Editing Essays into Solstice Speeches: Standing offer: if you have a speech to give at Solstice or other rationalist event, message me and I'll look at your script and/or video call you to critique your performance and help

It's still not been common until much more recently than that. Five years at most, which is not a new normal. It hasn't been working out that well for anyone except Trump himself and there's a decent chance it explodes on his death (which will almost certainly be before '32).

You don't seem to understand the distinction between 'mislead' and 'lie to the face.' Or, you know, you're lying to our face. Which would be a stupid plan, but maybe you're doing it on purpose.

Even politicians lying to their audience's faces is quite rare, in the USA. That it's become meaningfully popular in the last 10 years is considered extremely notable and a sign of the decay and apocalypse of the United States, and to lesser degrees other Western countries. Bill Clinton got caught lying to Congress's face about a trivial but embarrassing matter and "I did not have sexual relations with that woman" might as well be the words on his tombstone even though he suffered no real material consequences for it, because it's very rare and no one will forget it.

Lying is hard. Being misinformed, or interpreting true fair information in a biased way to reach a biased conclusion, or writing the bottom line first, or quoting true statements in such a way as to give a deliberately misleading impression, are easy. Politicians do them all the time, and this is expected. Journalists did it a little until the Web revolution and now do it more than that, and this is, again, considered very notable and a sign of the decay of the industry. Scientists and experts do it a little, and almost always the first two types, but when they do even a little of the latter types this makes people riot and consider it very notable and a sign of the decay of the field.

Oh, fantastic, I just went looking. Thanks for investigating this!

It may not be a necessary condition, but if you want to present it as obvious, it is necessary. Anything short of an exact match is only allegedly the same until you have some research results that don't currently exist to demonstrate strong reason to believe your lesser conditions are sufficient.

CAST is a great idea and seems like the most promising way forward with architectures similar to the ones we have, but I do not see any reason to believe we could, if we had a corrigibility meter, build an AI that implemented corrigibility with reasonable robustness within a year. Five years would probably be enough but at that point you're looking for at least one, and maybe 2-3, major insights.

In particular, it looks like we’re close enough to being able to implement corrigibility that the largest obstacle involves being able to observe how corrigible an AI is.

That's a wild claim to make without reference to specific papers or milestones. I'm not fully up on 'superalignment' progress but last I looked no one on the modern paradigm side was seriously attempting to study corrigibility, let alone making this kind of progress. And results like Golden Gate Claude and the 'buggy code -> evil' transformation indicated it was probably just as hard and unnatural vs in the MIRI paradigm.

In the scenario where he knows that prophecy foretells an outcome incompatible with his success here, his major decision point is long-past; he has reason to do it anyway (presumably prophetic reason). I see nothing he could do which is obviously better, and the conversation may itself be part of the keyhole future path.

If he doesn't, this is still far from an "approximately-worst option." It's still a really good trap unless Quirrell knows the Mirror is going to be the trap, knows Harry's Cloak is the genuine article that will still hide him from the Mirror, and can trick and coerce Harry into coming with him, which is three different things Dumbledore has good reason to think he probably doesn't know. The latter two are both achieved only through adventures Dumbledore doesn't know about - Azkaban, and Harry using up his time-loop password on the first day. As Lucius told Draco - any plan that relies on three things going right for you is at the limit of possible plans, and the real limit is two.

I still do not know what you think he should have done, either in the scenario where he knows he will fail due to prophecy, or the one where he does not.

I have two questions that come to mind, both about Chapter 86: 

• What else did the Dark Mark do? Did you work any of it out?

As many people have probably wondered since they read

"But as you can see, the Dark Lord was quite cunning." His gaze grew more distant. "Oh," Severus breathed, "he was very cunning indeed..."

and

• If Harry hadn't been interrupted by Mad-Eye arriving, would he have conceded enough of the point, and explained enough about how dangerous he would be fighting a war, to shock McGonagall and Snape into believing him? Would this have been catastrophic to the prophecy, probably via convincing Snape that Voldemort hadn't ever been fighting to win?

Because I reread this scene a couple weeks ago, and it seemed like there were a bunch of reasons for him to find a way to elaborate on why he thought a true Dark Rationalist fighting a war against Dumbledore would win in hours rather than years, and his internal debate on this seems to be trending solidly in that direction, but then when he's going to reach for paper, the meeting is interrupted and he never gets back to it. Which made me think that Dumbledore had arranged things for prophecy's sake to prevent him, and consider why.