Reports a federal ruling upholding the Pentagon’s defense-supply-chain blacklist of Claude over weapons and surveillance safeguards.
Anthropic launches a Claude plugin submission portal with review and usage tracking as MCP-based plugins scale.
Claude Code now gets a fixed weekly-limit allowance to reach a graceful stopping point rather than halt mid-edit.
Shares a study finding structural features distinguish unseen AI-written blog posts from human ones with 98% accuracy.
Anthropic cuts Opus 5.5 token pricing 20% and cache-read pricing 60%, alongside a Claude Code task-cost calculator.
BridgeBench says it has day-one Opus 5.5 results and will retest amid claims Anthropic silently degraded the model.
Reports T3 Code usage flipping in two weeks from Codex leading Claude to Claude holding a 2× lead.
Text Arena analysis finds Opus 5.5 improved on 10 of 12 writing measures versus Opus 5 in high-reasoning outputs.
Documents Opus 5.5 advancing a TypeScript-to-Rust port in 10 hours after four months of manual work stalled at 35% tests.
DHH reports a one-shot Opus 5.5 Rust-to-x86-64 assembly port of ttfx that runs up to 17× faster.