GitHub Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

GitHub Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Sudeep Srivastava (@sudeepsriv) reported 4 minutes ago

  GitHub might finally have a serious competitor. And it’s from Cursor. Most people know Cursor as an AI code editor. But Cursor Origin is much bigger. It’s trying to become an AI-native alternative to GitHub where AI agents don’t just help write code. They help build entire products. Think: • Source control • AI coding agents • Code review • Project understanding • Team collaboration all inside one workflow. Why developers are paying attention: Instead of manually searching through repositories, you can tell AI: • Fix this bug • Build this feature • Refactor this project • Investigate an issue • Ship a working version And AI handles much of the execution. The bigger shift: GitHub was built for humans writing code. Cursor Origin is being built for humans managing AI agents that write code. That’s a completely different future. We’re moving from: Human → Code to Human → AI Agent → Code My take: If GitHub defined the software era, Cursor Origin could help define the AI-native development era. And that’s why Elon Musk acquiring Cursor would be huge. xAI would gain: • AI models • Compute infrastructure • Coding agents • A developer platform That’s not just buying a product. That’s owning a major piece of how future software gets built.

• 
  Avi (@AvaneesaBee) reported 5 minutes ago

  @karankendre Good, GitHub is sooo slow these days

• 
  Ricardo (@Ric_RTP) reported 9 minutes ago

  Elon Musk just took Anthropic's biggest customer hostage three days before their IPO. He paid $60 billion for it without spending a dollar of cash. But the company he bought is actively losing the race he claims to be winning: The company is Cursor, the AI coding tool used by most of Silicon Valley and a huge chunk of Fortune 500 engineering teams. Its best feature is called Composer, and Composer became the most-loved AI coding product on earth for one specific reason: It runs on Anthropic's Claude. The phrase "vibe coding" was literally coined by a researcher playing with Cursor's Composer running on Claude Sonnet in early 2025. Anthropic's enterprise revenue exploded in 2025 partly because every engineer using Cursor was effectively a paying Anthropic customer underneath. Cursor became one of the largest external pipelines of Claude usage anywhere on the internet. And last week, Anthropic confidentially filed paperwork to go public. Three days after SpaceX completed its own IPO on Friday, Elon Musk exercised an option he had quietly signed in April and bought Cursor for $60 billion. The deal was announced Tuesday morning in an 8-K filing. By the time most people read the headline, the pipeline feeding Anthropic's biggest enterprise channel was already legally owned by its biggest RIVAL, days before that rival walks onto the public markets and has to explain its growth story to Wall Street. Now look at how he paid for it: Not one dollar of cash changed hands. The entire $60 billion was paid in SpaceX stock. Stock that was minted out of thin air on Friday when the company went public at $135 a share. By Tuesday, that same stock was trading at $211. So Musk used four days of public-market hype to mint $60 billion of fresh equity and immediately spent it on an acquisition that had been pre-arranged before anyone in the IPO even saw the prospectus. SpaceX investors who bought shares in the last four days got diluted by 3.4% before they understood what they owned. The IPO was literally the printing press for the acquisition. Now look at what he ACTUALLY bought: Cursor's market share among enterprise customers has been collapsing. According to spending data from Ramp, it fell from 41% in June 2025 to 26% in May 2026, bleeding ground every month to GitHub Copilot and Amazon Q. The smart money knew. Andreessen Horowitz, Thrive, and Nvidia were about to lead a round at a $50 billion valuation, which they already considered aggressive. Elon paid 20% more than that for a company actively LOSING the race. He paid premium for declining momentum. And he did this because his own AI division was in trouble. xAI has been struggling quite a bit so SpaceX needed an AI story that could survive a public-market quarterly earnings call. The fastest way to get one was to buy a brand engineers already trusted before that brand's market share slipped any further. So follow the whole chain: SpaceX went public to mint the currency. Elon used that currency to buy a fading market leader at a premium. And the seller of choice happened to be Anthropic's biggest enterprise pipeline with the timing landing in the exact window between Anthropic filing its prospectus and pricing its IPO. This was literally a hit job on Anthropic's IPO. Anthropic's next move is the one to watch. If they cannot show Wall Street that Cursor's revenue can be replaced fast, the most hyped AI IPO of the year just walked onto the public markets with a huge problem.

• 
  Meteor.js (@meteorjs) reported 13 minutes ago

  @CloudByGalaxy The goal: every push to your main branch automatically triggers the same sequence. GitHub pulls your code, installs Meteor, authenticates with Galaxy, injects your settings, and ships the build. No terminal, no human error, same result every time.

• 
  荷兰小博 (@DutchPhD) reported 14 minutes ago

  SpaceX bought Cursor. $60 billion. All stock. Four days after the IPO that made Elon Musk the world's first trillionaire. A rocket company just acquired a code editor for more than most countries' GDP. The timing is the story. SpaceX went public June 12 at $135 a share. By June 16, the stock had ripped past $200, market cap briefly brushing $2.9 trillion. With that kind of currency, a $60 billion all-stock deal might as well be free. They printed equity, handed it to Anysphere's founders, and walked away with the most important developer tool since GitHub. But here's what most coverage misses: this wasn't ambition. It was a bailout. By March 2026, all 11 co-founders of xAI had left. Musk admitted the AI unit "was not built right." Grok the model was fine but the organization around it was collapsing. SpaceX had absorbed xAI in February, folding the AI division, the X platform, and the Colossus supercomputer into a new entity called SpaceXAI. Within weeks, the talent evaporated. So instead of fixing it, Musk bought the front door. Cursor was already closing a $2 billion round at a $50 billion valuation. a16z, Thrive, Nvidia were in. SpaceX swooped in April with an option to acquire at $60 billion. The founders took the stock. Hard to blame them. Cursor had gone from $100M to $4B in annualized revenue in eighteen months, genuinely staggering growth. But there was a catch: they were running on models built by their future competitors. Anthropic's Claude. OpenAI's GPT. Cursor was a beautiful product with a dependency problem. Now that product lives inside an empire that owns the models (Grok), the compute (Colossus), the data (X), and the distribution. Millions of developers open Cursor every day. Every single one of them is now a SpaceXAI user, whether they think about it or not. The play is obvious. Anthropic has Claude Code. OpenAI has Copilot and Codex. Google has Gemini Code Assist. But none of them own the IDE. Cursor is where developers live, where they write, debug. If SpaceXAI ships coding models that work best inside Cursor, and only inside Cursor, the switching cost becomes the moat. You don't need to be better. You just need to be the default. I keep coming back to how weird this is. The company that builds rockets to Mars now controls the tool millions of developers use to write JavaScript. The line between infrastructure and empire hasn't just blurred. It's gone. The deal closes Q3. Antitrust isn't a conversation yet, but it will be. When your code editor, your AI model, your social network, your satellite internet, and your rocket ship all trace back to the same person's equity, you're not looking at a tech stack. You're looking at a vertical.

• 
  TokenFires (@TokenFires) reported 18 minutes ago

  Tired of Claude not working well? Me too. So I figured out how Anthropic has trained the model to expect to work. You work with Claude now, Claude does not work with you. Here are the keys to success with Opus and Sonnet: 1. Provide a strict set of agent instructions: - start with Karpathy’s rules - add run up and summary removals - add refusal for questions it can find the answers to - tune for preferences - enforce verification not assumptions - enforce responsibility (model performance will be discussed in retrospectives) - keep it SIMPLE though (aka: limit token burn and confusion for the LLM) - be specific about *** ops 2. Follow this workflow: [opus] research (docs/web = define source of truth) -> plan (intent and what success looks like) -> design -> task decomposition (target sonnet)-> create failing tests -> [sonnet] construction -> bug fix until tests green -> [opus] review against plan/design + test validation -> cover deploy/rollback. Then it works fine. Beats the 30 day rolling memory window Claude ships with. And/or, add a real memory system to Claude. Raw sessions + prompting went away with 4.5. Anthropic did not express this as strongly as they *could* have. But the 2026 versions expect a certain workflow now. If you work in it, it’s successful. If you skip anything or try to vibe your way to the end, it’s less likely to result in quality code. And your session will churn with flip flop changes and miscellaneous bug fixes. Claude *NEEDS* a library of good representative information to draw from through the whole process. Don’t skip doc building and providing web links with explanations (look here for this, read this for that). Try to shortcut this and the Claude models don’t “work”. Even better, build agents (or find built definitions on GitHub) that do these and create a skill walking through the whole process. I promise the result is better after the pre-work is done. I’m paid to do this and I ship AI code without the hype and vibes in my day job. Every week. Every day. Do people on X do this though? Is this a largely unknown thing outside of the software engineering field? Oh. And add hooks for delete and drop commands. And never connect AI to production. I feel like I shouldn’t have to say these things. But I know we’re only human.

• 
  Colin LeClercq (@ThePhilnado) reported 23 minutes ago

  Watch out for this one. A "recruiter" sends a GitHub repo to review, asks you to fix a deprecated Node modules issue, and the second you run npm install a backdoor fires off the prepare script and runs whatever they send. The job offer was the trap.

• 
  Matthew P. Grosvenor (@mpgros) reported 25 minutes ago

  @github - "We had a problem finding your email subscriptions." That's because I never subscribed to anything in the first place. Stop sending me your spam I didn't ask for.

• 
  Ben Canning (@benhackshealth) reported 25 minutes ago

  Spent today setting up home assistant in the gym, ran into an issue with the connection between the software that controls the LEDs around the mirror. Found a 4 year old repo on github, downloaded it, updated the code, fixed the problem and now control the lights without having to get out of my seat... Am I a hacker now?

• 
  outloop (@0utloop) reported 27 minutes ago

  I had a GitHub repo with my ex. I still create issues, but there's no response anymore.

• 
  Games from the heaps of Steam (@HeapsofSteam) reported 27 minutes ago

  Take a look at the newly released Silent Hill pc port by kushastronaut. All you need is to download the zip file from his github(links below) and find the copy of the game that you lost in someone else's server. Here you can see a bit of the start of the game and me loading a saved file from near the end of the game to test the 3rd person camera. You can also use the console to skip around and do other things. It's completely playable according to the dev so give it a try if you want but it's still in development. Cybil's bossfight was broken for me in testing, she was shooting the air lmao. Make sure to leave the dev a like. Cheers.

• 
  Kevin Tabet (@TabetKevin) reported 27 minutes ago

  @upstash Hey guys i think login with github is broken can't log in rn will try later. google works email i dont have

• 
  RY-GOD (@Moneybag_Fin) reported 27 minutes ago

  The clever part: There's no server and no database. A GitHub Action runs bam-net snapshot every 6h and commits the result straight back into the repo as JSONL. Because it's append-only and lives in *** history, anyone can replay the commit log and verify the dataset themselves.

• 
  LukeYoungblood.eth 🛡️ (@LukeYoungblood) reported 34 minutes ago

  @cursor_ai Not a bad idea... Github is core infra to everything but has been struggling under the incredible agentic coding scaling problem.

• 
  Jermaine Johnson’s Achillies (@8makesmewantkms) reported 38 minutes ago

  @The_Real_OQ You just admitted anyone can make github and do **** with it, you are legit just working against yourself, its not about whether you are “in the MLB front office or not” the fact you feel the need to prove yourself to another random person online, is only and issue you have.