Killed the vector DB for SQLite BM25 (0.35ms) and wrote a 3.6ns Rust sentinel for agent loops
OpenComplaint524 · reddit · 2026-10-26
After harsh community feedback on their open-source multi-agent hedge fund experiment, the team rebuilt three failure modes in autonomous agent pipelines:
- Raw metrics guarantee hallucinated reasoning. Feeding models PER=85 or volatility=65% led to positive justifications of risk signals. A Factor Ontology Schema pre-evaluates 11 quant factors with polarity contracts (HIGHERISBULLISH / LOWERISBULLISH / EXTREMEISRISK) before agents see data — polarity hallucinations dropped to 0%.
- Vector DBs are overkill for reflection memory. Embedding API + cosine similarity: 200-450ms round-trip plus API cost; SQLite WAL + FTS5 BM25: 0.35ms, zero dependencies. Keyword relevance also matched post-mortem lessons ("semiconductor pre-earnings", "volume spike") far more reliably than fuzzy embeddings.
- Python can't be your safety wire. A pure Rust sentinel with Struct-of-Arrays layout and forbid(unsafecode) evaluates a single stop-loss in 3.63ns, 1,000 positions in 3.62µs, with 2.88µs Python-Rust pipe IPC.
Takeaway: LLMs should only handle contextual synthesis and debate; everything deterministic belongs in schemas, fast local databases, and Rust.
More from coding & agent
- NVIDIA open-sources SkillSpector, a security scanner for AI agent skills — NVIDIA · 2026-10-26
- Huashu Design 3.0 turns one agent prompt into deliverable design, 26k stars — alchaincyf · 2026-10-26
- OpenBot is a local-first desktop workspace for persistent AI teammates — nightly-labs · 2026-10-26
- Open-source agentic social scheduler Postiz hits 38k GitHub stars — gitroomhq · 2026-10-26
- How Open-Source AI Agents Cut Costs While Boosting Control and Flexibility — ankur_maker · 2026-10-26
- Claude + HyperFrames can clone any short-video style in minutes — petergyang · 2026-10-26