Anthropic engineer: 90% of our engineers already use self-improving loops as prompting era ends
Roger_M_Taylor · x · 2026-09-18
A retweet chain surfaces a full course, "Loops and Graphs," on building Claude Code workflows.
- An Anthropic engineer claims 90% of the company's engineers already use self-improving loops for daily coding, and "prompting is basically over."
- The course builds a complete Claude Code system live from an empty terminal in 10 minutes, structured as Agents → Loops → Harness → Graphs.
- Skipping any layer makes failures resurface as "model errors," costing a week to debug.
- The pitch: stop babysitting agents step by step — loops let them iterate autonomously while graph structures manage dependencies so you only approve the final step.
More from coding & agent
- Distributed-Slides: an MCP server that compiles agent-written talks into offline presentations — arthurcolle · 2026-09-18
- Vercel skills CLI adds Notion-hosted agent skills, no Git repo required — ivanhzhao · 2026-09-18
- Vercel teams have long written GTM and on-call skills in Notion — ivanhzhao · 2026-09-18
- eve adds automatic model selection: agents pick models by task difficulty before inference — cramforce · 2026-09-18
- DHH on Lex Fridman: Nov 2024 split coding into two universes, agentic coding is transformative — zakkohane · 2026-09-18
- Vercel CLI now deploys static artifacts in under one second, build step skipped — cramforce · 2026-09-18