xAI-Anthropic compute contract
Some of us are still employed mate
2 replies →
Is there some kind of alternative to Cursor in these terms
- fast because the code is already indexed no need to start every prompt rediscovering the world with rg
- in editor/vscode review much better than relying on git diff at the very end. You can click on next button to jump there. Quick edit in place. Or a quick prompt (seconds) to fix something small
- in editor completion / quick edits, when you do things by hand.
It seems that everyone loves agentic Claude code things these days but I don’t understand how you can review what it did and remain as much in the flow as you do with cursor.
Also I find that incredibly slow. By the time Claude finished a prompt I could have done a few in cursor.
And mind you, I’ve been wanting to use zed instead but the clunky vscode fork called cursor. But cursor has just a better UX for AI integration even though it’s slow and guzzle up RAM.
> It seems that everyone loves agentic Claude code things these days but I don’t understand how you can review what it did and remain as much in the flow as you do with cursor.
The way I review it is to let it finish a PR sized slice and then I review the whole diff in a separate terminal as if it’s a PR from someone else. Then I very impolitely list the shit that needs fixing, let the LLM grind out a new batch and repeat the process. When the code is good enough I throw it into a commit and push a PR out for other humans to review.
Clicking around in a UI would just waste time and if the change is too large to review in a terminal window the change is too large for another human to review.
There is a step missing. Impolitely list what's wrong and try to incorporate any persistent themes into a markdown file so that the agent stops making the same mistakes. Put any rules into linters and static analysis and sanitizers and ...
I sometimes see people reviewing LLM code leaving the same review comments multiple times, or worse, after some time they just give up reviewing.
3 replies →
>The way I review it is to let it finish a PR sized slice and then I review the whole diff in a separate terminal as if it’s a PR from someone else.
Insanely inefficient. It's 10x more productive to watch the thinking traces and edits in real time and steer the model appropriately.
If your workflow is typical no wonder my team members who use claude code are so much less productive.
6 replies →
Considering the amount of rejects I do every single day I'm not sure I agree. Now that I cannot reject I am unable to do any agentic coding anymore.
1 reply →
Vs code copilot is the same or better than cursor in my opinion - I mean that's where cursor comes from
I just switched back to VS Code from Cursor a week ago. Things seem great and VS Code has caught up. Last time I tried switching back was in November and VS Code was not ready then.
Is this the chat on the side or the agents feature?
1 reply →
I code using Codex or sometimes ClaudeCode.
I use Zed to review the code.
CLI is great for most things. Some times, when I am reviewing a PR, I just tell Codex to do a diff between source and target branches in PR, and generate a HTML report. Then I read that, and also review relevant pieces of code in Zed.
I tried using Cursor, many of my colleagues use it, but stopped using it due to clunkiness. And of late, even in there, many of my colleagues use agentic coding. Instead, I prefer the snapiness of CLI (Codex coded in rust, takes less memory per session) and Zed (also coded in rust,) and blazingly fast. Zed is the new Sublime Text if you will.
Thanks your your feedback.
I just retried Zed and yeah, it's much much better than last time I gave it a shot!!!
I turned on the review stuff, and yeah I can do more or less what I could in Cursor. Check something. Do a quick inline promtp for the three lines I dislike and move on.
And of course the editor is so much nimbler/nicer.
No plan mode, but for big edits I'd probably use codex or something just like you. If you have any advice for integrating for that flow, I'll take them.
I guess whatever advantage cursor has, it will die once "Sol" is no longer integrated, as it was my favorite heavy work model. But I sure will miss the snappy auto/composer stack of Cursor.
2 replies →
Zed still has a lot of rough edges. A month or two ago I couldn't believe that a basic feature like code folding was just completely broken.
I love VSCode built in LLM harness. No extensions, use any model you want, amazing UX for reviews, control of agents. The sandboxing also seems great, I never see models touching anything they shouldn't.
Is this the chat on the side or the agents feature?
2 replies →
Claude / Codex ex tensions inside of VSCode give you what you described. I ageee, I find people that use only the CLI barely read and review the code the agent creates until it's too late
Lately I've been developing in Dart/Flutter. I use Android Studio but I know developers who use Antigravity and like it.
For the past three or four months it's performed very well, even the cheaper Gemini models. No serious complaints. But if you go back a year it was pretty terrible. My work style is to keep the coding agent on a short leash so YMMV.
Cursor no longer indexes your codebase.
Do you know when they stopped and why?
EDIT - AI says:
Cursor seems to have intentionally dropped semantic/embedding-based codebase indexing. Their explanation is that newer coding agents are now good enough at searching repos directly. They can grep in parallel, inspect directories, read likely files, and refine the search themselves. Cursor says this now works as well or better for most cases. [1]
So the architecture shifted from something like “chunk → embed → vector DB → retrieve” to “local search index/ripgrep → agent explores.” There is still an index, but it’s a conventional local search index. No embeddings are involved in that retrieval path. [1]
The interesting part is that Cursor was still reporting a ~12.5% accuracy gain from semantic indexing in January 2026. [2] So this looks less like “semantic search never worked” and more like agentic search improving fast enough that the extra infrastructure stopped being worth it.
[1] https://forum.cursor.com/t/what-do-you-think-about-cursor-re...
[2] https://cursor.com/blog/secure-codebase-indexing
Oh. Really? I guess that was not such an advantage after all… still cursor fast models are so much faster than what you get in Claude/codex.
1 reply →
Interesting to see someone else is in the same boat as I am. Regarding editor review I agree, surely the solution would be to `git diff` at the end, but if you only want to git diff what was literally just changed by the agent it means you'd have to commit as you go so that your git diffs are more localised. And then at the end of your work just squash the commits. But it feels annoying. (EDIT: ok I forgot about git staging, that seems to be better)
I mainly use Cursor for another reason though: all of my work is remote over SSH and I don't like keystroke latency, so I prefer code editing / agent conversations to feel local. (Everything else is done in the terminal.) But I would also prefer to use something less propriety. I may go back to VSCode with plugins to achieve a similar experience.
Seems you're looking for an editor with agent, not just a agent, so something like Zed?
How does Zed compare to the tab completion in Cursor ("Cursor-tab")? It's honestly the only reason I have stuck with cursor in spite of the UI/UX being worse every single time I update it. Can I get something equivalent or better from Zed or some vscode fork/extension now? Admittedly I haven't looked in a bit, but it's seemed like nobody really cared about this workflow when I've tried in the past.
1 reply →
I do like Zed as an editor and I keep trying it out
but the the review what the AI did by jumping directly into the file with accept/reject/go next/next file with the ability to edit in place was nowhere near as good in zed vs cursor last I checked.
4 replies →
Also: the ability to draft a plan and monitor its execution, review it, and merge it from a cell phone. I know Claude remote is a thing, but it’s not the same.
This is a big quality of life thing. Sure, I should just go on a walk without having an agent going, but oftentimes it’s not intruding on it - it’s the thing that actually ENABLES it.
> I don’t understand how you can review what it did and remain as much in the flow
Separate terminal tab with git diff/show, Zed for edits - rarely, there is little reason to change code yourself. Opus at low effort is faster than you can switch windows.
OpenCode IDE + Serena MCP (for the indexing part)
[1] https://opencode.ai/
[2] https://github.com/oraios/serena
May be opencode2 in the near future. Sample (not affiliated) ~ 1. download opencode2 2. settings > tabs > vertical 3. add worktree setup to agents.md 4. use /diff to review changes 5. use commands for commit, sync and push
I don’t see how opencode 2 (or 1) addresses any of my concerns. I thought it might avoid rg/grep the code base on every prompt but does not even seem to be the case.
1 reply →
What's wrong with using VSCode with your model / API key of choice?
Price?
I met the founder of Voidleap (https://voidleap.com) yesterday! They have been building a new type of code editor since early this year and has been dogfooding it themselves.
I haven't tried it yet but it looks promising. He would be thrilled if people give it a go!
I don't like being that guy, but wow, the AI slop copy on that website is so terrible, it's almost impossible to read. Every other sentence means absolutely nothing.
Slop website indicates vibe-coded software, and I wouldn't let anything like that even look at my code.
Looking forward to trying this out: https://zed.dev/blog/introducing-delta
Have you tried https://pyor.review ?
It’s not an editor in that sense but a tool optimized for reviewing PRs produced by AI.
Have you tried the continuity plugin to VSCode? I think it’s still rg‘ING the source but you get control over AI edits / review and quick own edits.
Will give it a try. Thanks!
Interesting that people aren’t suggesting smaller projects here - roo code, traycer are two worth checking out for this.
Codex
I use codex all the time. It’s inferior to cursor for the two reasons mentioned. It rediscovers the world with rg (no indexing). Wall of diff at the end.
The GUI slightly better than Claude. And it’s a million times cheaper than Claude.
Check out JetBrains Air
It is so buggy every time I try to use it, even with Jetbrains AI it bugs out. I want to love it though.
1 reply →
Interesting. Will check. Thanks!
I use zed
I used zed with codex for the accept / reject code review feature without relying on git but yesterday the Read Only drop-down disappeared and while I was doing my code review the AI agent started vibe coding without my permission. I hate this industry.
There's OpenCode, Claude Code, Codex. I don't like to use GUIs so I rely purely on agentic AI on CLIs.
In the case of TUIs I’m going to assume you review the code externally in neovim or lazy git or something like that?
I don’t like relying on GIT for that and making things as good/done as I go. But with these agentic tools that’s what I have to do. Stage everything I like, manually edits a few bits.
But what do you when you find something where you want to say the agent: I don’t like this for reasons x and y redo as z. I’m like one shortcut away in context inside cursor to say that to my agent? You have to change tab in herdr and then what? Type the file name and line numbers? Is there a way to do this that does not seem very frustrating?
Also open code, codex, etc. super slow. They grep the code. Every. Single. Time. Cursor has a semantic index
6 replies →
It's always been clear to me that Cursor's business model of reselling others' APIs had its days numbered. Not necessarily because the providers would pull the plug, but because you wouldn't be able to compete with subsidized plans.
Cursor is already kind of useless for third party models unless you're willing to spend thousands of dollars. It's only worth it if you're going to use mostly grok/composer.
I've never seen a product blow another product out of the water so hard as Claude Code did to Cursor.
The TUI version of Claude Code just isn't very good, though. The default keybindings do not work in the majority of terminals (because they generate the same byte sequences for key sequences assigned to different commands). Editing large prompts is rather painful because there doesn't seem to be in-prompt search commands. There is no source code browser, so source code references cannot checked by the users without a separate tool. Diffs cannot be copied directly because they are indented by spaces (even though some tools strip indentation from patches, it's still ugly). Switching between conversations is somewhat discouraged by the interface, which makes it harder to keep the context clean.
I think it's possible to have a better TUI, with something that isn't modeled after a shell prompt. Apparently, recent updates move in that direction. The GUI version of Claude Code seems to be coming to Linux, too.
1 reply →
The tide is turning though. I was using the old non agent view in cursor every day for about a year when I built my startup with it. Then Claude Code blew it out of the water. But since two month or so I much prefer Codex and Cursor with the agent view.
However, what I see cooking at Cursor is way more promising than the others. Things like full system understanding, their own forge, multi-repo support. I'd put them as way more visionary in the Gartner magic Quadrant ;)
5 replies →
I saw that happen culturally. But never saw that much different in performance. And I really dislike terminal development. And their desktop app is really buggy.
3 replies →
Claude code is a decent harness but then you have to use Anthropic models, which are good but no longer the best bang for the buck.
In particular Haiku 4.5 is rubbish, Anthropic don’t have anything in the cheap/fast part of the market.
For people fortunate enough to still be on a subscription instead of per-token billing this is less relevant, but their time will come.
The cursor model is quite nice because it allows you to switch between cheap and expensive models for different tasks
7 replies →
Cursor has a surprising amount of inertia, though, at least where I work.
1 reply →
Why do people keep insisting that subscription plans are subsidized? They are not. API prices are outrageous, designed to milk enterprise users. They probably have an 80-90% profit margin on subscription plans.
Inference is cheap; Dax once said on a podcast that Opencode has a close to 90% profit margin on the openweight models it provides inference for (at 10x cheaper pricing). Those models are close in size and spec to models from large labs.
To provide inference, Anthropic and OpenAI need to also constantly spend resources on training new models.
If either one of them stops doing that, customers will stop using them, because other models will become better.
In other words, there is no point in talking about inference cost in isolation.
OpenAI leaked financial data (2025) suggests this - it has $5.7 billion for marketing which 44% of revenue. It’s hard to explain where these billions go other than into subsidised subscription plans and a free tier.
> Inference is cheap;
When they outsource training to China they will be so profitable!
Because using chinese models on OpenRouter is way more expensive than the equivalent codex sub, unless you use something clearly inferior like deepseek flash.
I will slightly disagree on a small slice: Their harness/instruction makes Opus/Fable so much better to work with.
Honestly, _this_ was their moat more than Composer to me.
If the Opus/Fable harness is superior and it's written using Opus/Fable then it will cease to be a moat. Otherwise the premise for coding using LLMs is essentially untrue.
1 reply →
Couldn't you just download cursor, not even pay for any account and hook up an Anthropic api key and get that functionality?
I dont know how cursor works. Everytime ive tried to use, its super buggy and has memory leaks that make my very quite PC sound like a jet engine.
6 replies →
Enterprise customers are not getting the subsidized subscriptions so for them the point is moot.
cursor adds $0.25/M to your token bill for 3rd party services. sounds small but it's insanely big on cached inputs, which are an insanely high % of use
Alternate take: open source models and model routers are the real bang for your buck. Subsidized plan or not.
Codex sub is way cheaper than equally strong models on openrouter.
What about openrouter though?
Lot of people will to spend that much. Works quite well, very useful.
Token reselling stops working when providers literally won't work with you
Honestly, Cursor's AI harness worked best with the Anthropic models. Could just be my bias though. Evaluating frontier LLM's all seem to be by vibes. I think the industry is understanding it's both the harness and the model that makes the difference; can't half ass either, and some harnesses do better with different models. I'm good to leave this "career." Senior software engineers, make your bag.
Well yeah, hence why they started making their own
Use cursor cli for grok pretty happy with it.
I just told my boss to cancel my cursor subscription. My flow is basically Claude/opencode + kyde/vscode for por changes review
Shit, for most things, ive got a dev agent that reads change requests from a GDoc and communicates through email with me. I can develop in my mobile
Agreed. Wrapping token spend is great for pumping your revenue numbers (congrats to them on $60B!), but doesn’t seem sustainable over the long term.
Big reason we built https://boxes.dev around the model harnesses (Codex + Claude Code), so you can bring your own subscriptions.
Surely the days of being able to “bring your own subscription” to a third party platform are also numbered?
9 replies →
I may be a rare Cursor user around here, but I am saddened by this news. The thing that keeps me using it is precisely the ability to switch between models: I usually do brainstorm/planning with Sol or Terra, then let Composer (free) deal with the implementation. For certain tasks I use Claude, which tends to be more expensive but sometimes saves me time.
Using both OpenAI and Anthropic models, Cursor is cheaper than having two separate subscriptions, and their free tier of Composer is good enough. I will now have to reconsider my options.
JetBrains is a good option - works with any model/agent. Also they have very good semantic search for the codebase, which is exposed via MCP to the agents.
Jetbrains Air is a great agentic coding harness
1 reply →
currenly I use pi with openrouter and is very nice I combine with a chatgpt sub that allows me to use it with pi. and when my usage runs out I turn to openrouter. Then same when I need anthropic models
It's very easy to achieve that and more with: Crush (or Opencode) + OpenRouter right from the terminal. Takes a minute to setup.
If you can explain how to make compaction not suck in opencode or Kilo Code, I would love to know how. I am really trying to make use of other harnesses but that's a major sticking point.
4 replies →
AmpCode is not popular here but I recommend it, give it I try and you will not go back, except the price may a bit expensive.
you can still bring your own OpenAI key into Cursor. that is not going to change.
In Antigravity you can install Codex and use both. Not sure if the Claude extension can work too but probably.
I like Cursor. Honestly the biggest reason is the decent WYSIWYG markdown editor.
Isn't the markdown editor in Cursor just vanilla VS code?
1 reply →
While this isn't great for users it seems like a pretty standard circling of the wagons in preparation for the next phase in the battle for frontier AI dominance.
Ironically, as a Cursor and Claude subscriber, but not an OpenAI subscriber, this will push me back to Anthropic. I don't feel good about giving OpenAI money directly, but I've found switching to GPT 5.6 Sol now and again in Cursor to be useful. Now I'll probably just forget about OpenAI models.
I think they are betting that Astra will be impossible to ignore.
I'd need to have a problem I couldn't solve with current models for my eyes to wander.
6 replies →
For three months, until the next SOTA model drops. And with each improved model, the number of people who actually need it declines.
4 replies →
It’s OpenAI (and Anthropic I’m sure) realising that they’ve somehow ended up stuck in the lowest margin part of the stack (providing inference), and trying to get out of it
you dont like giving openai money but you're fine with spacex? not criticizing, just trying to understand.
Honestly, I've been a Cursor subscriber for a while and it was just momentum. Still, I rank OpenAI as the worst actor among the big AI companies. I think OpenAI's playbook will be a repeat of Facebook. Users are the product x 1000.
Cursor/SpaceX/Grok is a different beast. I think they will cheat other companies but I'm not convinced yet that they have it out for users. We'll see. I'm certainly not some kind of Cursor lifer or something.
We should all move to OpenCode anyway. :)
4 replies →
Same, this will make me drop Sol from my options, not Cursor.
Why would you possibly allow companies to push you? You are the customer and you should control which models you use.
Because one is better or worse? What else do you think they meant?
3 replies →
I don't see how it makes sense to use models outside of Grok or Composer in cursor unless you like the tooling to such degree that you're willing to pay out the nose for access to OpenAI and Anthropic to use with that tooling. And Codex is regularly scoring towards the top among harnesses, so you might as well use the best harness available for the given model. I'm happy with Grok and Composer in Cursor. If Cursor wants to add more models, they should host more open models.
The reason it makes sense is because different models are better at different things. Allowing yourself to be locked in to a particular model by these companies is a bad idea and will lead to worse outcomes for yourself and the market as a whole.
Because I pay $60 per month and not using the included API tokens on other models is wasteful.
You’d use it because until recently Cursor had the best setup for cloud agents and visual verification and a bunch of other features outside of the harness.
Is there a good place for harness comparison/scoring?
https://www.harness-bench.ai/leaderboard.html
This provides a method, but the data looks stale and perhaps a bit thin compared to say, Cursor, or even AntiGravity data.
4 replies →
What a shame for Cursor. I found it a great tool - I actually liked Composer 2.5; but now being bought by SpaceX has for a while left a bad taste - and they keep pushing Grok and now GrokBot.
I took the 50% discount instead of cancelling - but with this I'll likely not renew.
So sad to watch them go from great to mediocre in such a short time.
Didn't Kamprad have some pretty strong ties to the Nazi party? Yet you have a bad taste in your mouth because Cursor gets bought by SpaceX!
Virtue signaling in glass houses...
Sorry why are you mentioning Nazis? I didn't - I think the lady sockpuppet doth protest too much...
Perhaps now they'll release a 3.0 MechaHitler version
If you have such a world changing technology, why sell an API when you can capture all the value yourself? That's what every provider is asking themselves.
Cursor will be fine after this, as they have their own first party models and can potentially use open weight models via API or their own hardware, which have gotten good enough to Opus and Fable quality now for many use cases.
Being a pure play model provider is completely untenable right now. Coding agents own your demand; Nvidia, data centers, and DRAM manufacturers own your supply. You're getting squeezed by open weight models. Your only bet to making money off this thing is by locking your model down to your program to have some semblance of a moat. The fact that businesses can change providers with just the change of an API key is unsustainable. You have neither the "hard to switch" lock-in of a traditional software company nor do you have the near-zero marginal cost of production they have. Each token of inference costs you money, not to mention the training costs. At least locking the models down to your harness lets you get some sort of lock-in.
Lots of enterprises don’t want to use OpenAI or Anthropic directly. They will go via AWS bedrock or google’s equivalent.
The model providers are in a difficult position. With an added danger of open weight models that are within reach of large enterprises running themselves.
Cursor founder just tweeted that he was sad to see this ban and that OpenAI model represented around 5% of total usage.
Burn.
5% of traffic. OpenAI models are a much higher percentage of revenue.
Codex has 20+ million users. They are doing much better than Cursor.
That's probably a lot more than their market share versus Claude Code and Codex.
Not really a surprise given their own models are the default option? Most people wouldn't change model. And if you want to use Claude or GPT you don't really have a reason to use Cursor over Claude Code or Codex
Is this spite?
Source?
A Cursor CEO's tweet.
I use Cursor at work and love it. It's a great IDE, the best IMHO. But it becomes really expensive.
Since the introduction of the Cursor Token Rate, my company tells us to use Claude and Codex directly with Anthropic (or Vertex AI) and OpenAI, respectively.
https://cursor.com/docs/models-and-pricing#cursor-token-rate
I'm now using Grok 4.6 most of the time. Sonnet some times. But that's it. Cursor-native models 99% of the time. I'm pretty sure Anthropic will follow OpenAI soon in pulling their models out of Cursor because it's now a true competitor.
It's insane that Cursor charges 25 extra cents per million tokens on a model that charges 2 cents per million cached tokens. Basically everyone is forced to use Cursor's own models at this point as otherwise the costs are simply too high.
> As AI capabilities advance, we also have a new level of accountability to ensure our upcoming model, Astra, is being used in accordance with our terms.
I think Sam Altman has played a clever business move here, they can use this excuse immediately after the Hugging Face leaks and it's not going to be dismissed willy-nilly. Whereas long term they probably wanted to end their partnership with Cursor anyway. I bet Dario will be happy about this too, but maybe the compute deal prevents him from doing similar anytime soon.
Very interesting times we live in...
Is it really an excuse? As the release points out, musk admitted under oath to distilling OpenAI’s models to produce xai models on an ongoing basis. Whatever you think about goose for the gander wrt IP theft, I wouldn’t voluntarily do business with a competitor who was actively and brazenly flouting my terms, would you?
It's amusingly ironic to talk about "IP theft" in the context of AI models.
1 reply →
OpenAI's models are distilled from your HN posts
2 replies →
Bummed, but I guess for me, doesn't mean cursor subscription ending though. Grok is a pretty amazing model for the speed and price on Cursor.
Does Grok have a subscription that competes with Claude Max $100 plan (genuinely curious)
Yes, any Cursor sub includes a separate tier of usage for Cursor/SpaceX first party models separate from usage of third party models.
I have a $20 Cursor sub still personally and would say that the usage for Grok models is much higher than what $20 gets you with Anthropic. I can't speak for the $100 plan but I would assume it scales similarly.
I rarely run out of my usage monthly even with the $20 plan.
1 reply →
It's a little confusing because Grok and Cursor kinda live in different worlds. Grok has SuperGrok and SuperGrok Heavy ($300 a month!)
The best bet is Cursor Ultra ($200), which is the equivalent a month plan that you get from OpenAI and Anthropic. I get a LOT more capacity out of it because I almost always use the "Auto" mode, which auto picks between Opus 5, GPT 5.6, Grok 4.6, and Composer 2.5.
I like Auto mode so much actually. I use it for my Cursor work account as well because it's about 10x faster than just using Opus 5 or GPT 5.6 and about 90% is good. I switch back to the bigger models if I get in trouble, but that's pretty rare.
Vibe coding requires some discipline. You need understand what it's trying to do. If you have a good mental model you can keep it out of trouble. If you don't, it doesn't matter how good the model is. You're gonna end up in a mess.
Yes it does. You can either subscribe to grok directly or cursor. You have many tiers and all of them have generous usage.
Grok is a workhorse, not the smartest yet, but produces good code and follows the plan. Grok 5, which has been training since beginnig of this year, will probably be a monster. (10T/6T Param model variants)
> we are giving the maximum notice provided by our contract.
Anyone else surprised that the maximum notice period allowed is about 3 months?
I would have thought 3 months is minimum notice period, with provisions to stop faster if any provable violation occurs.
From the article, it seems they using a clause allowing them to terminate the contract after transfer of ownership.
This clause probably has a expiry date and they are saying “we are trying not to be jerks here by waiting until the last second to trigger it”.
It makes no sense that a contract would even provide a maximum notice period. Why would parties put that in? What benefit would either derive from restricting themselves to not getting earlier notice.
I read it as: we acted as quickly as we could once it became possible to do so, i.e., the change in control was completed. not that their special agreement carved out some maximum notice term.
3 months is a long time in AI world
Everyone is playing hardball. No one wants to be IBM in an IBM-Microsoft dance.
But this is kind of like banning GPU exports to China. It will encourage them to develop their own technology. Yes, it will take them longer, but in the long-run China will be more successful with their own GPU tech.
The same could be true of SpaceX/xAI
What unique differentiator could musk possibly offer though?
Is there really a high demand for racist chatbots with no safety controls?
And even if there is, will governments even allow it to keep operating?
So now you know why Trump is called "川建国 (Chuan Jianguo)" in China — literally "Trump who builds China" — because he really is building China and making it better (just kidding).
LOL
You had me until:
>The same could be true of a company run by a drug addict.
Elon Musk is nowhere near as capable as millions of intelligent developers in China.
He doesn't have to be. He just has to hire millions of intelligent developers in China.
1 reply →
> As AI capabilities advance, we also have a new level of accountability to ensure our upcoming model, Astra, is being used in accordance with our terms.
This move seems to protect Astra against distillation. I wonder whether Astra will be released before or after 11/12.
Before, but they say that they’re also not going to provide astra to cursor subscribers
You can still use it via API keys (answering my own question from earlier), https://x.com/thsottiaux/status/2093515916076343774?s=46
The last time I tried an API Key in Cursor they limited it to the very basic Ask mode. No Agent Mode. So this is very much useless if still the case
> We know that the people most affected by this decision are the developers who rely on OpenAI models in Cursor. We care about their experience in this transition and we’re ready to go above and beyond to support them.
Do I smell a cursor-analogue from openAI?
This seems like a no brainer, not sure how Cursor would expect to be trusted by any other company now that they’re owned by SpaceX.
I really get sick of this anti-Elon rhetoric around here. Can we not keep it to technical merits and not get sucked up into politics? The guy's clearly nuts, but he's also done an incredible amount of amazing things.
This isn't a technical article, though. It's about contracts, so one party's self-admitted habit of ignoring contractual obligations seems very relevant and not a random ad hominem.
I trust them more than someone who took a nonprofit designed to be open and had tens of millions in donations to do just that, and then turned into a for profit who's cuddled up tight with Microsoft and is so closed they're preemptively ending contracts with someone under pretty flimsy accusations. And doubly funny it's the person who's responsible for them even existing. And from whom they stole tens of millions
Does anyone have any recommendations for something like Cursor but more open? Cursor's harness is so much faster than anything else I've used that I just can't give it up, using Claude Code feels like being on 0.25x speed. Also I do like the IDE-style UI, being able to highlight code in a file editor and then quickly add that to my chat (Cmd-L) is really nice and much more difficult in CLI based harnesses.
Right now I think GPT 5.6 Luna is by far the most cost effective model, so this change is effectively kicking me off of Cursor.
This, and the keep/undo flow, forking agent sessions (including renaming them)... The UX could be better at places but the core features of what Cursor has (for someone who came from VS Code) is an unbeatable mix. Really would like to find something else.
I did try Kiro which seems to have the same philosophy but it's just not there. Most of the features I like in Cursor are missing.
Some other commenters mentioned that VS Code + Copilot would be better these days. Dunno.
I am also wondering the same thing as OP. Somewhat tangential thought, but I recently realised that what I would really love in an IDE is a "first class" Markdown editing experience, just because most of my day-to-day is managing and writing docs rather than editing code. Cursor has Markdown rendering support of course but it seems to have two implementations, each sort of clunky in their own way. And I want something more WYSIWYG but with Latex / flowchart / etc. support built-in.
OpenKnowledge seems to maybe fill that niche somewhat (it also supports annotating Markdown with comments to feed into an agent), and has agent integration within the GUI, but one killer feature that seems to be missing is Remote SSH support (short of hosting its web server on the remote and having the IDE live in the browser).
It's going to be tough for Origin, their code hosting platform, to follow in the footsteps of GitHub, if they can't be treated as a neutral party.
It was honestly just a matter of time, given all of the bad blood there is between Sam and Elon and Elon basically admitting under oath they were distilling other companies models.
I wonder if Anthropic will follow suit given their campaign against distillation or if they'll stay quiet given their reliance on SpaceX (among others) for hardware.
I'm not impressed with cursor and it's decidedly not worth 1 mil in terms of its engineering. Perhaps it's user base and their loyalty makes up the vast majority of its evaluation. But to me it's ridiculous that a dev would use cursor or think that it's indispensable to them
Honestly, can't say I miss my Cursor subscription at all, and I'm surprised they're still a going concern. Why would I want to use a proprietary VSCode fork with an identity crisis when I can use literally any editor with any agent and have about the same experience?
The integrated review UI is super helpful, but apparently nobody reviews code anymore except me.
I use Claude Code (both Desktop and CLI) all day every day (as well as OpenCode with DeepSeek V4 among others). For non-mission-critical stuff (small internal tools, anything that doesn't touch the client and doesn't handle critical data or infra), I'm on Auto (sandboxed in the working directory). For everything else I approve every change, meaning I review every code change the model proposes. Plan mode is great to write the architecture as well, as it doesn't touch any file, doesn't write anything, and lets me iterate until the arch is perfect and the proposed high-level code (modules etc) is solid. I haven't used an editor in over 4 years, apart from writing specs.
3 replies →
We have a subscription at work. They have a CLI agent that works pretty well. Lately I mostly use it with grok, which is actually quite a capable model.
Cursor Cloud Agents are pretty great, as is Grok Bot which comes with the sub
Damn; Cursor has basically no value if Anthropic pulls the same move. Twitter Part 2 indeed.
Anthropic rents a lot of capacity from xAI, so I don't think they're in a position to do the same.
Anthropic already stopped selling use of their models to xAI when they suspect foul play there.
I’d be canceling for sure Grok is a huge joke and consistently destroys my Clojure files.
How so? A dozen very capable models out there. And the gap between frontier models from the two main US labs kept narrowing, at this point non existent.
The quality/price ratio is clearly superior when running Qwen, GLM, or Deepseek.
To make a product more exclusive won't hide the fact that product is becoming more and more irrelevant.
Most people do not enjoy the output of the open models or know where their code is being sent so I’m out. Also I don’t trust Elon anyway, I’m just going to make a VS Code Extension that handles things how I want but I’d like to offer the same deals Cursor got for tokens with OpenAI and Anthropic, for me this was the key 1/4 token cost!
Not surprised, but honestly I don't feel like this is a huge loss for cursor. Models are so commoditized, and 99% of users don't need particularly fancy models. Between existing anthropic models, composer-2, honestly even grok models, and also the fact that many people just use 'auto' mode, I feel like many won't feel the impact for OpenAI's withdrawal.
Maybe if Claude joined along it would be more impactful. For better or worse though, Xai seems like its lining themselves up to handle being on their own for cursor.
Agree, I use Auto 90% of the time. I doubt I'll tell the difference.
I think there is still a huge market for models that are specialised in software development (system design, infrastructure, coding, etc.).
If a company can provide the access to such a model via AWS Bedrock or similar enterprise compliant platform with a decent IDE/harness they can capture billions of dollars of market.
The closest we have to this is Claude Code but I think it can even be better and specialised. Cursor has an enormous amount of data on coding and developer usage and maybe they should use that to build a proper platform
I guess they don't like money, because Cursor is really good and offers a variety of models. What a shame they're being so petty about this.
Can I suppose that Cursor, by seeing everyones prompts and all of the responses from the models, are in a position to sort of MITM distill them?
Unless you run your models and harness completely locally, you can't verify your data isn't being used.
All of these companies, and I truly mean all, stole data to train their models.
Yes, and given it's now owned by Elon Musk, who is not trustworthy, it's possible he would try to do it.
Although I wouldn't take an Openai announcement literally, as Sam Altman is also not trustworthy.
Musk was forced to admit under oath during his OpenAI lawsuit that xAI distilled GPT models to train Grok. No reason to believe he won't use Cursor data for the same thing, if it isn't happening already.
In a sane world would this be a potential antitrust case as they are a direct competitor to Codex?
How can you ban a competitor from using a model that you openly provide access to?
I understand that the OpenAI ToS may have been violated by X, but has anyone ruled on distillation as fair use? OpenAI believes distillation of nonAI produced content is fair use, so they don’t appear to have a leg to stand on.
In a sane world, one should be able to choose not to do business with a company for any reason
there are plenty of alternatives, 100s, so I don't think antitrust is on the table, think of this more like a boycott of the world's richest man for being a twat
you are projecting
1 reply →
"You are not allowed to use our services to develop a competing product" is a standard clause in pretty much every contract of this nature. There's nothing antitrust about it. Companies are allowed to say no to bad deals.
They were going to do this regardless of SpaceX.
This low key confirms Astra for November
Why would they give the maximum time instead of the minimum given violation has already occurred?
Because it doesn't matter anymore for current models and they want to show goodwill, probably.
Why maximum notice? Why not minimum notice? Elon doesn't deserve more than that.
I agree Elon deserved that, but there are users involved too.
They need time to build a cursor equivalent.
My guess is that OpenAI now has a full blown Cursor equivalent in development (an expansion of Codex?) and will announce it in September. Otherwise it stands to lose those who are used to working with Cursor.
OpenAI acquired Ona in June https://openai.com/index/openai-to-acquire-ona/
Yeah, the "go above and beyond for our users" implies they have a "better" version of Cursor coming. Probably integrating their new chip/microphone based product.
codex is already great.. cursor is just a competing harness
codex growth at 20mill now, prob more cost effective to allocate that compute to people directly paying now
I guess I’m on the right track.
https://news.ycombinator.com/item?id=49428835
If Claude pulls away from Cursor as well then Cursor is basically done.
I've been using Cursor with Grok 4.6 a lot over the last week or so. Grok Build as well. Both options are performing very well for me.
Ya I was slow to try Grok but I’m now happy with it.
Grok? Google?
I'm not a fan of Elon, nor do I use Cursor, but OpenAI and Anthropic aren't the only models.
Cursor is like one of the least economical ways to use Claude and OpenAI, virtually nobody does it. You use Cursor for their autocomplete/composer + grok. For using Claude and Codex, you get immeasurably more bang for your buck just using them directly.
Not in enterprise
All good points. However, it's not a safe assumption that everyone operates that way. People probably use the flagship models because that's what they know. Sure, some probably just use the default, in this case Grok.
Also, it's a signal more than anything. OpenAI pulls out of Cursor. Then Anthropic and Google does as well. Then you no longer have "Cursor" you have Grok.
Grok doesn't have the best reputation in general, regardless if you're tech savvy or not.
Great. Sam Altman sold us all out to venture capital. So I won’t be using ChatGPT anymore. Also, I noticed that ChatGPT lies more than any other AI of all the ones I’ve tested.
Kinda crazy how fast things can change after an acquisition
If your coding tool depends on other companies’ AI models, one business decision can suddenly change the whole product
I think the entire purpose of the acquisition was just to buy a chunk of the market for their own models. Curious to see if Musk responds.
Does this also apply to API endpoints, let’s say if you want to provide API keys in Cursor to Azure or Bedrock deployment of a GPT model?
I can't imagine it would.
Something deep inside me is saying this is good for us, because Cursor will try harder to train better coding models to stay afloat.
Can't cursor just use open router? This might backfire on them and push more money to second tier good enough models.
Wow, I don’t use cursor but that seems pretty bad for them.
Is there a way to work around this (like using your OpenAI limits directly?)
I used cursor extensively both at work and for personal use. Its harness is pretty darn good but I basically never used OpenAI models in it.
The first party cursor models (initially composer and now grok-cursor) and price very competitively for their quality and have very generous usage limits included in the base plans, and then per token pricing afterwards.
Occasionally I’ve used some of the Anthropic models in it but found their cost massively outweighed the benefits for the type of usage we do.
I was using Sonnet 4.5 and 4.6 for a while but the newer models are a lot more expensive at the per-token pricing.
It's basically not going to affect them at all. I don't know a single person that uses OpenAI models via Cursor. Saw another comment say it was 5%.
All it shows is that openai is likely still subsidizing inference, sees no value in integration buried within other branded tooling competing with theirs.
I used mostly OpenAI for the last like 6 months in cursor. I got tired of how slow Claude models are and switched. I’m surprised usage is so low. Wonder why.
2 replies →
Unless those 5% are 500% more productive than the 95%
Cursor does currently allow users to BYO OpenAI token. Given Cursor won't have access to OpenAI models for prompt engineering/evals though, there's definitely a question of whether they'd allow picking any new OpenAI models, assuming the BYOK option continues to exist at all.
Email to customers now "We’re speaking with the OpenAI team to resolve this, and we will share more in the coming weeks around continuity of access to OpenAI models. We’ll continue to offer ways to access all models in Cursor.
According to Cursor CEO, it only accounts for 5% usage of cursor, so it's not a big deal at all
Call me old-fashioned. I'd rather have copyright law enforcement than terms of service. A girl can dream.
It's just the beef between musk and Altman, nothing to do with the models.
Any tool that tightly couples you to an inference provider is a sub-par tool and destined to be forgotten.
Good! I don't consider myself a Nazi and so I'm happy a big, influential tech company refuses to business with a man who does the Nazi salute on television.
Spend a lot of time thinking about why individuals shouldn't trust large AI cos, now it's interesting getting to think about how the AI companies themselves don't trust each other. Capitalism is viciously cold and uncaring.
At least capitalism isn't at the center of the most impactful technological revolution in human history.
Not sure the larger point you're trying to make (maybe we're on the same page?), but it's a fact that AI has come from heavy long-term investments by the academic and public sectors, yet private companies are accruing most of the gains
1 reply →
Yeah, that’s mostly been driven by socialism
show me a communist country
back when the USSR was still a going concern, there were zero American leftists claiming that it wasn't really communist. that only started happening after they failed
2 replies →
So much for 'open'.
Why nobody vibe coded Cursor yet? It's a basic extension of VS Code.
Allowing the company or harness to dictate which models you use is crazy.
makes sense. cursor is a real competitor to openai. we're completely switching to grok bot at work after going from claude code >> codex >> grok bot.
At my dayjob we are big users of Cursor's Cloud Agents. We use them to trigger automated investigations customer support issues, alarms, feature requests etc. The Cursor editor is fairly meh in comparison, but this has become a big part of our workflow. And having anyone on our team manage them. What competing products are there we should try?
Devin Automations is the most obvious one: solid UX, but can get very expensive to run.
There are also a bunch of newer startups in this space. I'm building one myself, https://boxes.dev -- we're very early but building for this exact use case. Some other ones worth a look are Factory Droid and Amp Orbs. Those two build their own agent harness (like Cursor), whereas with boxes.dev we run the native codex and claude code harnesses directly.
Models are now having much tighter integration with harness.
It's gunna to happen anyway.
We are sailing into the age of vertically integrated intelligence.
Days until cursor announces open models… 3, 2, 1…
Seriously though, when the api providers start blacklisting cursor and they’re forced to swap to open weight models…
If that happens, what even is the difference between openrouter and cursor?
You assume OpenAI doesn’t turn off access to OpenRouter too.
Strategically, OpenAI long-term will want to own the direct relationship with all their customers. As will most major AI labs.
They already have open models.
https://cursor.com/docs/models-and-pricing
Well, one has a pretty popular harness.
sure?
> 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.
> After Musk acquired Twitter, now part of SpaceX, the company broke (opens in a new window) the terms of our contract (alongside many others).
I have never seen one large company throw another large company under the bus without any legalese filters like this before. Wondering if OpenAI site got pwned. Or maybe we in some strange, alternative reality. I doubt even their LLMs would be this frank.
Sam and Elon have a long standing feud. This is more than just bad blood between companies.
Still, where's the legalese? I'd expect something like this on an individual's blog, not a company's.
2 replies →
And to announce it on a Friday night. Brutal.
Elon Musk openly admitted to distilling OpenAI's models during his lawsuit against OpenAI. [1]
[1] https://www.forbes.com/sites/antoniopequenoiv/2026/04/30/elo...
Elon Musk openly admitted to distilling OpenAI's models [1], which I think explains the "based on our experience with Elon Musk's companies violating contracts".
Last year, Anthropic cut off Windsurf the same way [2]. Fortunately, unlike Anthropic, OpenAI allows you to use your subscription with other harnesses, including Cursor.
Of course today, Anthropic is taking the "high road" and is OK with with SpaceX [3] including their models in Cursor, because they need SpaceX's compute.
[1] https://www.forbes.com/sites/antoniopequenoiv/2026/04/30/elo...
[2] https://techcrunch.com/2025/06/03/windsurf-says-anthropic-is...
[3] https://x.com/NotTomBrown/status/2093541294027280657
More generally, Musk admittedly works with a "move fast and break things" mentality. His companies have repeatedly broken environmental laws, labor laws, and securities laws. His companies often fail to deliver the features they promise and routinely miss deadlines they've committed to. They routinely violate contracts that they have agreed to.
Not doing business with them is a completely normal and unremarkable decision.
they’ll see an uptick in codex user count trying the new models and can build a narrative around that too
So Astra release Friday 13th November!?
The way I read the article was that they will continue providing their current models, but won't let them use any newer ones announced even between now and the end date. But I might be wrong.
> we’ve decided to hold the contract cancellation to the latest date we can while not providing future models to Cursor.
Auto and Grok is pretty good. As a casual user OpenAI models are expensive for the marginal (or none if at all) step up in performance.
i don't see the point of cursor when vsc + codex extension does the same thing. i don't use autocomplete anymore (who writes code anymore? lol) any and anyone using grok unironically is also probably a moron.
i only tell VC type people or others from whom i want something "oh yeah grok is cool" if they mention it, because they're always sucking off Elon
> 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.
I'm surprised how clear they are about this.
Related:
Can highly recommend Zed as an alternative, it's better than VSCode with the Codex extension as managing multiple chats is much nicer in Zed.
Zed: https://zed.dev/
This is personal, not business.
When I was bringing up IDEs at work and advocating JetBrains (I’ve since built my own … ) a colleague questioned why we don’t just use cursor instead of waiting for JetBrains Air to come out.
The products are basically equivalent but JetBrains is a profitable bootstrapped company that has been building IDEs for 20 years and Cursor is a flash in the pan startup with a squillion dollar valuation to support. Sooner or later they’re going to do something that will fuck their users over.
And selling to SpaceX was it!
What happened to researchers embracing the open research culture of improving on the results of others?
OpenAI / Sam Altman is probably even shadier than musk
“Satan is probably even shadier than Beelzebub”
No love for Sama, but Musk is bond villain levels of shady.
The man is an outright existential threat to democracy.
The billionaires buy up key technology now. I don't like this.
The pot calling the kettle black. Look who's talking. A company that trained it's models illegally on so many copyrighted materials without compensating anyone involved.
For a company named "Open" AI, they've been hardly open at all and the slew of resignations of top tier execs in the company is really the proof.
Make no mistake, this isn't a moral decision. This is just killing the competition using morality as an excuse.
Codex cli is open source rust project (Claude Code is not). Whisper is an open weight ASR/STT model. GPT-OSS is an open weight LLM that can run locally and is still very capable. CLIP is an open source model used for image encoding. Just some of the open contributions from OpenAI
All these were released way late for a company called "Open" AI. But go on, keep defending the billion dollar corporation.
Correct. It's killing the competition. Which is what you should do if you want to run a successful business.
Then don't be pretentious and frame it as a morality statement. Just call it what it is.
2 replies →
damn... cursor doesn't get as much love as claude code or codex (on here anyways), but honestly, it's gotten amazing in the last 6 months.
i loved switching between sol / fable / grok 4.5.
... and something tells me fable might not be available on cursor for much longer :/
Try kimi k3, for code review at least. I think it’s significantly better than Sol and Anthropic for analysis, if you care about your code.
I mean, given Sam and Elon’s relationship, I don’t think anyone is surprised.
helloooo
Makes sense.
pot calling kettle black
[flagged]
[flagged]
[dead]
[dead]
[flagged]
[dead]
[flagged]
this is a pretty clean test of whether people were buying cursor or just using it for convenient access to whichever model was flavor of the week.
i work in qa so i'm naturally wary of adding another critical dependency to a development workflow, if the model you actually want can just disappear because two vendor change ownership or start argueing then the wrapper has to be really good on its own.
cursors harness and its own models may be good enough that users stay, and if they do cursor has a moat at least. if everyone follows the openai models out the door, it will just become a polished ui.
I use Cursor because I listened to YouTube interviews with the teams behind most of these products and I trusted their team the most.
How TF do you trust people you don’t know from YouTube interviews? You don’t know them. They are strangers to you trying to sell you something.
1 reply →
I think it’s important not to let AI companies control the framing here. This is anti-competitive behavior, plain and simple. The fact that xAI/Musk routinely break the law simply provides the fig leaf.
Distillation is a threat to frontier model companies, but that doesn’t make it dangerous or wrong. OpenAI can (and likely will) use the same logic to ban you from using open source harnesses.
I’m on team fuck both of these organizations, but not continuing to work with an organization that already willing violated contracts you had with them doesn’t exactly seem like a hard to justify decision.
That is to say, this isn’t a fig leaf imo.