I think Erdős said of his own mathematical ability (no doubt with some humility) that it was very much a 'bag of tricks' he would apply to the problems his multitudinous collaborators were working on. I think a lot of philosophers use this approach too.

I think the value of polymaths is more 'synthetic', not necessarily 'analytic'. Many useful discoveries or insights are the result of an almost magical use of what C.S. Peirce called abduction: the process of generating hypotheses.

Knowledge in many domains, connected to the right problem, can lead to profoundly effective abduction. Sometimes this transfer can take the form of a persistent metaphor, or maybe an experience that makes itself known to you while you're thinking about something else, or simply your attention is guided to look at something in a particular light.

I feel like the post does a great job critiquing those high-status polymaths who emit an irresponsible amount of certainty. But there are plenty who use their wide learnings in a less concrete way, looking to add to their bag of tricks, and to put things together in interesting ways.