Okay so the idea is similar to ICO(initial coin offering) in which different AI models will be tokenized and funds will be raised.<p>The idea is to create a platform where engineers can raise funds on the models created by them.<p>please review & suggest!
Is this like: I buy $500 for a token that entitles me to 5,000,000 API calls?<p>In the future the retail price of an API call could be ten times that, so even if I don't want to do the API calls I could sell it to someone who would for $5000.<p>One trouble is I see a lot of variance in terms of how an individual A.I. model will turn out, particularly as judged by a user. Buying into one of these looks like a risky business so if I was going to buy into one I might want to buy into ten.<p>Whoever is bankrolling the model should have some input into the test data if not the training data.
What is worth investing in are not the models, but the people behind them who can continue to drive improvement.<p>A model gets surpassed by another pretty quickly now.<p>Similar to a car engine, which is a consumable item, it's not worth investing in.
You’re going to have the same problem that ICO’s and existing AI models have: there’s no moat. It’s technically trivial to copy successful AI models/web3 projects.