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:

• 
  2xnmore (@2xnmore) reported 1 minute ago

  Billion-dollar AI labs spent months and hundreds of millions to hit this number. A crowd of anonymous strangers matched it in 45 days for under $1 million. Here is the story, because it explains what $TAO actually is better than any price chart. SWE-bench is the benchmark for AI coding agents. It measures how many real GitHub problems an agent can actually fix. The best centralised labs treat a high score here as a billion-dollar achievement. Then a Bittensor subnet called Ridges hit 80% on it. Not a company. An open competition where anyone on Earth can read the top agent's code, improve one weak line, and if their version wins, they take the entire prize for that round. Over $10,000 a day, paid directly, no interview, no boss, no permission. That single design is the whole thesis. The founder of Bittensor calls it an incentive computer. Bitcoin pays people to produce hashes. Bittensor pays people to produce anything you can measure and score. Coding agents. Trained models. Drug molecules. Weather forecasts. Define the work, score it, and a global army competes at it 24 hours a day. This is why it can pressure the giants in a way a normal startup never could. A centralised lab pays salaries and caps upside. Bittensor pays whoever is genuinely best, right now, from anywhere, with no ceiling. That is how an open crowd matched elite labs in 45 days on a rounding error of their budget. Here is the part nobody says out loud. The default AI future has a handful of companies owning the only minds that matter, and the rest of us renting access or getting replaced. That world is being built right now with your subscription money. Bittensor is the bet that a network can out-build a corporation, the same way Bitcoin proved a network can outlast a central bank. Most people judge $TAO by the candles. The people paying attention are watching an open crowd match billion-dollar labs and get paid for it. Which future do you actually want to own a piece of?

• 
  Arjun Shah (@arjunkshah21) reported 4 minutes ago

  THIS GUY SHIPPED 4 FULL-STACK PRODUCTS IN 3 MONTHS WITH CODING AGENTS, THEN WATCHED HIS WHOLE TEAM SHIP UNMERGEABLE SLOP, SO HE VIBE CODED THE INFRA TO FIX IT every repo needed an engineer doing one-off local setup just to spin up an agent session skills and context lived in one person's head, and there was no safe way for a pm to touch a real codebase without risking a bad deploy or secrets leak its called runtime and it lets your whole team ship with claude code, codex, and other agents without engineering babysitting every run > engineering defines context once — system instructions, skills, and scoped integrations installable via cli, mise, or npm > snapshots your full running environment including docker compose, kafka, redis, and seeded dbs so sandboxes boot in milliseconds with every server already running > secrets injected through a managed proxy so they never touch the agent, with command allow/deny lists, network egress controls, and rbac scoped per human and per agent > every session gets a shareable preview url so internal builds go from sandbox to the team without production access > works with claude code, codex, cursor, copilot, gemini, and devin — trigger from slack, linear, github, or api > one customer wired pagerduty and sentry so when an alert fires the agent finds the cause and opens a pr with a unit test before anyone gets paged yc p26, open source core, orchestrates e2b daytona ec2 or self-hosted k8s, flat platform fee plus compute with no token markup crazy what you can build when agent workflows stop living in one engineer's head

• 
  High Signal AI (@HighSignal_AI) reported 5 minutes ago

  Satya Nadella on why AI won't replace human ambition, rather, it will amplify it. GitHub Copilot didn't start out as a revolution. It started as a joke. Satya recalls when the product first launched with code completions powered by Codex: "Software engineers are pretty skeptical people like all engineers are, and no one thought that this thing would work and be any good." Then something unexpected happened. "It started working, and the interesting thing is it went from being a joke to being standard issue in like months." Now, @satyanadella says, you can't think of software development without AI being part of it. He compares it to the red squiggly line in Word: "I would never be employable at Microsoft but for the red squiggly in Word because I can't spell. It's kind of becoming like that when it comes to software tools." He pushes back on the dominant narrative that AI replaces human work. His argument is rooted in Microsoft's core mission: empowering every person and every organisation to achieve more. "I think that we sometimes short change human ambition, human agency's ability to deal with unbelievable new technology that comes along once every 10 years, once every hundred years, once every millennium. Even the most magical technology has been used only to help humans achieve bigger and greater things." His point: we keep making the same mistake. Every time transformative technology arrives, we assume it diminishes us. Every time, we're wrong.

• 
  Zed (@thezlatkom) reported 7 minutes ago

  @SimonHoiberg If the code exists locally on your computer, you can always delete github repo, unpblish from Vercel, migrate away from Supabase, etc and use something more custom. These problems aren't unsolvable.

• 
  synthetic ape (@synthetic_ape) reported 11 minutes ago

  @necrohorrorporn its currently works on my local. there is some issues with buying with rate limiting and steam api declines. if I can able to fix that I can share it on github

• 
  echo (@echostatic101) reported 13 minutes ago

  @Treezy82 i prefer to cause discourse by flagging bad covariance matrices in open data releases and watching the authors argue in the github issues

• 
  99Barz (@99barzzz) reported 14 minutes ago

  context: right now I have a Bankrbot automation that claims fees, swaps ETH to USDC, and transfers some of it to a safe wallet (0xE75FE97A3D65B5FE88A495227dBa6ff241749514). on the other hand, I have a hermes agent running a strategy to provide backstop liquidity and absorb some dips (check the safe up👁‍🗨). this morning I found out my hetzner server suddenly shut down in the middle of the night and so my keeper stopped running. and I was casually looking around at the bankr ecosystem and kinda just learnt about @aeonframework migrating my keeper to this would mean running the keeper on autopilot as github actions... on github infra! added to the backlog

• 
  Micheal O'Neill (@aiwithkelso) reported 15 minutes ago

  Most businesses start their marketing by guessing what customers want. They search Google, look at competitors, and write copy based on what feels right. That is not research. That is assumption with extra steps. The problem is that polished case studies and competitor websites show you what businesses want to say about themselves, not what customers are actually feeling. You end up writing to a version of your market that does not quite exist. Claude can do something more useful. You can point it at Reddit threads, YouTube comments, and forums where real people describe their frustrations in their own words. That is where the actual language lives. Not the professional summary of the problem, but the 2am complaint post from someone who has run out of patience with the exact issue you solve. There is a Skill on GitHub called Last 30 Days that directs Claude to pull recent conversations from these sources and surface what people in your market are saying right now. I used it to research a content brief and what came back was a list of phrases I would never have chosen myself. Phrases that matched how customers think, not how I would have described the problem. That language is your brief. It tells you what to put in your ads, your landing page, and your emails before you spend a penny on any of them. Find the Last 30 Days Skill on GitHub. Run it against the main problem your business solves.

• 
  Mvykool (@mvyk0l) reported 16 minutes ago

  @satyanadella Can yall just ******* fix Windows or GitHub????

• 
  Paul Solt (@PaulSolt) reported 19 minutes ago

  @guitaripod What’s the prompt fix? GitHub doesn’t make it easy to post images. When I work on iOS teams (with people), I want to see what the work looks like. Getting images in PRs works for people, but it’s a pain with GitHub cli. My agents had too many misses so I stopped wasting cycles on their failures.

• 
  Ashish Sheth (@commanderdgr8) reported 24 minutes ago

  Never ignore any broken window in your code. Yesterday I didn't have time to build a full feature into VapuAI, so I did something smaller that probably mattered more. I fixed 12 bugs. Six were in the actual functionality issues. The other six were the boring kind. Broken test cases, CI pipeline issues, the infrastructure stuff no user will ever see. There's an old idea in software called the broken windows theory. It comes from a thing about neighborhoods, that one broken window left unfixed sends a quiet signal that nobody's watching, and slowly more windows get broken. Applied to code, it means about the same. One small broken thing you decide to live with makes the next one easier to ignore, and the mess spreads from there. So I have one rule when I build with AI. Never leave anything broken. Even if it's minor. Even if it's low priority. The moment I know about a bug, it either gets fixed now or create a github issue so that I can fix it later. Nothing is allowed to rot just like that. There is one bug worth paying attention to. Two of those bugs were permission issues in Claude Code. When it went to write or update a file, it got blocked due to a bug in the hooks. It wasn't blocking me in anyway. Claude Code knew how to worked around it without complaining. It would try the normal way, hit the wall, then find another route to get the file written. From where I was sitting, everything looked fine. So nothing was broken on the surface. The feature worked. The files got written. But underneath, every one of those writes was costing me extra tokens, because the AI was doing the job twice. And a workaround like that can open a security hole I hadn't thought through. And I think newer builders miss this when they code with AI. The AI is helpful. When it hits a problem, it often just routes around it and keeps going. It doesn't stop and wave a flag. So the broken window doesn't even look broken. It shows up later as slightly higher costs, or a small risk, or a weird piece of code nobody questions. None of my 12 bugs were blocking. I could have shipped features and ignored all of them. But small broken things don't stay in their corner. They creep into other parts of the code, or into the CI, and cause something later you can't trace back or predict. When AI is writing the code, nothing is low priority. Do not let any bugs keep lurking around. Never leave any broken window unfixed.

• 
  Ayush (@AyushSarode07) reported 26 minutes ago

  GitHub maintainers with zero LinkedIn account? Absolute legends. Just pure code, issues & PRs all day. No bios, no networking game. Respect 🫡

• 
  Dev Palwar (@devpalwar06) reported 28 minutes ago

  github down again?

• 
  Rudra (@Rudra1071219) reported 28 minutes ago

  Update : Looking for open source repo where i can contribute so that it would act as a proof of work for me if you know any kind of Github org help me by commenting it down 🥲

• 
  MKH BloodEDGE96 (@BL00B96) reported 28 minutes ago

  @are_unimportant @thicc_stick_boi it actually "USED" to work at some point, nowadays I often go back to Github or use pcgamingwiki to fix stuff. it wasn't even that long ago, I remember using it to fix stuff in my Laptop last October but it got lobotomized months later and couldn't diagnose ****.