I mean, if you’re going to attempt to kill with a bite, I’d assume they’re going for the neck or other particularly vulnerable place. I’d argue commoner bite attacks should have very low chance to hit, but could plausibly be lethal. Most places they get a good bite won’t be able to do any damage, except maybe cause a Con save for disease… I’d say an improvised attack with Disadvantage for 1d6 damage probably works pretty well.

There is a long range tazer design that uses this principle. It uses a tiny burst laser to plasmify the air, which then becomes the conductor for the electrical payload to the target.

8 plus minutes to name the game, and not even in the description. Cool.

Pathfinder 1st Edition was a branch of DnD 3.5, it is occasionally called 3.5.5 or 3.75. It is pretty much a 3rd party patch for 3.5 but uses the same core systems. That, 3.5 and PF1e, is kind of a mess.

I’m not surprised you find PF2e confusing, but from a design standpoint I would call it clean, considering everything that is going on. It is deep, but well organized. As opposed to DnD 5e, which is relatively shallow, which can make it easier to jump into, but not as well organized. The messiest part of 5e is the “natural language” philosophy they went with, which can leave a lot of rules ambiguous. It was supposed to make it easier to intuitively pick up and play, but it also makes it much easier to have misconceptions and anything that is slightly unintuitive can easily be accidentally used the wrong way for ages. PF2e might have a lot of interconnected rules depth, but it also has a less ambiguous guide for dealing with it, which is what enfranchised players will generally mean by “clean.”

Well, in many games there is a Speak With Animals spells or equivalent. They are capable of communication, we just usually can’t understand.

What level of abstraction is enough? Training doesn’t store or reference the work at all. It derives a set of weights from it automatically. But what if you had a legion of interns manually deriving the weights and entering them in instead? Besides the impracticality of it, if I look at a picture, write down a long list of small adjustments, -2.343, -.02, +5.327, etc etc etc, and adjust the parameters of the algorithm without ever scanning it in, is that legal? If that is, does that mean the automation of that process is the illegal part?

Sentience is the little hump that we can at least sort of see some evidence of, judging by how similar regions of brains activate in certain circumstances. Sapience is the real tricky one.