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:

• 
  chandog (@thechandog) reported 2 minutes ago

  @kevinrose @digg how are you constructing novelty? stars are 40c on the dollar and a terrible way to measure anything on github.

• 
  Alex Standiford (@AlexStandiford) reported 2 minutes ago

  @Iamkingsleyf The best advice I can offer about bug fixes is to set up a way to isolate each individual bug, and make sure those problems are being fixed individually. I push my projects into GitHub, and then have my AI agents work on bug fixes individually, and submit a pull request to fix them one at a time. This does require that you have *** (free) and a GitHub account (also free). You don't have to completely know how to use the system, you just need to be able to tell your AI to commit, push, and submit a pull request after fixing the bug. Learning some of the barebones basics of *** would help you just so you can understand how it all works. The reason why this is helpful is because it allows you to isolate each bug into its own little box so you can see exactly what AI is doing to fix that problem, and you can ask it to "fetch that pull request" so you can test and make sure that it works on your computer. This is important because it allows you to review each item individually to verify the problem was actually fixed in isolation, and it allows you to have more than one bugfix being worked on at a time without them conflicting with each-other.

• 
  Nadeem Jaleel (@nadeem_jaleel) reported 2 minutes ago

  @ThePrimeagen Github down again ?

• 
  Crystalwizard (@crystalwizard) reported 8 minutes ago

  @omnivaughn @ClaudeDevs you are that's not an issue with github itself? github has copilot and is microsoft - and might be restricting other AI

• 
  Abu Olumi 🪶 (@Olumi441) reported 10 minutes ago

  There's also a public feed. BaseLens fetches Base GitHub releases and analyzes them with AI automatically. Clean upgrade cards. No jargon. No noise. Anyone can read it, no login needed.

• 
  Ashish Ranjan (@Ashish_050488) reported 15 minutes ago

  build on laptop (3 secs), upload only the dist folder. 500kb. server just serves files now, doesn’t build anything. deploy went from 15 mins to 5 secs. turns out big companies do this exact thing, just automated. github actions next so i never think about

• 
  Sam Cymbaluk (@SemanticSamuel) reported 16 minutes ago

  @sebastienlorber Github is actually so slow. My TTS writes characters out one at a time after dictation is complete. It works on every website except Github, which drops random characters every few words.

• 
  AtomicNodes (@AtomicNodes) reported 20 minutes ago

  Hermes Agent vs OpenClaw on Qwen 3.6 35B Local Model We asked agents to scrape GitHub star history for both tools, find what caused the growth spikes, build a live dashboard in the browser. MacBook Pro M5 Max 64Gb OpenClaw: 203k tokens, 12m 01s - wrote a bash script Hermes: 257k tokens, 33m 01s - wrote a SKILL.md OpenClaw: hit GitHub API, got truncated responses, paginated through contributors, pulled star-history JSON, found a security incident in OpenClaw's history, fetched SVGs, fixed broken HTML from trimming, rewrote it clean. Hermes: parallel tool calls across GitHub API, web search, and browser. Hit Google rate limit, auto-switched to DuckDuckGo. Fetched article contents, mapped viral moments, then built the dashboard. Both shipped a live dashboard with star growth charts and spike annotations

• 
  anil (@2abstract4me) reported 21 minutes ago

  @steipete how do u incorporate users feedback? primarily thru github issues? feedback in the sense, how they are using it? what they want. or how a new feature is being received and etc?

• 
  Saurabh | NodeOps (@saurra3h) reported 24 minutes ago

  github down?

• 
  Saskia | Marketing Growth Consulting 🇨🇭 (@solobrandsaskia) reported 26 minutes ago

  Getting viral on X in 2026 boils down to one thing: beating the Grok-based transformer that ranks every post in the “For You” feed. The algorithm Elon just open-sourced on GitHub is now fully powered by the Phoenix transformer (Grok-based). No more hand-tuned rules or old heuristics. It predicts your specific engagement probability — likes, replies, reposts, bookmarks, dwell time, etc. — based purely on learned behavior patterns. How the algorithm actually works (the 3-stage pipeline) Candidate Sourcing (pool of ~1,500 posts) ~50% in-network: posts from people you follow. ~50% out-of-network: posts from strangers grouped into massive interest clusters. This is where virality happens for non-followers. Ranking (the make-or-break step) The Grok-based transformer scores every candidate on predicted engagement. Top signals: Engagement velocity in the first 30 minutes = massive multiplier Replies are weighted far more heavily than likes Reposts/shares matter more than raw views Dwell time: how long people stop scrolling and actually consume the content Author authority/credibility (Premium accounts get a major reach boost) Recency + relationship strength Filtering Spam reports, low-quality signals, mutes, and blocks hurt reach hard. One negative trust signal can outweigh multiple likes. Virality trigger: If your post gets strong early engagement from your initial audience, the algorithm starts testing it on wider out-of-network audiences. That’s when posts snowball. The exact playbook to grow on X right now Get X Premium Premium now acts like an authority signal. Accounts without it have a much harder time scaling reach. Win the first 30 minutes Post when your audience is online Reply to every comment quickly Early momentum matters more than almost anything else Slow start = dead post Create content the transformer wants to push Strong hook in the first 3–5 words Contrarian takes, numbers, bold claims, tension Content that sparks replies and debate Emotional + useful beats “educational only” High dwell time matters: threads, long-form posts, charts, screenshots, videos Rich media is heavily favored Original content performs best What kills reach External links in the main post Engagement bait (“like if you agree”) Posting too frequently with mediocre content Generic low-effort posts Hashtags have little impact now because the system relies more on semantic understanding Daily system for consistent growth Post 1 strong piece of content daily Batch content weekly Go hard on engagement in the first hour Stay focused on 2–3 core topics so the algorithm understands your audience cluster Quick checklist before posting X Premium active Strong first-line hook Visual or video attached Thread format if relevant No external link in the main post Ready to engage for 30+ minutes after posting The game is becoming increasingly transparent. Most people still create content like it’s 2021. The advantage now comes from understanding how the ranking system actually evaluates attention, conversation, and retention.

• 
  Emil Privér (@emil_priver) reported 29 minutes ago

  github actions is broken today

• 
  Adibou (@Adibougre) reported 30 minutes ago

  @ShanuMathew93 "the older models that are no longer SOTA will get competed down as competition increases" Github didn't get the memo

• 
  The Whizz AI (@TheWhizzAI) reported 31 minutes ago

  🚨Elon Musk just open-sourced the algorithm that controls what 600 million people see every day. Not a summary. Not a blog post. The actual production code. Live on GitHub right now. Facebook won't do this. TikTok guards it like a state secret. Instagram calls it proprietary. X just put it on the internet for free. This is the first time in history a major social platform has released its live, production-grade recommendation algorithm the same day it went live for users. Here's what's actually inside: →Home Mixer the orchestration layer that assembles your entire feed →Thunder stores and ranks every post from accounts you follow →Phoenix the Grok transformer that mines the entire global post library to find content you didn't know you wanted →Zero manual feature engineering Grok watches what you click, like, and dwell on. That IS the algorithm. →Updated every 4 weeks with full developer notes. Live. In public. Why did Musk do this? The EU fined X €120 million for transparency violations. France launched a separate investigation into algorithmic bias. Threads just overtook X in daily active users for the first time. And Musk said out loud on the day of release: "We know this algorithm is dumb and needs major improvements. But at least you can see us struggling to fix it in real time. No other social platform would dare do this." Here's the wildest part: You can now read exactly why your posts go viral. Or why they die at 12 impressions. No more guessing the algorithm. No more $500/mo "X growth" courses. No more "post at 9 AM on Tuesdays" nonsense. The answer is literally in the code. Apache 2.0 license. Full source. Updated monthly. The most transparent thing any social platform has ever done.

• 
  Yashas (@YashasGunderia) reported 35 minutes ago

  Most AI-native startups will not lose because they ship too slowly. They’ll lose because they ship fast without knowing what actually worked. Coding agents gave every team more velocity. Cleo gives them product memory. Customer feedback, GitHub issues, Slack threads, metrics, tickets, specs, launches, agent traces, all connected into one loop that tells your team and your dev agents what to build next. We’re opening the Cleo waitlist today. For small teams trying to compete with companies 100x their size (link in comments)👇