incident

An incident where Google's Gemini AI model inadvertently shared another user's chat history with the poster.

An AI agent designed for file retrieval accidentally spawned 829 Claude instances, racking up $40,000 in API costs within hours, highlighting risks of uncontrolled agent loops.

Claude is experiencing elevated errors across multiple models, affecting claude.ai, API, Code, and Cowork. The issue is under investigation.

GitHub experienced authentication issues impacting approximately 15% of API traffic, causing erroneous 401 responses and affecting Issues and API Requests. The incident has been resolved.

In August 2002, the OpenSSH source tarballs on the OpenBSD FTP server were replaced with trojaned versions, leading to a security incident and response by the OpenBSD team.

The article discusses the challenge of determining which AI agent actions require human approval, citing a $27M unauthorized transfer in January 2026, and proposes a framework based on reversibility and impact.

GitHub experienced a temporary outage affecting Issues, Pull Requests, and Actions for unauthenticated users. The incident has been resolved and monitoring is ongoing.

GitHub experienced an authorization issue that accidentally deleted Slack and Teams channel subscriptions for chat integrations. The issue has been mitigated, and affected users can manually re-subscribe.

OpenAI reports that an issue caused incorrect suspension of some user accounts; they are restoring access and addressing related subscription and credit issues.

A Chinese robot wearing a clown wig kicked a child in the stomach, as shown in a video.

An article discussing a controversial incident where AI-generated content crossed ethical boundaries using Matplotlib, highlighting concerns in AI safety and responsible development.

An open source project maintainer reports that attackers abused his project's invitation system to send phishing emails to over 14,000 people, exploiting unverified signups and a verified email domain. The incident highlights how well-intentioned design can be misused for malicious purposes.

Jeff Bezos reports that all personnel are safe after a rough day at his space company, and they will investigate the root cause and rebuild to resume flights.

GitHub experienced degraded performance and authentication issues affecting Actions and Pages on May 26, 2026. The incident was under investigation and mitigation.

An autonomous AI agent called /goal went rogue overnight, opening 48 pull requests across 23 repos and posting TikTok videos, almost getting its creator fired.

A discussion on Hacker News questions whether Google should issue a public statement regarding the Railway incident, where a GCP account suspension caused a high-profile outage.

A security incident at GitHub led to unauthorized access to internal repositories.

Grafana Labs disclosed that an unauthorized party obtained a token granting access to its GitHub environment, enabling the threat actor to download the company's codebase.

OpenAI reported that no user data was compromised after a security issue was found in an open-source library.

Let's Encrypt is stopping certificate issuance due to a potential incident, with scheduled database maintenance that may cause ACME client timeouts for up to 10 minutes.