Claude Opus 4.8 — One Step on Capability, One on Policy: A More Honest Model at the Same Price, Plus a 2028 Scenario
Tags
21 pages
Anthropic
Anthropic's April 23 Postmortem — Three Overlapping Regressions and What Engineers on Claude Should Take Away
Anthropic's Teaching Claude Why — Reasoning Beats Demonstration, Blackmail Drops to 0%
Anthropic Leases All of SpaceX's Colossus 1 — What the Claude Rate-Limit Bump Actually Means
Claude Design, Tested on a Real Project — PopCon UI Refinement
Claude Tuner — Stop Guessing Your Rate Limit, Start Tracking It
Drew Bent on Using AI Well: It's Not About Speed
Project Glasswing and Claude Mythos Preview — Anthropic's Bet on Proactive Cybersecurity
Measuring Claude Code's Hidden Token Tax — 11 Bugs, 30K Requests, One Flag Override
Reading the Claude Code Hidden Problem Analysis — 11 Bugs, Proxy Data, and a Quota Blind Spot
