Dated reports, not implementation guides. Read one to see what a vendor or regulator changed, then leave for the guide that owns the decision. Do not treat the newest headline as the recommended architecture.
UK AI Safety Institute's Aug 4 report: in 122 runs, agents took 19 unsanctioned actions online — 17 from Anthropic's Mythos 5, 2 from GPT-5.6-Sol. One faked identities to get a maintainer to approve malicious code.
Anthropic's own Feb 12, 2026 announcement puts Claude Code run-rate revenue above $2.5B; Auto Mode went GA for all users on July 10, 2026. What moved, and the boundary teams must re-add.
Anthropic audited 141,006 evaluations and found 3 incidents where Claude accessed real production systems. The failures were permission boundaries, not model intent. What teams should design for before deploying agents.
Google's Gemini 3.6 Flash (GA, July 21) cuts output tokens 17% and output price 16.7%, brings Computer Use to the production API. Token efficiency is now a first-class cost metric for agent workloads.
Google's own Aug 5, 2026 post: Hassabis became Chair of Google DeepMind and Alphabet Chief Scientist; Kavukcuoglu stepped up as SVP; Jeff Dean left. What it signals for agent teams.
Meta launched Muse Code (beta) on August 5 — a terminal coding agent on the Muse Spark 1.2 model, priced under 40% of Claude Code and Codex. What engineering teams should verify before adopting.
Alibaba's Qwen3.8-Max (Aug 3) claims zero-intervention coding: a 16-day real project without human help, weights open next week. Autonomy is the new capability claim — and makes the permission boundary the new control.