Right* (depending on jurisdiction, ymmv). Getting from it articles that you cannot have otherwise is (except for the vanishingly few articles in the public domain, which are not really likely to be found on Sci-Hub anyway. I tried).

Saying that the work could not have been done without it is basically admitting that you got papers outside the subscriptions of your institution. It is also against the policy of all the institutions I know, so could be grounds for termination. And again, this could apply to your co-authors as well.

I am really not endorsing the subscription model, but putting something like that in your articles' acknowledgments really is not smart.

Can't tell if it's intentional (some variation on Poe's law applies here), but this is a near-perfect illustration of the difference between a rivalrous (space in an intersection) and non-rivalrous (copies of a digital file) good.

I will allow that there may be some harms which can result from unlimited copying of anything you want— but they bear no resemblance to the collision of automobiles in the midst of a busy intersection.

That metaphor backfirews pretty hard IMO, is that intentional?

Doing it and claiming to have done it in a professional setting are quite different though. I am perfectly willing to believe that a lot of researchers use Sci-Hub (though from my experience they tend to use Researchgate more); I am certain that the number of people who would shout about doing it on the rooftops is nowhere near even a small minority.

If you don't have the courage, you can also do it indirectly: simply reference a paper that states that it would not be possible without Scihub. Then, your paper would not be possible without Scihub either :)

Yes, that would fit better.