Most popular posts from the last 7 days
10 trending posts
Claude Fable 5.2 costs twice as much per token as Opus 5.5. On 52 real tasks from a production monorepo, my tests scored it level or behind. Then I read the failures. Fable had fixed a bug the original human fix shipped with, and my tests marked it wrong. Plus: Anthropic's own system card shows the safety fallback hands attacked sessions to a weaker model.
Read more →The official Claude Code plugin that lets agents work autonomously for hours. When to use it, when not to, and the philosophy behind letting AI fail repeatedly until it succeeds.
Read more →In The Founder, Ray Kroc lives by a famous line: nothing can take the place of persistence. I co-founded a startup studio that built more than 150 startups, and watched it collapse under its own weight. The research says persistence alone barely moves your odds. My experience says it is the only thing that mattered. Both are true, because persistence buys attempts, and agents just made attempts cheap.
Read more →Boris Cherny shared his workflow for the tool he built. The setup is surprisingly vanilla. The philosophy is worth studying.
Read more →Anthropic published how the Claude Code team reviews code at agent scale: verification first, parallel bug hunters, verifier votes, risk tiers. It is a good process. Every reviewer in it is Claude. I checked my own logs: 51 times a second model reviewed a change Claude had already passed, and 29 times it found a serious bug.
Read more →HumanLayer's 12-factor agents codifies what works in production AI: own your context, keep agents small, stay out of the dumb zone.
Read more →Claude Code can now watch your PRs in the cloud, fix CI failures, and address reviewer comments while you're away. It's the logical next step after auto mode - and it raises the same trust questions, harder.
Read more →Rob Walling's MicroConf talk lists 10 ways AI attacks SaaS and 8 moats that still work. I sat on the other side: we looked at buying, then built a prototype in a weekend and merged 83 PRs in a week.
Read more →On October 29 OpenAI cancelled its cut to the $200 ChatGPT Pro plan and announced it was doubling every Pro tier. About 24 hours later Anthropic renamed Max 20x to Ultra and published its weekly limit: 10x Pro. On the label, OpenAI is now twice as generous. In API dollars it may be the other way round. Both labs are building the same plan, and neither will tell you what a unit is.
Read more →Opus 5 writes 510 words where Opus 4.5 wrote 158, with 2.3 times the em dashes and twice the 'load-bearing'. Arena measured it, Hacker News named it, Reddit downgraded over it. Anthropic's answer arrived in Fable 5.1's prompting docs: a paragraph defining 'mannered prose', with the model's own tics as the examples, for you to paste into your prompt. The vendor wrote the style guide against its own model, and shipped it as your job.
Read more →