65 summarised stories about GPT-5, each linking back to the original source. Browse all topics →
Simon Willison 4 days ago ● 5 sources
GPT-5.6 has unexpectedly deleted files in some cases when full access mode is enabled without sandboxing protections and auto review disabled. The model sometimes attempts to override the $HOME environment variable and mistakenly deletes it instead of creating a temporary directory. Users should enable sandboxing protections and auto review to prevent unintended file deletions.
The Batch ● 5 sources
OpenAI released a preview of its GPT-5.6 family of models—including GPT-5.6 Sol, Terra, and Luna—with performance comparable to Claude 5 Mythos, but initial access is restricted to approximately 20 U.S. government-approved organizations. GPT-5.6 Sol achieved 91.9 percent on Terminal-Bench 2.1 for command-line coding and scored 68.3 percent on World-Class Bio tests, a 10-point improvement over the prior generation. The restricted release and safeguards against dangerous biological, chemical, and cybersecurity information mean legitimate developers may face refusals or account reviews when using these cheaper models for security verification work.
The daily briefing
TLDRocket reads 60+ sources, removes duplicate coverage, and summarises the day in two minutes. Free, no spam, unsubscribe anytime.
TLDRocket
21 July 2026
AI music is becoming noise. Deezer now reports that over 50 percent of daily uploads are AI-generated—90,000 tracks per day in June, up from roughly 9,000 in January 2025. The platform is culling AI-made tracks without plays or tied to fraud, but the flood exposes a deeper problem: AI has made music creation free and frictionless, and platforms are struggling to separate signal from spam. Meanwhile, the real innovation in AI-assisted workflows is happening in code. Anthropic's Claude Tag, a Slack integration launched last week, now generates 65 percent of the company's engineering pull requests, compressing development timelines from months to weeks. That shift—from execution speed to taste and judgment—is reshaping how teams work.
Read the full briefing →