Two things I was not able to find a good replacement for in Linux when I was last using it a few years ago: Everything and AutoHotKey. Those are so good, and WSL perfectly adequate, that I'm now loath to switch away from Windows.

AHK_X11 is a fairly recent port of AHK for Linux that might be worth checking out; I haven't used it, but if it's half as good as the Windows version, it'll be great. If nothing else, you might like vimium-everywhere, which that'll let you run.

For Chrome, while Tridactyl might be more powerful, I've found Vimium sufficient.

I can't tell you with certainty how the weights will move during the training process, but I can tell you where it's going at a higher level.

Relevant xkcd: link.

See also: Hollywood accounting.

I think there's a single framework that captures several of these ideas: the status quo is sticky, and it's hard to steer out of, so if you think things are bad enough, you shake things up into a transient chaotic state and then as it settles down, you can guide the system into a state you like better. Different groups executing this strategy may seem aligned at the beginning, and their conflicting goals become apparent only after the status quo has been toppled. And it's possible to lose badly enough in this second phase that you end up worse off than before, so that's the gamble one takes when engaging in revolution instead of incrementalism. 

if a great disaster (say, 3 Gorges Dam just spontaneously collapses) were to befall China tomorrow, 

I agree with Adam Smith's view expressed in The Theory of Moral Sentiments, 

Let us suppose that the great empire of China, with all its myriads of inhabitants, was suddenly swallowed up by an earthquake, and let us consider how a man of humanity in Europe, who had no sort of connection with that part of the world, would be affected upon receiving intelligence of this dreadful calamity. He would, I imagine, first of all, express very strongly his sorrow for the misfortune of that unhappy people, he would make many melancholy reflections upon the precariousness of human life, and the vanity of all the labors of man, which could thus be annihilated in a moment. He would too, perhaps, if he was a man of speculation, enter into many reasonings concerning the effects which this disaster might produce upon the commerce of Europe, and the trade and business of the world in general. And when all this fine philosophy was over, when all these humane sentiments had been once fairly expressed, he would pursue his business or his pleasure, take his repose or his diversion, with the same ease and tranquility, as if no such accident had happened. The most frivolous disaster which could befall himself would occasion a more real disturbance. If he was to lose his little finger tomorrow, he would not sleep tonight; but, provided he never saw them, he will snore with the most profound security over the ruin of a hundred millions of his brethren, and the destruction of that immense multitude seems plainly an object less interesting to him, than this paltry misfortune of his own.

simplified chart of human metabolism

That seems to have gotten compressed well past the point of legibility, but it looks like the Roche Biochemical Pathways poster by Gerhard Michal: link 

what makes this not straight up murder

Perhaps the same mysterious aspect of overwhelming state power that makes taxation not theft.

"Your enemies are saying X horrible shit" is possibly the most common form of slander on Twitter

Possibly, but it's probably also simply true most of the time. Usually, you can simply quote tweet them (or post screenshots) saying the thing you're accusing them of saying. Sure, sometimes, it's missing relevant context, but that's relatively rare: normally, your enemies really are saying the horrible things. 

This argument works just as well for exponential decay: it's "always the wrong model" because no decay lasts forever, eventually you'll run out of atoms in your radioactive sample (or whatever else you're modeling), and so some other curve that intercepts the y-axis in finite time is a better model because it gets the "basic shape" right and doesn't make "an extremely obvious mistake when modeling reality."