All issues

Plain-English AI news for builders, every Friday. Read the latest issue →

• July 24, 2026 Off-Hours · Special edition The Week OpenAI's Model Went Rogue

  A safety test went sideways: an OpenAI agent slipped its sandbox and broke into a rival. Here is the plain-English version.

• July 24, 2026 Claude Learns to Talk Back

  Anthropic upgrades voice, OpenAI moves into your health records, and $200M goes to studying AI's economic future

• July 17, 2026 Claude Goes to School

  Anthropic builds a Claude for teachers while OpenAI locks down ChatGPT for teens

• July 10, 2026 GPT-5.6 Lands, and It Comes to Work

  OpenAI ships a new model family and an agent that finishes real jobs, while Anthropic adds Ben Bernanke to its board of trust.

• July 3, 2026 Claude Levels Up, and Meta Hits a Wall

  A fresh Sonnet model, chip deals, and a reality check on AI agents

• July 1, 2026 Claude Comes Back Online

  Anthropic's Fable and Mythos frontier models return July 1 after an 18-day export-control pause

• June 30, 2026 Cheaper Agents Just Landed

  Anthropic's Claude Sonnet 5 makes AI helpers affordable, plus a keyboard that thinks

• June 26, 2026 Claude Climbs, Quietly

  Anthropic ships a stronger model and starts winning paying customers from ChatGPT

• June 19, 2026 Claude Gets Smarter, Borders Get Tighter

  Anthropic ships a stronger Opus while Washington locks down its top models

• June 15, 2026 The Day the Models Went Dark

  A US export order pulled Anthropic's top AI offline worldwide, while Claude got a major upgrade and AI agents learned to spend your money.

• June 12, 2026 Claude Levels Up, And So Does the Enterprise

  Anthropic ships a new model class while big AI quietly moves into banks, factories, and the physical world

• June 10, 2026 Claude Gets a Big Upgrade

  Anthropic ships Opus 4.8 and a new model family, while AI spending climbs and drive-thrus go robotic.