Wow, that sounds like it could be a really great project! I was not familiar with Claude 2 (or any version of it), so I am playing catch-up. I signed up for Claude but even the subscription model or how to access Claude 2 is not clear to me. I'm also wondering why specifically that LLM? Rather than Bard or ChatGPT 4. I will have to look at your other thread on it too (just noticed some recent entries). I found this interesting comparison: . It doesn't help with my initial quesiton.
Insofar as this Fed research, it does sound like it could be very interesting. I'm not sure where to begin in order to get the data and TBH I have some other more pressing backtesting that I need to get to. So that while inclined to say I'd love to explore it, I don't want to say that and then seem flaky if I can't realistically get to it right away. However, I am always up for learning more and I really appreciate the idea. Let me learn a bit more about Claude 2 and what might be involved.