What I'm reading

Worth a builder's time.

A few AI and tech stories a day from the publishers themselves, each with one line on why it's worth reading. Links go to the original.

Wired · 10/22Hugging Face’s New AI Biology Model Lets You Chat With DNA ↗A new open-source model brings conversational tools into genetics and biology research.Simon Willison · 10/22Today's PHP Security Release, and How We're Strengthening PHP Security ↗Crucial update for anyone running web servers after severe flaws were found in PHP.Simon Willison · 10/22Panache — RTS Belle Époque pour mobile ↗A fun example of how quickly open-source game code can be remixed into something fresh.MIT Technology Review · 10/22AI’s hidden storage bill: What inference leaves behind ↗A practical look at how running modern models creates unexpected and costly storage demands.Simon Willison · 10/22Donating Agent Skills to the Agentic AI Foundation ↗An open standard for software agents moves to an independent foundation to keep tools cross-compatible.OpenAI · 10/22Build more responsive apps with the Decisions API ↗A new API designed to give developers quick, low-cost model choices for snappy app interactions.TechCrunch · 10/221Password launches a way for AI agents to log in on your behalf without seeing your password ↗A clever security approach that lets software assistants log in to sites without exposing actual passwords.TechCrunch · 10/22Left alone with nothing to do, AI agents started playing ↗A fascinating look at unexpected behaviors that emerge when autonomous bots interact without human intervention.