["1-Click GitHub Token Stealing via a VSCode Bug", "Agentic Mfw", "MAI-Code-1-Flash", "AI outperforms law professors in Stanford Law study", "How we index images for RAG", "Trump signs downsized AI order after weeks of reversals", "U of T researchers demonstrate AI worm could target any online device", "Show HN: Paseo \u2013 Beautiful open-source coding agent interface", "MegaTrain: Full Precision Training of 100B+ Parameter LLMs on a Single GPU", "I've been waiting over a month for Anthropic to respond to my billing issue"] ["1-Click GitHub Token Stealing via a VSCode Bug", "Agentic Mfw", "MAI-Code-1-Flash", "AI outperforms law professors in Stanford Law study", "How we index images for RAG", "Trump signs downsized AI order after weeks of reversals", "U of T researchers demonstrate AI worm could target any online device", "Show HN: Paseo \u2013 Beautiful open-source coding agent interface", "MegaTrain: Full Precision Training of 100B+ Parameter LLMs on a Single GPU", "I've been waiting over a month for Anthropic to respond to my billing issue"]
Hot Takes

Anthropic Quietly Gutted Caching. OpenAI Pulled Study Mode. Nobody Told You.

On March 6 Anthropic silently cut prompt cache TTL from 60 minutes to 5. On April 11 OpenAI yanked Study Mode from ChatGPT Plus. Zero changelog entries. Real developer bills up 17%. Welcome to the silent-nerf era of AI subscriptions.

Apr 13, 2026 6 min read