This seems like a really interesting usecase of LLMs. I’m curious how you’ve validated the outputs and what gives you confidence that the forecasts are good and not affected by hallucinations or incomplete information?
The most interesting aspect of this is backtesting. Quant models get run on past data to see if their predictions work.
When you use LLMs agents, though, you run into their memorized knowledge of the world. And then there's the fact that they do their research on the open internet. It makes backtesting hard - but not impossible.
"International expansion that builds on 125% international growth since the beginning of the year"
"Investment in the core Pinterest service, especially mobile which has grown 50% since the beginning of the year to become more than three-fourths of all usage"
Aren't you always extending Facebook's core features when building with their API? The point of leveraging Facebook platform is to build amazing products that Facebook doesn't have the time or resources to create (see zuck's statement on the importance of Facebook platform during the earnings call).
Thanks for the feedback. I know I probably sound redundant at this point, but we are still very early in the development this product. These screenshots are meant to show a working prototype built on our API and also to spark your imagination!
@comex for example expressed his likes about the UI, do not take my opinions as simple critics about the style. But looks like an UI that every UI designer could come up to. Well, it's pretty and curated. Thinking about it you have to account for customers/users with strong expectation from such a service, it's not that stupid to go for the well-trodden path.
I hope you're not going to be flooded by a too common design...
