Replicating "pre-AI-winter" success is like trying to replicate "pre-dot-com" success, et cetera. The past is the past. Such things all depend not on technical reality but on what was the predominant belief among some non-technical decision makers. Old Lisp was funded by institutions. There were so many dialects because multiple institutions were actually developing Lisp using local talent. That required funds, and dispensation of funds requires that someone is rationalizing that dispensation based on some beliefs, and those beliefs are very distant from specific technical ideas like "multiple dispatch is a beneficial in OOP".

What remains is that the Lisp design contains numerous elements which are great ideas, which, if they are nicely implemented in their proper form, give you a great programming language.

That doesn't translate to being able to return to the time when you're showered with institutional money to just go wild hacking on whatever you like.