I’ve also gone down this rabbit hole of thinking. Here is the main roadblock I encounter when imagining it in practice: It might prevent users from having a co

I’ve also gone down this rabbit hole of thinking. Here is the main roadblock I encounter when imagining it in practice: It might prevent users from having a common ‘language’ for describing their interactions with the product. If everyone perceives the product from a different lens, then everyone’s experience will drastically differ and there isn’t an objective common ground for users to discuss their experience with one another. This problem seems to be rampant across the user experience with ChatGPT, wherein countless users claim that it offers profound value, and others claim that it is hallucinatory rubbage. It’s possible that none of this is an actual ‘problem’ though, and just the very nature of probabilistic software. Users and product owners just need to adapt to the mindset that their experience may vary. But this presents a whole new breed of A/B testing considerations.