Warp shows 3 ways software factories improve themselves: skills, memory, model routing
AI Engineer · youtube · 2026-09-27
At the AI Engineer conference, Suraj Gupta, who leads harness development at Warp, demos three practical mechanisms from Warp's open-source software factory:
- Self-improving skills: an outer-loop agent watches the triage agent's runs and feedback, then opens a PR to update its skill — every change is Git-tracked and human-reviewed.
- Persistent memory: a versioned, traceable store of facts so e.g. a Sentry agent doesn't rediscover root causes it already found; works across harnesses including Claude Code and Codex.
- Model routing: avoid paying Opus prices for triage or simple CI fixes; use Warp's auto models or custom rules, with internal evals showing UI tasks run well on GLM.
More from coding & agent
- Gemini now connects to Airtable, Linear, Adobe and more — 10 workflows to try — alifcoder · 2026-09-27
- User denies Chrome permission, Codex pivots to in-app browser instead — andimarafioti · 2026-09-27
- Opus 5.5 plus the right skills equals a junior video editor: 4 open-source picks — lxfater · 2026-09-27
- OmO V5 ships as desktop app with auto model selector teased next — jasonkneen · 2026-09-27
- They killed 27 of 30 agents and rebuilt around 3 — maintenance forced the rethink — AmosBarJoseph · 2026-09-27
- Dev hails Claude Code's Dynamic Workflows with Opus 5.5 as best multi-agent experience yet — daniel_mac8 · 2026-09-27