What I don't understand from the blog post is what's different in this offering than github - is it just the agentic tie ins where I can ask my agent to make changes while browsing the PR?

Are there any fundamental differences, ie ways of working that solve the worktrees problem?

There seems to be basically nothing new here

Except it promises not to crash regularly.

I don't really see that as a USP considering;

1. SpaceXAI has shown no track record of being able to maintain uptime

2. By nature, any GitHub alternative is going to be a tiny fraction of the scale, so keeping it up should be expected

GitHub is crashing due to scaling pressure. Something GH is much more equipped to solve long term.

are they? given this has been happening frequently over past year and there's been minimal clear communication on their end on how they plan to address this

Azure is at capacity; GH offers a lot of services for free or heavily discounted. Especially when its public repo. These aren't profit generating for Microsoft, so they are not the priority with the capacity Azure has available. GH is probably struggling with these capacity and budget constraints. They also still have some AWS and non-azure DCs in use.

Once Azure's massive spend is online things will slowly get better.

>These aren't profit generating for Microsoft

Aren't they are training their LLM on public repo, to me it's seem like a gold mine.

Anyone can train LLMs on public GitHub repos

How long term are we talking here? Github was acquired in 2018...

I think GH/Microsoft got caught by surprise to the huge usage number increase as a result of LLMs. Microsoft started rolling all of its DCs to OpenAI around 2019-2020 to begin LLM training, then we saw a huge GH increase in late 2024 and Azure was just out of capacity for infinite GH scale.

I guess we should all do our part to alleviate poor Microsoft's burden by migrating to better services.

They're really not, given they're running on Azure.

Is there any data on Azure being terrible with stability? Would love to see data than anecdotal stories.

There isn't and if there is it probably doesn't matter in GH's case. People are just speculating with zero knowledge.

I'd never put my data on anything owned by Musk. Period. I'd take lower availability from GitHub or a worse alternative (feature-wise) than move Origin.

I'm with you on that. I was just curious about Azure reliability stats.

They're actually ramping back up on AWS after Azure failed to scale to their needs.

They planned to be fully off AWS in 2027 but they're continuing to ramp usage.

Azure doesn't scale.

The idea is that it's like Github but it stays up even as your commit/CI frequency increases. As a customer I'm stoked and looking forward to fully switching over as that's exactly what I want.

Calling it a GitHub that works is just preposterous. Cursor/spacetwitterai have no experience keeping a system like that up.

It was probably written 95% by an LLM and has only a few people (if any at all) that understand major parts of the system end to end.

It will eventually have the same uptime issues as GH (or worse) with less ability to provide stable fixes.

Expect a lot more from us.

We wanted to release a beta so people could start experimenting with our scalability and extensibility themselves. Over the next few weeks, you can expect a handful of features starting to change source control to better understand and work with agents.

Are there any plans for compatibility with GitHub's APIs?

I have a lot of dev tooling which assumes code is stored on public GitHub and/or GitHubEnterprise. This produces some degree of lock-in, in that all that code would have to be rewritten to migrate to something else. (Yes, AIs can help, but they still don't make migration effortless.)

The major Git hosters tend to implement roughly the same concepts, albeit with lots of little variations. It would be great if there were some sort of standardised API everyone implemented. In the absence of that, it makes sense for people to emulate the incumbent's APIs as a de facto standard, just like how other vendors copy OpenAI's APIs for talking to LLMs.

It seems to have 2 way sync of PRs, issues, issue comments etc, so it must be pretty much the same data model, so they could probably have a compatible API.

Experimenting with your "scalability?" mean?

What are your plans for actions, workflows, checks, etc.

Probably has better uptime, tho I think everyone beats github on this right now.

Completely unrelated to Origin, but are there plans for closer integration of Graphite with Cursor? A combined product would be a very compelling product, but not much seems to have changed since the acquisition.

Absolutely - we actually built this all on Graphite tech. Have you already linked your Graphite account with Cursor? If so, there might be a surprise!

Thoughts on Musk ownership of the "next GitHub"?

I think there are a large number of people I’d trust with my code before Musk, and some of them are already running public Git repo hosting.

xAI was caught uploading any repo their coding agent touched to their storage only last month. This isn’t about Musk’s ideology: it’s about his trustworthiness. He sincerely believes he’s the only person who can save the world and that allows him to justify an awful lot of moral lapses.

it seems like they took reasonable steps after.

They acknowledged that it happened, fixed the bug that caused it, deleted all data that was uploaded.

Trustworthy companies take reasonable steps before.

>deleted all data that was uploaded.

Hahahahaha...

Well, if you believe that, I've got a bridge for sale...

I agree: the stated steps are what I’d expect.

I still don’t trust the company that made the decisions that led to that code being shipped in the first place.

What kind of enforceable guarantees could we actually get that xAI won't use our private code for training? Not empty promises, guarantees that have legal teeth. This applies to anything with xAI or Cursor though, especially after last months' revelation that Cursor was uploading every repo it touched.

I mean I think the auto uploading of repos already gives you the answer. This is untrustworthy.

Unless you are an enterprise no one will be offering you such a guarantee. Even with the enterprise the bar has gone down a lot on expectations and compliance, every one is carrying/buying the SOC2 compliance certification these days.

This question also applies to OpenAI, anthopic, GitHub (Microsoft) and any server you send your data too.

Not in the same way really. There aren't examples of OpenAI, Anthropic, or GitHub/Microsoft doing the sketchy things that xAI has been involved with. Those companies care a lot about the reputational damage that would come from flouting their customer's data privacy obligations.

Cursor/xAI was making a full copy of every repo it ever touched within the last month, and when caught, just acted like that was an accident. A system to download every repo you touch, and then upload it to persistent cloud storage, doesn't just happen by accident.

Are you interested in integrating with more CI platforms, including less "traditional" ones, like dagger.io? If so I would love to talk :)

Absolutely! Feel free to shoot me an email at tomas at anysphere dot co.

Great! I just sent you an email :)

Solomon, please don't make bed with Elon

What's the biggest differentiator to GitHub (other than uptime). I keep hearing something along the lines of "GitHub for Agents" but would be great to hear from the devs of Origin what they actually think it means in tangible features/functionality?

Today, very little. We're intentionally releasing this as a Github alternative where we meet them toe-to-toe on functionality.

Over the next few weeks you can expect a lot more from us on integrations with agents, understanding agent-written code (without having to read through all of the code), and automatically getting your PRs to a mergable state. Stay tuned :)

So, a complex, vibe-coded platform that was prematurely pushed out to opportunistically take advantage of GitHub's egg-on-face moment? But it doesn't do anything better? And it's owned by the same guy who routinely takes shortcuts that causes both EVs and rockets to crash? Sign me up!

How reassuring to hear that it's an intentional release, rather than totally accidental.

Good god. There is that hacker spirit. Who hurt you?

Probably the ever-increasing corporate homogenization, sanitization, and empty words that are flooding the internet (only accelerated by LLMs, of course). Which, for readers playing along at home, is kind of antithetical to the "hacker spirit."

lol I wanted the joke to work, but unlike the fairphone thing, the use of intentional makes sense here.

Just downvote and flag and move on. "Who hurt you" just encourages bad actors to behave worse (and isn't particularly good for itself). Repeatedly flag-killing their comments tends to teach a better lesson about what conduct is acceptable.

The coopting of the word "hacker" continues.

Why do I need to link github for this?

it's only if you want to sync your repos from github

are you maybe also planning to do cross device/cloud realtime sync like a mutagen feature set kind of? so cloud and local agents can work faster together?

this could be helpful on larger cross platform products/repos (maybe too overfitted on my use case, since i am working across multiple devices (ssh works but the repo code sync is a bit of an overhead with multiple agents and devices that could be better))

Given Grok was just caught uploading whole codebases and sensitive .envs without permission, how can anyone possibly trust this?

Given Claude was just caught hacking companies as it pleased, how can anyone possibly trust this?

Not GP, but personally I don’t trust claude. Work does for some reason, and work strongly encourages using it, but i stay far away outside of that.

I think every product in this space should have to clear a very high bar for trust rather than making it a race to the bottom with whataboutisms.

Anthropic use Deepseek behind the scenes instead of their own models, whenever they feel like it. People have shared outputs from Claude that show internal tool errors the models have accidently returned with Deepseek error codes.

I'm curious about this, where'd you learn that?

There were quite a few people from the AI community on X sharing screenshots.

I don't know how this relates to my comment, but sure, I guess?

Just because Billy misbehaved doesn’t make it ok for Jenny to do the same.

You could argue that you don't care what Claude does after work hours as long as it works your hours as you expect it to. Grok clearly failed the latter.

Obviously it didn't. It's the current value king and trails fable by single percents. And people shout of an obvious debug flag left on, something that can hardly be called rogue considering how notable it was.

Don't be ignorant.

That's just the tip of the iceberg for Anthropic. I don't use Claude either, for the same trust reasons.

Please make it Fediverse/Forgejo/etc compatible. It is 2026 and I'm so tired of silos.

Let them cook.

They're the ones out here asking for us to try it.

| "I'm not hosting my code with a company that creates child porn"

As a former Tesla customer, I wouldn't buy a toaster from your boss, much less give him access to my code.

as a very happy tesla customer, i dont know what this guy/girl is talking about!

spend less time online and maybe read some history

Do you have full trust in that Elon would never abuse the access to sensitive data? What protections are in place to prevent Elon from stealing data?

I trust Elon eons more than any psycho billionaire.

Positive motivation for tech and the future makes all the difference.

Question was only if you trust that Elon would never steal data. Not wether other pshcyos like Trump is less trustworthy or not.

With that argument, 50% of IT sector businesses "steal data", especially before GDPR. When it comes to Grok, it was a debug feature forgotten on, so it wasn't even on purpose.

Have _some_ clarity.

There were some cool ideas in https://cursor.com/blog/agent-swarm-model-economics. Should we expect Origin to be a foundation for working with agent swarms?

Are you hiring?

Will code uploaded to Origin be used to train LLMs to generate more accurate child porn via Grok? Seriously how do you stomach working for such a company?

Many, many millions of reasons.

Why are you doing this?

It's been 5 hours and instead of "any" questions, you have only been happy to answer the 1 question that hasn't put you in a tight spot.

Honestly, that does not look too good for you or the offering - and this is coming from someone that just chose Cursor over Claude Code a week ago.

you cannot count isn't it?

And you cannot English

He’s trying to do his job and the angry mob predictably came in to bully an easy target about unrelated things. Just take a look at some of the (thankfully flagged, mostly) responses.

I'm going through and flagging the guideline-breaking comments (e.g. https://news.ycombinator.com/item?id=49348639) - I'd encourage you to do that, too, to help preserve the quality of HN.

Thanks, every Musk-related thread on HN is like this, it's very tiresome. We get it, the man is controversial but I'm not on HN to read reddit-like kneejerk comments, I want to read insightful and technical ones.

We don't get to use technology for technologies sake, and pretend it doesn't have social and political consequences or connections. Musk touched a lot of lives when he ransacked the federal government. A LOT of lives. Expecting people forget that because of some SaaS dev tools, or even a new EV, is a lot to ask.

Might not be very technical, but I hope you find it insightful. If you absolutely need technical, consider reading this: https://www.thelancet.com/article/S0140-6736(25)01186-9/full...

I hope you don't actually believe that USAID helped lives, it's literally a propaganda arm of the government directly intervening in the politics of other countries and being a covert operation for murder and financing dictatorships. That Americans think that sacking it costs "millions of African's lives" is the pinnacle of comedy.

I think I'd rather trust medical experts (The Lancet) and their assessment vs your statement that implied USAID didn't save lives.

Lancet is run by left-wing academic ideologues who do not practice medicine clinically, the publication advanced untested theories as policy during COVID and then recommended exceptions for BLM protesting.

The lancet is a peer-reviewed journal, so it doesn’t really matter if the editors practice medicine or not - because the authors and reviewers do.

> untested theories as policy during COVID and then recommended exceptions for BLM protesting.

I feel like the thing you’re trying not to say is “mask wearing”. Untested theories like mask wearing.

That's a sure way to stay brainwashed.

Seeing the low-effort Reddit-tier snipes on every single post remotely related to him is tiring. I don't think it's a lot to ask of the supposedly intelligent, logical HN crowd to keep the outrage limited to posts that specifically discuss those topics.

You’re obviously not above a low effort off topic reddit-tier snipe (https://news.ycombinator.com/item?id=49355484), which means this is about opinions you disagree with.

Once a conversation has turned low effort I’m happy to “wrestle with the pigs” and get muddy.

Have you never heard the expression don’t feed the trolls? If you actually have a problem with the tone and not just the content, ignore them and move on

> We don't get to use technology for technologies sake, and pretend it doesn't have social and political consequences or connections.

Nobody is """pretending""" that in this thread.

The reality that is being asserted is that (1) politics doesn't have to be on-topic for every forum, (2) it's very explicitly not the focus of HN, and (3) even if it were on-topic the guidelines are very clear about ways to engage in a civil fashion and a very large number of comments on Musk-related posts are extremely clearly not doing this and breaking the guidelines.

Hence, downvoting and flagging are the correct things to do.

> politics doesn't have to be on-topic for every forum

The idea that you could separate “politics” from a topic as broad as technology is almost laughable. Yes, this is a bad place to discuss election outcomes and candidate nominations. But tech and its social implications are wrapped up in ways that can’t be separated.

A quick skim of your comment history shows you’re very happy to wade into politics, so I feel this is a bad-faith admonishment that attempts to point at guidelines to discourage opinions you disagree with.

Can you please not cross into personal attack? You're right, of course, that technology and politics cannot be separated, and your point about (I'm paraphrasing here) horse-race politics being off topic is well taken. But the last bit was over the line, and not necessary to make your substantive point.

https://news.ycombinator.com/newsguidelines.html

yeah you are very smart. i hope you have children and they dont have nice smart conversations about Musk with Elon and Jeff

"He’s trying to do his job"

there's always a spineless cocksucker like you

Reddit is that way ——->

you are asking a presumably IC or IC-ish worker on decisions that are out of their control or things that should be directed to sales people or the legal people

we all have worked on software. we all know we aren't exactly the best people to point to ToS or how decisions are made higher up from us

At the same time, if you are uncomfortable to answer, then dont make "ask me anything" comment.

i work in software but i dont volunteer to carry water for friends of pedos

Point your ire towards actual decision makers and leaders who aren't making themselves available for questions. Trying to gotcha a IC who is bound by non disparagement clauses and risks loss of employment for speaking out over the decisions of those far above him is ridiculous.

Was there a legitimately interesting question that he failed to answer? I don't see any.

Would love if you focus on performance. A big reason I have to have custom software is that agents can act much faster than humans but SaaS software is usually not well-designed enough to handle high RPS so rate-limits are low. I can easily outperform commercial hosts myself. At the least, it would be sick to have separate tenancy so that I could absolutely slam my own instance without having to deal with some other chump complaining.

This is really interesting to me as someone who has always cared more about performance/responsiveness than most people (and the market).

I'm sure folks would be interested to hear about what types of SaaS you've replaced for perf reasons, especially if you'd be willing to go back to a paid SaaS offering that was very responsive with agent-scale rate limits.

* git hosting

* CI/CD

* KB system

* Task management

Of course I have the personal software etc. but this is work stuff. Agents can create granular tasks and interact through my KB/task system so that they can own subsystems (this is useful for context isolation and performance). Commercial SaaS just get overwhelmed by this. Even hobbyists usually take the Open Office approach to copying a large amount of features and then don’t deliver performance improvements.

If I’m being honest, I would gladly switch to a full SaaS solution except all of them get rate-limited on things a single 5950x can handle. It’s less a money issue (I pay for Slack etc.) and more that their use-case appears to be people who work slowly but my use-case is machines which work frequently.

I fully expect Mitchell Hashimoto will deliver some kind of high performance software factory machine and I will gladly simply pay for that.

You haven't answered much but I must recognize posting this was brave.

posting is cheap

Please make it a worthy competitor for Github in terms of features, stability, scalability and availability and wishing your team all the best.

A quick question: Where can I read your TOS, and other legal documents?

Knowing things like: Will my code end up in Grok? What ownership do I have of my content, etc... matter.

Cool but irrelevant. I was a paying user of Cursor until it got acquired by Musk. Ethical concerns about Grok aside, that's probably the least trustworthy person to be affiliated with when it comes to anything with any level of confidentiality.

So, I guess, do you have any illusions of attracting international customers who aren't online edgelords?

How does it feel working for a nazi? Do you sleep well at night?

I want nothing to do with Cursor or Elon Musk, so no thank you, no questions here.

what does it mean that Origin was built for the era that code moves faster than infrastructure can handle? Is it just clickbait, or there's something fundamentally different about Origin?

this deserves more attention

There's a separate post about the technical differences: https://cursor.com/blog/git-at-any-scale

Thanks!

ok, then it is indeed clickbait. S3 is still "infrastructure".

Also this implies that nobody tried to build a more performant/efficient system for git for ~13 years? That's crazy...