re: #5 lawhawk

IMO, companies that have corralled a lot of data, and limit their AI to their curated content, have an advantage here. Limit the generative AI to that data set, and you have potential to limit the hallucinatory effects.

This is the primary actual purpose of AI. The smaller the problem domain, the more accurate the AI is at generating its results. Within that problem domain, the more data you can have, the better. As the problem domain grows in scope, the AI’s predictive power is reduced, until you get to a point where you’d get better results flipping a coin than from asking the AI.