The spoken word is probably the single most difficult common specific sound to filter out - it's probably the sound we're most hardwired to pick out from background noise and process, possibly second to the sounds of an aggressive wild animal.

Audible conversation is incredibly distracting - we put up with it in the workplace because workers need to collaborate and communicate. There's a reason it's heavily frowned upon in places where people intend to hear something specific (theatres, churches) or focus deeply (libraries).

Your Cherry MX Blue switches probably aren't going to be welcome in any of those spaces either though.

Tell that to people who intentionally go to cafes to get work done?

At least multiple conversations happening at once is easy to filter out. A single conversation in relative silence is certainly difficult.