Designer uses Claude with subagents on screencap+audio to make the whole world drivable
tomjohndesign · x · 2026-10-31
Designer tomjohndesign is attempting an ambitious project: making the whole world drivable from satellite imagery, road data and other sources, and admits it's harder than it sounds.
He shares a 20-minute clip of his workflow: instead of feeding individual screenshots to Claude, he annotates via screencap narration. Claude extracts the audio from the recording, then spawns subagents to investigate every issue he mentions in the transcript — much faster than one-off screenshots. He expects the build to be ready around tomorrow.
More from coding & agent
- Stanford launches CS329Z: Engineering AI Agents, with full materials public — stanfordnlp · 2026-10-31
- Cloudflare Durable Objects now defer code updates up to 300s by default — irvinebroque · 2026-10-31
- GitHub Copilot app is building richer dashboard widgets for personalized data views — mariorod1 · 2026-10-31
- Todoist's self-healing agent auto-fixes 1 in 5 Sentry issues and opens PRs — zeeg · 2026-10-31
- Dev's Multica 1.0 build log: 432M tokens, 89 agent runs, just $183 — jiayuan_jy · 2026-10-31
- Selective Context Memory skill helps AI coding agents retain key decisions across sessions — gringobrsa · 2026-10-31