Most LLMs are unreliable when it comes to trading and markets. They hallucinate strategies, indicators, and even backtests.
I’m currently doing a deep dive on how to fix this using RAG + real trading knowledge 🧵↓
Still early, but the takeaway so far:
RAG > fine-tuning for financial decision systems.
I’ll share learnings & outcomes as this progresses — across portfolio management, trading, & other use cases.
Follow along if you’re working seriously at the intersection of AI & markets
Finance is not a domain where “close enough” is acceptable.
If an AI system can’t cite its assumptions, data sources, and logic paths, it shouldn’t be trusted with capital.