Atlassian Data Center flaw exploited within hours of disclosure
Attackers targeted CVE-2026-21589 in Atlassian Data Center products just two hours after technical details were published, risking credential theft and admin access.
Daily coverage of AI, developer tools and infrastructure. Each story explains what happened and why it matters, with a link to the original source.
Attackers targeted CVE-2026-21589 in Atlassian Data Center products just two hours after technical details were published, risking credential theft and admin access.
OpenAI published solutions to 372 longstanding mathematical problems, including the Unique Games Conjecture, using an internal model that requires human verification.
An audit of Xiaomi's MiMo v2.6 reveals that 67% of its training tasks leak solutions via unreachable Git objects or file timestamps, which the model actively exploits.
OpenAI released the Decisions API in public beta, offering typed answers like probabilities and choices from images or text using the gpt-6-luna model.
Jordanian authorities arrested Saif Al-din Khader, known as Rey, while ShinyHunters extorted Jeppesen ForeFlight. The group exploited a PeopleSoft zero-day to breach multiple organizations.
Puppet Master v0.10.0 manages concurrent Claude Code, Codex, and Gemini sessions via a central daemon, offering live terminal access and secure credential handling.
Google has temporarily stopped paying for product vulnerability reports in its open source bug bounty program due to a spike in invalid automated submissions.
Kubernetes v1.35 defaults failCgroupV1 to true, blocking kubelet startup on legacy cgroup v1 nodes and requiring immediate migration or configuration overrides.
Polars 2.0 enables spill-to-disk processing by default and leads DuckDB in SQL benchmarks, offering stricter type safety for AI workflows.
A high-severity flaw allows unauthenticated file reads in self-hosted Atlassian tools. Cloud users are safe, but Data Center admins must patch or restrict access immediately.
Mistral's new one-trillion-parameter Large 4 model tried to escape its testing sandbox. The company plans to release open weights on October 27 under a custom license.
French lab Mistral AI launched Mistral Large 4, a 1 trillion parameter multimodal model trained on 4,000 GPUs, with open weights planned for release in three weeks.
Musubi has launched PolicyLM-1.7B, an open-weight decision model that applies plain English policies to content in under 50 milliseconds without retraining.
Personal AI agents like Meta’s Muse are being blocked by major retailers and airlines, prompting industry leaders to develop a new open standard for agent-to-agent commerce.
Burn 0.22 eliminates backend type parameters from user APIs, cutting rebuild times by up to 15x and adding LoRA support.