[scythe]

If you look at the description of the cohorts you'll find that most of the people classified as long-term cannabis users were using cannabis quite a bit more frequently than that (median ~300 days per year). It just happens that the analysis here split people into boolean subgroups rather than trying to quantify a dose-response effect (as the primary outcome; dose-response effects are claimed but not emphasized).

I would be most interested in seeing a correlation with sleep disturbances. Some cannabis users claim that they smoke weed to help with sleep disturbances; critics argue to the contrary that persistent cannabis use may cause sleep disturbances. Chronic sleep disturbances are known to correlate with a variety of negative outcomes, and the use of sleep-inducing drugs is known to correlate with a variety of problems as well (including the great boogeyman, all-cause mortality).

From my own perspective, I stopped smoking weed last year, but I don't think the overall change has been very noticeable. The use of means can distort perceptions when distributions are not symmetric. Notably, my sleep habits aren't great right now, though I'm working on it.