The value spoken of here is that of slowly building up (evolving) directed knowledge over time. The community evolves links using work and coherence into actionable information/knowledge whereas AI currently don't have an idea of leadership or direction into which to take that knowledge, so they're just creating more related information which is interpreted as "adjacent noise". Choosing a path and building off of it to create direction is where the promise lies. Of course some paths may wither and die, but the community will manage that whereas the AI would indiscriminately keep building in all directions without the value of demise within the system.

9/8g The card index technique is based on the experience that one cannot think without writing – at least not in demanding, selectively accessing memory-based contexts.

This also means that one cannot think without making allowances for differences.

I like this slightly more differentiated instantiation for thinking better than Ahren's assertion that one can't think without writing. Luhmann qualifies it over and above Ahrens who elides meaning if this was the source he may have been tangentially referencing. (Was it an explicit reference? check...)

https://en.wikipedia.org/wiki/Po_(lateral_thinking)

Adler has juxtaposed the ideas of genius and commonplace together, but somehow doesn't notice the ratchet that ties them together to be one and the same tool for making them so productive?!?

Commonplacing, florilegia, anthologies, miscellanies, zettelkasten are a ratchet for thinking.

Well made notes, revisited can act as a ratchet to elevate and expand a conversation rather than devolving back to the mundane.