the biggest mistake people make with AI investing / trading is asking it what to buy. you could instead use it to find profitable wallets that win a high % of trades using on-chain data. you could get it to track these wallets. you could combine the data you receive to find live actionable trends in the market. you could plug this in to compare it to categories and narratives that are trending. you could then find coins that interest you and then run a separate skill that conducts institutional grade research. then you could find upcoming high quality catalysts for that coin of interest. then when bitcoin has a red day and you see your coin of interest has relative strength, good catalysts coming, profitable wallets are buying it, and it passes with flying colours via your institutional grade research skill then, and only then do you decide to buy it.