The real rub is throwing the baby out with the bathwater. How can the signal be isolated/retained/copied easily so the noise can be purged? This is especially difficult for very old data that is spread across innumerable channels whose value could be lost and the team left unawares.

I think you need a culture of summarising this type of info & saving somewhere that isn't ephemeral (e.g. README, team wiki). Else even with the history maintained, it's still a pain to try to a) play pattern matching with Slack's search engine and b) weed out all the false positives of old info that's no longer relevant.

This feels like the type of thing that maybe could be farmed out to a trained LLM in the future — "summarise today's discussion on X in channel Y and post into the wiki".