Have long hoped someone would do this thoroughly, thank you.

Much cheaper, though still hokey, ideas that you should have already thought of at some point:

• A "formalization office" that checks and formalizes results by alignment researchers. It should not take months for a John Wentworth result to get formalized by someone else.
• Mathopedia.
• Alignment-specific outreach at campuses/conventions with top cybersecurity people.

Maybe! I'm most interested in math because of its utility for AI alignment and because math (especially advanced math) is notoriously considered "hard" or "impenetrable" by many people (even people who otherwise consider themselves smart/competent). Part of that is probably lack of good math-intuitions (grokking-by-playing-with-concept, maths-is-about-abstract-objects, law-thinking, etc.).

Yeah, we'd hope there's a good bit of existing pedagogy that applies to this. Not much stood out to me, but maybe I haven't looked hard enough at the field.

We ought to have a new word, besides "steelmanning", for "I think this idea is bad, but it made me think of another, much stronger idea that sounds similar, and I want to look at that idea now and ignore the first idea and probably whoever was advocating it".

Good points, thanks! (Mainly the list part)

Thank you! Another person pointed this out on LW.

This post/cause seems sorely underrated; e.g. what org exists can someone donate to, for mass case detection? It has such a high potential lives-saved-per-$1,000!

OK, thanks! Also, after more consideration and object-level thinking about the questions, I will probably write a good bit of prose anyway.

How would you respond to essays that are substantially or mostly in the form of bullet-points, lists, tables, and other information organization methods besides prose? (Prior discussion here, here, and here, to get a sense of why I'm interested in doing this.)