AI coding in one year: from unusable to full automation with no code review
angelaqdai · x · 2026-11-12
Synthesis AI founder Matthias Niessner traces a clear capability curve across one year: AI coding was unusable 1 year ago (Copilot as vibe-coding, too many bugs), agentic tasks mostly failed 6 months ago (though Opus 4.6 was a turning point once plans were well-specified), and now agentic tasks work so well he describes the state as "nobody checks code + full automation." His key method: iterate on the plan until it's well-specified before handing tasks to the model.
More from coding & agent
- United Airlines to Ban Personal AI Agents From Booking Flights for Customers — Polymarket · 2026-11-12
- Agentic Search explained: a 4-search loop costs about 5x one search — and queries leak to more companies — adnan_hashmi · 2026-11-12
- AWS says thousands of autonomous AI agents wrote every line of AWS Emulator — SumitGup · 2026-11-12
- freeCodeCamp launches free 2-hour course on agentic coding with Codex — adnan_hashmi · 2026-11-12
- A criminally underrated resource comparing tool-calling reliability across models and providers — ricklamers · 2026-11-12
- Organizing 55k data points is easy: enterprise data prep for agents is the hard part — BenBajarin · 2026-11-12