It certainly seems very likely that what you're doing with said screen matters quite a bit, but this is also an extremely complex thing to try to study. Not only can the net result (i.e. a hypothetical positive or negative result) of a given activity change based on the individual instance of that activity, but it can also change based on the person doing that activity (whose mindset could also change depending on the instance, which could effect the net result of a given instance of a given activity).

The point is, while it might be "pretty obvious that not all screen time is the same," it would also not be that surprising if screen time was still negatively related to the average number of common screen activities over a certain amount of time (which is what you would expect from a study like this).

I'll of course hedge my statement, by saying that the Hacker News community is likely not a representative sample when it comes to common screen activities relative to the general population.