Jefferies 的 Laurence Alexander 团队主张，Aristotle 的 Poetics 可为 AI 辅助量化投资提供叙事框架：LLM 能在吸收财报、数千条新闻与社群资料、价格数据及行为金融文献后，近乎即时地量化投资者信念，并将情绪标记为「euphoria」或「capitulation」等周期阶段。

核心数据在于叙事周期压缩：过去 Genesis 到 Adoption 通常需 6 至 18 个月，AI 驱动系统可缩短至数周；Euphoria 阶段更短但更剧烈，First Cracks 则因模型快速调仓而更难被察觉。Jefferies 还举例要求模型揭示类型、准确度与关键因子，如 f(ISM, PMI, OECD LEI): RF、88% R-2、80% OOB。

主要风险是 thesis lock-in：一旦投资论点被写入提示词或流程，AI 会偏向寻找支持证据、延长判准、重新诠释负面资料，并忽略竞争叙事。Jefferies 认为 AI 适合在既有框架内确认与追踪趋势，而人类分析师更擅长判断框架何时必须改变；若未在 Phase 2 Validation 前进场，迟到成本与成为退出流动性的风险正在升高。

Jefferies’ team led by Laurence Alexander argues that Aristotle’s Poetics can offer a narrative framework for AI-enabled quantitative investment: after ingesting earnings reports, thousands of news and social feeds, price data, and behavioral-finance literature, LLMs can quantify investor conviction almost instantly and label sentiment as cycle stages such as “euphoria” or “capitulation.”

The central numerical point is narrative-cycle compression: Genesis to Adoption used to take 6 to 18 months, while AI-driven systems can shorten it to weeks; Euphoria becomes shorter but more violent, and First Cracks become harder to detect because models reposition quickly. Jefferies also cites model disclosure such as f(ISM, PMI, OECD LEI): RF, 88% R-2, 80% OOB.

The main risk is thesis lock-in: once an investment thesis is embedded in prompts or workflows, AI tends to seek confirming evidence, move goalposts, reinterpret negative data, and miss competing narratives. Jefferies concludes that AI is useful for confirming and tracking trends within an existing frame, while human analysts are better at recognizing when the frame itself must change; if investors are not in by Phase 2 Validation, the cost of being late and the risk of providing exit liquidity are rising.