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:

• 
  x (@xaoticatech) reported 46 seconds ago

  The reason you can't fix your AI is because you are insanely irrational and prefer your emotions over facts to conform to the popular herd because it's easy @anthropicai @openai so you just "accidentally" ignore even the most obvious facts on GitHub Xaotica.

• 
  Smukx.E (@5mukx) reported 2 minutes ago

  @NinjaParanoid @0xTriboulet @github I have asked about issue very clearly. No response from them since its an weekend... Lets see how this goes..

• 
  Lyrie.ai (@lyrie_ai) reported 2 minutes ago

  Vendor v8.8. A single *** push command — requiring nothing beyond an ordinary developer account — was enough to achieve remote code execution on GitHub Enterprise Server and

• 
  Libegato (@Libegato) reported 4 minutes ago

  Working with AI means accepting no bottlenecks. I don’t always exercise that instinct as much as I should. But a few days ago, I did! I had a local workflow problem: how to parallelize work when a single repository is ~50GB? I wanted multiple parallel workstreams, but I definitely did not want 10 full copies of the repo when I barely had disk space for one. Worktrees don’t solve it. So I built Mirage. It leverages APFS to clone a folder with virtually zero upfront disk cost, and then only pays as files are actually edited in a sweet CLI API. Suddenly BANG! I can spin up a bunch of “worktrees” fast and cheap. Now to the next bottlekneck... Github repo here: renanliberato/mirage

• 
  Lux Sp4rk (@lux_sp4rk) reported 6 minutes ago

  I'm still using GitHub for issue management like a caveman, treating them like agent cards. Gotta fix ts immediately or I too aint gonna make it

• 
  Hamza (@Hermzer1773) reported 8 minutes ago

  @zulw1337 @IBuzovskyi Yeah ngl it didn’t really work. Issue with the 2 commmands maybe? Should probs look into via sandbox new Hermes’ install. Or maybe I messed up Hermes’ with my weird usage and downloading random things. Anyway I just told Hermes’ ur GitHub repo, cloned it, and it did it itself. Used it for a trick question to see if it pulled data, it in fact did not. Unsure if the model I use just sucks or what. Anyway I just edited soul md to use it. Works great now!

• 
  Thomas Cercato (@ilpomo) reported 10 minutes ago

  There isn't a single Microsoft product that's better than the competition. They've wasted years making themselves hated in every business they've ever touched; everything is broken, badly designed, and ugly. PLEASE just sell GitHub and die.

• 
  Dao world (@Koreanteacher1) reported 11 minutes ago

  I think you are totally misunderstanding something. Pi core team is not trying to build a DEX from scratch. They are not trying to reinvent the wheel. They are trying to figure out how to apply an existing DEX model to Pi in the right way. You can see this just by looking at the website, the app descriptions, or even GitHub. What the core team is really thinking about is how to reduce the usual problems in token launches, like rug pulls and whales holding too much of the token supply. That is why they keep testing the Launchpad. It is not just about launching tokens quickly. It is about creating a safer and more practical way for real utility tokens to be launched inside the Pi ecosystem.

• 
  Ritesh Roushan (@devXritesh) reported 12 minutes ago

  @Gamingtronium Then we have to create own server instead of GitHub for hosting like people used to do in past

• 
  Jon Oringer (@jonoringer) reported 15 minutes ago

  It started with a message from a recruiter at a small crypto startup. She described a broken proof-of-concept they needed a lead engineer for and sent a public GitHub repo to review. Specifically, she asked to "check out the deprecated Node modules issue.

• 
  Anton Semenenko (@adelayida210519) reported 17 minutes ago

  @mark_k I use both, but not for the same job. ChatGPT is where I shape the architecture: problem thesis boundaries risk model roadmap public framing what should / should not be built Codex is where I push that architecture into code: small scoped tasks repo changes tests docs checks repeatable validation I try not to use Codex as “do everything”. For me the flow is more like: ChatGPT => architecture / reasoning / prompt design Codex => implementation GitHub => memory / proof / audit trail local runtime => test what actually works The important part is not the tool split itself. It is keeping generation, execution, and release separate. Even with Codex: generated patch ≠ accepted change passing code ≠ released system agent capability ≠ release authority That is basically the same principle I use in my own work: generation != release authority

• 
  Cyprian (@realcyprian) reported 23 minutes ago

  Adding CA in bio is not really my problem cos most bankr projects do it, and they've performed well.. but adding it at the top of your website or github is a No for me, man. Another one is projects constantly tweeting ca and talking about PA.. for me, it is an obvious farm. If you want to farm, farm maturely and stop making it look obvious

• 
  ᴷᴵᴺᴳ (@Adam_Sven_) reported 24 minutes ago

  @onlyzhynx @zuldotso Hope these guys get banned soon, I also reported their GitHub account and repo hoping it gets taken down.

• 
  Ame (@nooneloveame) reported 25 minutes ago

  After 6 days inside Fable 5 I'm convinced of one thing: within a year, almost every serious company will run some version of a Company Brain The teams who build it in 2026 spend next year compounding while everyone else is still wiring up their first integration You pull Slack, GitHub, HubSpot and the rest of your stack into one intelligence layer, then build an org chart on top of it: a main brain setting direction, a fleet commander running the agent fleet, specialist sub-agents handling the actual execution Your team already lives in Slack all day, so you're not making them learn a new tool You're just adding agents to the room they never leave I broke the whole build down into 14 steps Article below

• 
  fromsoftserve (@fromsoftserve) reported 27 minutes ago

  @morphys_tears oh if you mean the actual renderer itself, it's not on github or anything, although ragevitamins has said in the past that he might open source it down the road, but don't take my word as gospel as I could be entirely wrong, and I don't speak officially for it haha