Thariq Shihipar

Reports say Thariq Shihipar, one of the most visible Claude Code engineers, has left Anthropic after more than a year. Here is what is claimed, what we could and could not verify, and what changes for people who use his guides.

Stacking several $20 accounts looks like a cheap way to get more AI coding usage. The math says a $200 plan buys twice as much per dollar at Claude, Cursor and Google, the terms say stacking is risky, and the real leak is almost always stale context. Here are the numbers, a re-read calculator, and a habit list that works on every vendor.

On October 2, 2026, Andrej Karpathy posted a short ladder for understanding what language models write: controlled-language text, then diagrams, then HTML pages, then bespoke explainer videos. Here is what each rung means, the exact prompts to try, and where the advice breaks down.

On October 1, 2026 ClaudeDevs said claude.dev is the new home for people building with Claude. The site already held months of team writing. This is a map of what belongs there versus docs.claude.com, plus the terminal easter eggs that are not Claude Code commands.

"The right way to use model capabilities is not to ship 10x more features to prod" — Anthropic's Thariq Shihipar made a direct case against feature-velocity as the point of better models, arguing the actual leverage is in spending saved time on understanding users and prototyping, not output volume. The replies split between agreement and a pointed question: does this survive contact with a business that just wants more shipped?