The rule for AI pipelines: never depend on a model you can't swap in a day
ccerrato147 · x · 2026-09-21
The author proposes a floor rule for production AI systems: no node in the pipeline should depend on a model you can't swap out within a day. The accompanying "floor" stack: open-weights models (3-6 months behind on benchmarks, near parity on most production jobs once priced), small self-trained, self-hosted decision models for routing/scoring/approval/escalation (milliseconds, no API), and your own outcome data and evals.
More from coding & agent
- SemiAnalysis: Persistent agents now eat 45% of our AI spend, and Perplexity's harness beats Claude Tags — dylan522p · 2026-11-12
- Ora's Ranker 3.0 grades whether agents can actually use your site; scores drop 40+ points — yenkel · 2026-11-12
- Gemini API Managed Agents now support self-hosted sandboxes, incl. Vercel — steren · 2026-11-12
- Rayline launches ARC, a per-turn model router claiming 90% cost cuts — testingcatalog · 2026-11-12
- Karpathy's llm-wiki: a pattern for building personal knowledge bases with LLM agents — stochasticchasm · 2026-11-12
- Agent self-improvement with frozen weights: failure traces add just 4pp, blind search gets 60% — burny_tech · 2026-11-12