OR, I see a lot of hype around for perplexity but no so much development towards that. I could build an open-source version of Perplexity with much much more features and with all the hype I could easily get some traffic.
What do you say?
You seem to have sectors you'd like to solve problems for (banking, finance). How many conversations you've had with how many people in these sectors? What have you learned about problems they have and are willing to pay for (as in, they're already paying to solve them but the solutions are unsatisfactory)?
Have you narrowed these down to something you can deliver that they agreed to pay for and, maybe, even agreed to finance the development for?
From what I understood from what you wrote is that you seem to start from technology (LLM, features to an eventual open-source alternative to a technology) when, in my opinion, it ought to start from a (potential) customer's problems.