I can't speak for the previous commenter but it does seem there is a different character to the research work happening at FAANG-scale companies (which is still generally in service of their corporate strategy) vs the work that happened at Bell Labs back in the day (which was more abstract and less directly-commercializable). Bell Labs was more of a large team of brilliant researchers with ample resources to support their work, and it just had to be useful to somebody rather than be part of a product/service that could be sold directly.

Brian Kernighan's book, "Unix: A History and a Memoir" paints a compelling picture of just how open-ended research at Bell Labs can be.

FAANG do a lot of research, but the budgeting for those efforts is much more tightly controlled and connected to corporate strategy.