One thing to bear in mind is how few people ever used a Lisp Machine, so for all intents and purposes, there is no “there” to get back to when it comes to a widespread audience.
Symbolics in particular was a tour de force though. If you take all the nice things you like about programming in a Lisp today, Symbolics almost certainly had an equivalent, and it was all coherently integrated and amp’d up to 11. e.g. if you like Datomic, Symbolics had Statice, an OODB fully integrated with CLOS with a Datalog-based query language.
The compromises today are that all this stuff is available multi-language, multi-platform, mostly open source, more powerful and cheaper hardware/cloud. But not quite as coherent, not quite as capable. A sustained source of funding would be needed to overcome those compromises, but it’s not clear to me what the incentives would be other than a nerdy altruism.