Quantocracy Daily Wrap: State-Space Models, AI Research Agents, and Quant Edge vs Noise

Quantocracy's July 16, 2026 wrap features four articles on quantitative finance, starting with CryptoMamba, a USC-built selective state-space model that predicts Bitcoin's next-day close from 14 days of OHLCV data. A House of Quants podcast episode discusses how researchers distinguish genuine market edge from noise, covering survivorship bias, timestamp errors, and false strategy creation.

Why it matters: Researchers from Google, Boston College, and Columbia propose an autonomous LLM framework that conducts financial hypothesis discovery rather than merely summarizing existing research.