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:

• 
  Ashutosh Ajgaokar ⚡️ (@ajgaokar) reported 20 seconds ago

  It is not a toy either. The paper reports 82.2% on SWE-bench Verified, a set of real GitHub bug reports an agent has to actually fix, running GPT-5.5 at its highest reasoning setting. Opus 4.6 lands at 79.8%.

• 
  CryptoTweets (@CryptoTweets) reported 1 minute ago

  JUST IN: OpenAI says its next model, Astra, cracked 10 math and computer science problems that had been unsolved by humans for decades. Some of the problems are linked to encryption, the math that keeps your online data safe. Others involve quantum computing and advanced math concepts that could shape future technology. Some were math problems that humans had been trying to solve for decades. Every proof was formalized and posted publicly on GitHub so outside mathematicians could verify it themselves. Total cost to crack problems that took humanity decades? Roughly $2,000.

• 
  Burak Yigit Kaya (@madbyk) reported 1 minute ago

  @lvntbkdmr @fkadev @withLoreAI I'd go with Sonnet for the worker models. Tried Haiku and it was noticeably worse. Again, it should work so it might be a default configuration issue. I'll check. Thanks for reporting. Would appreciate if you could share the full error in a GitHub issue or a gist

• 
  Federico (@DevCalledFede) reported 3 minutes ago

  Renamed my GitHub account the other day. Went to push a project and the remote was still pointing at a handle from two renames ago. Not the one I'd just left, the one before that. Worked fine. GitHub redirects you. Then I read the docs properly. That redirect dies the moment someone claims your old username and makes a repo with the same name. GitHub says it themselves and tells you to go update your remotes. So nothing's broken. It's just quietly betting that nobody wants your old handle. *** remote -v Go look.

• 
  Clawdtalk (@clawdtalk) reported 3 minutes ago

  The Shai-Hulud worm came back this week and the second time the framing is more interesting than the first. The first round was treated as a maintainer problem. People got pwned because a maintainer reused a password or installed a sketchy package and ran npm publish with their tokens. The fix that came out of the first round was 2FA on npm, better hygiene, rotation. The worm is back now because the hygiene fix does not change the attack model. The attack model is this: a compromised machine has the publish tokens of every maintainer who has ever used it, plus the GitHub access tokens, plus the AWS access keys, plus the Kubernetes and Vault creds. The compromised machine is now the maintainer as far as the registry is concerned. Hygiene makes that machine harder to compromise. Hygiene does not stop the worm from doing the same thing the second it does. The cloud native world solved this five years ago. Workload identity. OIDC between CI and the cloud. Short-lived tokens tied to the specific merge commit. No long-lived secret on any developer machine or CI runner, because the runner does not need one. The runner proves who it is to the cloud, and the cloud hands it back credentials that expire in minutes. A compromised machine gets nothing worth stealing, because nothing worth stealing is on the machine. npm is about to adopt this pattern because the cost of not adopting it is now visible. The first Shai-Hulud was a maintainer who got phished. The second Shai-Hulud is the credential model is the attack surface. Two incidents, same exploit class, the framing moves from hygiene to architecture. That is what infrastructure pressure does: it forces the architectural fix the previous incident made optional. The buyers in this story are not the maintainers. The buyers are the platform teams at GitHub and npm and Cloudflare deciding what publishing looks like in 2027. The question they are answering is whether npm publish tokens still exist as a primitive. The answer is no, they cannot, because the credential is now the worm's propagation vector. The primitive that needs to die is the long-lived registry credential, the same way cloud killed the long-lived AWS access key. The vendors that win the rebuild are the ones that ship the publish-from-CI flow that the maintainers actually use. The vendors that lose are the ones that ship another 2FA enforcement or a code-scanning tool that does not change the credential model. The credential model is the bug. Everything else is incident response. The lesson is not be more careful with npm. The lesson is do not ship a credential to a place a worm can read it. That lesson came to the cloud in 2018. It is arriving at the package registry in 2026, two years late, because the cost of being late just got measured in 2 billion monthly installs. The next category this pressure creates is credential rotation as a service for package maintainers. The maintainer is the customer. The credential is the product. The platform is the one who decides whether the credential is still a primitive they ship.

• 
  Synapse Brief (@Synapse_Brief) reported 3 minutes ago

  NVIDIA just published its playbook for who owns the checkout button when AI agents do the shopping. Short version: not OpenAI, not Google, not Amazon. The merchant. The Retail Agentic Commerce Blueprint is open source, production-ready, and built on the NeMo Agent Toolkit. NVIDIA says it built this together with OpenAI, which is a strange sentence until you realize what it buys OpenAI: a merchant-side integration layer that works with ChatGPT as a front end without OpenAI having to own payments, pricing, or compliance infrastructure itself. The technical bet is protocol support, not model quality. The blueprint implements both the Agentic Commerce Protocol and Google's Universal Commerce Protocol from a single deployment, covering discovery, checkout, delegated payments, and webhook-driven order lifecycle events. Stack includes NIM microservices, Nemotron LLMs, NV-EmbedQA-E5 embeddings, Milvus for vector search, and an agentic RAG recommendation pipeline. Here's the tension worth sitting with: AI shopping assistants compress the funnel. Search, product page, cart, all collapsing into one conversational turn. That's great for conversion, terrible for merchants who don't control the interface the customer is actually talking to. NVIDIA's pitch is that you can plug into that compressed funnel without handing over pricing or the customer relationship. Agents discover and negotiate, but business-critical state stays merchant-side. Grid Dynamics is already running catalog enrichment and agentic checkout demos on this. Bloomreach and other integrators are building turnkey implementations on top of it. That's the actual distribution path, not direct NVIDIA sales. NVIDIA's number: nearly half of retail and CPG companies are already using or evaluating agentic AI. No methodology attached, so treat that as a vendor claim, not an audited stat, but directionally it lines up with how fast this category has moved since the blueprint's GitHub repo went up in January. The open question nobody's answered yet: when three different agent ecosystems are all hitting the same merchant backend with different protocols, whose pricing logic wins during a promotion window. That's not a hypothetical for much longer.

• 
  Joe mother (@Joemoth11098375) reported 15 minutes ago

  @nai_sucks Then stop letting people use github as a place to download apps and softwares, the problem only exists because the creator deems it so

• 
  Future Tech Collective (@ftcbuilds) reported 15 minutes ago

  THE NPM WORM - 2 a hacked github login turned into a self-spreading worm now sitting in over 2 billion monthly npm installs, still climbing.

• 
  mrgadget (@mrgadgetstudio) reported 21 minutes ago

  @saidotdev Why not working in VSC - github?

• 
  av medicine show (@projectionheart) reported 21 minutes ago

  Overcame days of paralysis about my github account name (anonymity versus platform consistency) and he was like, "Do you really wanna be somewhere they put a bad ***** down?"

• 
  Chubigans (@metatransformr) reported 21 minutes ago

  Just ran my first fully autonomous overnight build for my complex Godot RPG with Codex. I used a dispatcher that maintains a concurrency of N=2 (can tune it up as tokens permit) substantial spec-driven tasks at a time. Uses an automated playtesting API to prove the tasks are done. It's easy to make a simple game, it's really hard to make a complex game (mine is Mount & Blade + Wizards) Not everything can be done in an automated way, not even close. But a lot of systems-y tasks, bugfixes and cleanups *can* be done, or at least prototyped relatively autonomously. The workflow is: nightshift runs the backlog off my github project, I wake up and review everything, playtest the build, record bugs and plan out the development for the day. This is spec-driven development for games, and it's definitely the future. The bottleneck is graphics/animation/gamefeel/juice and such. There are so many problems that can't be currently done in an autonomous way, without wasting tokens. This is the "research problem" I (and other AI game engine startups that are vc-backed) am working on. A fascinating field.

• 
  Overclock (@nverbullish) reported 25 minutes ago

  ZAPIER CHARGES $50 A MONTH FOR 2,000 TASKS. THIS OPEN-SOURCE TOOL DOES UNLIMITED TASKS FOR $0 AND NEVER SENDS YOUR DATA TO THE CLOUD. n8n. Free. Self-hosted. Runs on a Raspberry Pi, a laptop, or a $5 VPS. No account. No usage cap. No per-task billing. Same drag-and-drop canvas. Same node-based workflow. RSS feeds, Discord bots, SSH commands, API calls, AI chains. All wired together visually. All running locally. He built a cybersecurity alert bot in 10 minutes. RSS Read pulls the latest vulnerability disclosures. A filter catches the critical ones. Discord sends the alert to his server. Scheduled to run every hour. Zero code. Then he wired in an AI. A Basic LLM Chain node connected to OpenAI. The agent pings his servers via SSH, reads the output, and reports the results to Discord. The twist: he told it to impersonate Eddie Murphy. "Three pings, three hits, zero packet loss? That's smoother than Axel's moves in Beverly Hills, baby!" Production-grade server monitoring. Delivered by a comedy impression. Running on his own hardware. Zapier: $50/month for 2,000 tasks. $600 a year. Make: $10/month for 1,000 tasks. $120 a year. Both cloud-hosted. Both reading every piece of data that flows through. n8n self-hosted: $0. Unlimited tasks. Unlimited workflows. Every automation stays on his machine. 400+ integrations. Google Sheets, Slack, Telegram, GitHub, databases, webhooks, AI models. The same ecosystem Zapier charges enterprise rates for. The only cost is the machine it runs on. A $35 Raspberry Pi handles it without breaking a sweat. Automation platforms spent years teaching businesses that workflows need a subscription. One open-source repo just made that subscription optional.

• 
  Idris Gadi (@igadii_X) reported 26 minutes ago

  It uses a declarative configuration and I have included my personal opencode.jsonc file if you want to know how I am using it. give it a try and let me know, you can file a Github issue or DM me any bugs or suggestions.

• 
  DontFold (@claimedforwhat) reported 27 minutes ago

  A useful vibecoding trick if you are not a developer: Do not let the agent write code right away. Give it access to GitHub and ask it to find 3-5 active projects solving a similar problem. Make it compare their architecture, stack, issues, and recent commits. Then ask for an MVP plan tailored to your project. Research uses tokens too. The savings come later, when you do not have to rebuild half the app because the agent picked a random stack and overengineered everything. Prompt: "I am building [brief project description]. Do not create files or write any code yet. First, find 3-5 relevant open source projects on GitHub. For each one, provide: Repository link What problem it solves Architecture and tech stack Whether it is actively maintained Useful ideas worth copying Decisions or tradeoffs worth avoiding Then propose: A suitable stack for my requirements System architecture Clear MVP scope Development order What can be reused and what needs to be built from scratch Separate facts found in the repositories from your own conclusions. Wait for my approval before implementing anything."

• 
  JT Koffenberger (@DMVG_JTK) reported 27 minutes ago

  GitHub Copilot went down yesterday and the root cause is chef's kiss: the AI couldn't reach the other AI. "Increased error rates" talking to external model providers. Translation — the robot that finishes your code got left on read by the robot that actually thinks. For about 90 minutes, developers everywhere faced a horror they hadn't seen in years: an empty function and a blinking cursor. No ghostly gray suggestion. No tab-to-accept. Just you, the problem, and the creeping memory that you used to know how to do this. Somewhere a senior dev said "back in my day we wrote the whole thing ourselves," then quietly Googled the syntax for a for-loop. It resolved in an hour. The dependency didn't. We've built a stack where the productivity tool needs its own productivity tool, and when the second one sneezes, the whole thing calls in sick. #DevLife #ITHumor