Some thoughts as someone who has been eating plant-based for the past year and who thinks about the ethics constantly for fun:

• I know someone on a mostly-beef diet who will probably develop health issues if they stay on a diet with plant-based food other than fruit for too long.
• Beef seems to be less bad from an animal welfare perspective than most meats, at least in farms in Australia. I would probably still pay a premium for extra-ethical beef if I stopped eating . Dairy is probably bad everywhere except India.
• Kangaroo meat seems like a clear-cut example of a "vegan meat" in practice. The Australian government sets quotas for culling them (they are overpopulated in deforested areas), very precise sharpshooters shoot them in the head, and some of the meat from culled kangaroos is sold. Demand for kangaroo meat has no effect on quotas. There is no counterfactual animal harm from buying kangaroo meat. Kangaroo suffering (and negative externalities in general) resulting from culling is very low at any rate, save for the occasional accidental killing of a mother kangaroo. I've had kangaroo meat maybe five or six times since going vegan, usually either to deal with the odd craving, to signal to family that I'm not a purist, or because someone had cooked it or a restaurant had it on a menu.
• Avoiding honey/oysters seems like it has less impact than it's worth for the time I've spent pondering the subject.
• You're right that wild-caught fish are not necessarily humanely treated. The ikejime process for killing fish is very humane compared to the cheaper and more suffocation method, but I've never seen fish advertised as ikejime (and they are probably an order of magnitude more expensive).
• Social pressures in either direction are a thing. The time I spend with different groups of friends has changed since I went vegan, even though nobody has really objected to it. My flatmate reduced their meat intake significantly almost immediately after I went vegan, but they still eat meat when going out to restaurants with others. People seem to care about food a lot. People also seem to take cues from those around them.
• If eating vegan had noticeably worsened my health or my concentration, I would have stopped, because I can probably do more net good in the world without those problems even if I had to eat the flesh of sentient creatures. Eating vegan and spending your time on other things are not orthogonal for everyone.
• If you rely on evolutionary heuristics for diet you're probably not going to be vegan, especially since we know B12 is a thing that's difficult to get without animal products, and there could be other things you aren't getting. I'm not really worried about this since I don't see many short term issues for myself, the medium-term issues seem to be more reliably mitigated against now, and if anything I've heard that it helps longevity.
• Vegans trying to gain muscle for whatever reason aren't necessarily going to have a hard time as long as they remember to eat protein at all. Vegans trying to cut (lose fat without losing too much muscle) are probably going to not have quite the same success if they were an omnivore or otherwise have a very boring time with the food they eat (unless they eat kangaroo).
• I seem to need to eat more to actually have enough energy if I'm not eating eggs/meat. It's more expensive if you don't like cooking, but the extra fibre is usually a good thing.

I'd say that not eating animal products seems like the correct choice if you can pull it off without too much trouble. It probably makes more and more sense the older you get. If you can afford to change your diet but not exclude animal products entirely, consider eating kangaroo if your country doesn't ban its import. Consider eating kangaroo anyway.

There's an analogy being drawn between the power of a hypothetical advanced alien civilization and the power of a superintelligent AI. If you agree that the hypothetical AI would be more powerful, and that an alien civilization capable of travelling to Earth would be a threat, then it follows that superintelligent AI is a threat.

I think most people here are in agreement that AI poses a huge risk, but are differ on how likely it is that we're all going to die. A 20% chance we're all going to die is very much worth trying to mitigate sensibly, and the OP says still it's worth trying to mitigate a 99.9999% chance of human extinction in a similarly level-headed manner (even if the mechanics of doing the work are slightly different at that point).

Given how long it took me to conclude whether these were Eliezer's true thoughts or a representation of his predicted thoughts in a somewhat probable future, I'm not sure whether I'd use the label "candid" to describe the post, at least without qualification.

While the post does contain a genuinely useful way of framing near-hopeless situations and a nuanced and relatively terse lesson in practical ethics, I would describe the post as an extremely next-level play in terms of its broader purpose (and leave it at that).

I think I'm more motivated by the thought that I am going to die soon, any children I might have in the future will die soon, my family, my friends, and their children are going to die soon, and any QALYs I think I'm buying are around 40% as valuable as I thought, more than undoing the income tax deduction I get for them.

It seems like wrangling my ADHD brain into looking for way to prevent catastrophe could be more worthwhile than working a high-paid job I can currently hyper-focus on (and probably more virtuous, too), unless I find that the probability of success is literally 0% despite what I think I know about Bayesian reasoning, in which case I'll probably go into art or something.