Astra tip: shift reasoning levels mid-conversation without invalidating the cache
intellectronica · x · 2026-09-14
- Replying to Dan Shipper, @intellectronica shares a non-obvious Astra trick: you can change the reasoning level mid-conversation without invalidating the cache.
- He calls it "riding it like manual gear": start on light, jack up to medium when things get interesting, high for planning, then back to light for execution — all within the same session.
- For heavy agent users, this is a practical way to trade cost vs. quality per task phase instead of locking in one level per session.
More from coding & agent
- Phantom: deploy a Slack AI agent that learns your workflow with one docker command — JafarNajafov · 2026-09-14
- Open-source Phantom gives AI its own computer, email, and self-rewriting config — JafarNajafov · 2026-09-14
- Memory is a curation problem, not a capacity problem—agent memory thread wrap-up — bigaiguy · 2026-09-14
- Cold-start test: a fresh agent should reconstruct the task from state alone — bigaiguy · 2026-09-14
- Agent memory trick: scan memory for contradictions before acting — bigaiguy · 2026-09-14
- Multi-agent handoff trick: pass four fields, not full transcripts — bigaiguy · 2026-09-14