I was thinking about a hybrid group/hashtag system. Groups and hashtags have different noise/signal profiles but, theoretically, they can act as two connected water bodies than can feed back each other when there are problems in one of them.

For example, a hashtag can have a high amount of noise, a high amount of signal and a relatively low signal/noise ratio; a group can be the opposite: low noise, low signal and a relatively high signal/noise ratio, curated by a group of people that know about the subject. In this sense, hashtags and groups could be complementary depending on how much time you want to employ searching for the thing you need. Also, the utility of the hybrid system would be greater over time because you would be slowly building and expanding a web of trust