Stop Writing Perfect Prompts: Let AI Watch You Work and Infer Intent
dfinke · x · 2026-08-18
Doug Finke, a 16-time Microsoft MVP, will host a live session "What If AI Watched You Work?" exploring a development loop that differs from writing perfect instructions:
- Start with a small tool that's deliberately missing capabilities, use it the way you wish it worked, and try workflows it wasn't built for.
- Then say a single word — "feedback" — and watch AI examine the evidence, infer what you were trying to achieve, and suggest the smallest useful improvements.
- Approve selected changes, rerun the workflow, and see whether the project learned something for the next iteration.
The loop: demonstrate behavior, observe the gap, infer intent, confirm direction, improve the tool, learn for next time. Deeper code continues in his Agentic AI Forge community.
More from coding & agent
- Automating a personal knowledge base with Claude: a minimal agent workflow — anirbanbandyo · 2026-08-18
- Claude Code 2.1.234 Released: GitLab Integration, Security Fixes, and CLI Enhancements — ClaudeCodeLog · 2026-08-18
- Supastarter Launches AI-Native SaaS Boilerplate with Agent Skills — jonathan_wilke · 2026-08-18
- Building an E2E testing framework for agents deploying real infra to Cloudflare — samgoodwin89 · 2026-08-18
- Open-Source Pi Plugin Rewrites AI-Speak Into Plain Human Language — wtfzambo · 2026-08-18
- ByteDance Paper: Are AI Agents Actually Controllable? — rohanpaul_ai · 2026-08-18