Comment by jcranmer

Disclaimer: this mostly happened before, or at best shortly after, I was born, so this isn't drawn from personal recollection but rather attempting to synthesize from others' recollections, often from people who have some bias.

One of the major trends in computing in the 80's and 90's is that high-end systems lost out to the growth in capabilities of low-end systems, and this happens in pretty much every level in the computing stack. Several people responded to this trend by writing articles sniffling that their high-end systems lost to mass market garbage, often by focusing on the garbage of the mass market garbage and conveniently avoiding analysis as to why the high-end systems failed to be competitive in the mass market. The wonders of Lisp is one of the major topics of this genre.

Most famously, Lisp was tarred by its association with AI during the concomitant collapse of AI that led to the AI Winter, though it's less often explored why AI failed. In short, it didn't work. But more than just AI at the time, people also felt that the future of programming in general was based around the concept of something like rules-based systems: you have a set of rules that correspond to all of the necessary business logic, and a framework of program logic that's making those rules actually take effect--you can see how a language like Lisp works very well in such a world. But programming doesn't have a clean separation between business logic and program logic in practice, and attempts to make that separation cleaner have largely failed.

So Lisp has a strong competitive advantage in a feature that hasn't proven to actually be compelling (separating business from program logic). Outside of that feature, most of its other features are rather less unique and have seeped into most mainstream programming languages. Functional paradigms, REPLs, smart debuggers, garbage collection--these are all pretty widespread nowadays. Where Lisp had good ideas, they've been extensively borrowed. Where those ideas haven't pulled their weight... they've languished, and most of the people wistfully wishing for a return to Lisp haven't acknowledged that the limitations of these features.