gwern comments on Harry Potter and the Methods of Rationality discussion thread, part 16, chapter 85 - Less Wrong Discussion
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (1106)
The introspective morality-dump chapters are not my favorites (eg. I find the 'imagine distant descendants' to be entirely useless intuitively, and would prefer versions of the update-now argument which are more like 'decide now how you would update your beliefs based on predictions you make now failing or succeeding, since once they actually fail or succeed you'll be embarrassed & biased'), but oh well let's begin analysis.
This is actually a pretty bad example. The Australians and New Guineans etc. were not necessarily incompetent (witness the boomerang, or the independent invention of the blow-gun), and specifically, throwing-sticks (atlatl) are really fearsome weapons which can throw darts or rocks insane distances more comparable to English longbows than anything else. Throwing sticks for spears were in military use in ancient Greece or Egypt, areas which always had bows-and-arrows.
A better example would be Tasmania and technology it lost, like making fire.
This would seem to indicate Harry over-estimated the magnitude of his inference in the early chapters about the implication of so few Muggleborns at Hogwarts, but immediately raises the question of what do those lands do with their Muggleborns.
Finally:
Sybil is now definitely the bearer of at least one unvoiced prophecy, and if I'm counting right, at least two - she woke up without speaking in some earlier chapter as well.
My theory would be they just have some weird powers and never really find out what it means to be a wizard. Various Mediums are probably unknowing Muggleborns.
Burn them.
If you knew that a woman in your village was communing via socially unapproved rituals with a transhuman intelligence of unknown nature and preferences, would you convince your village to burn her to death? Ideally you'd just use the Object Class: Roko Containment Protocol, but then her own soul remains at risk—burning her alive at least gives her strong incentive to repent.
But where would we get that many D-class personnel?
For the record, I don't think that was a good idea under any of the plausible scenarios.
Edit: do the people upvoting this have any clue what I'm referring to?
I know I don't. What are you referring to?
As far as I can tell from my limited research, it appears to be a combination of the SCP Foundation's "Object Classes" with a hypothetical new object class "Roko" which I believe to be named for an LW user who appears to no longer exist, but made a post at some point (the best I can establish is that it had to be prior to December 2010), presenting some idea which later came to be called a "basilisk", because the very knowledge of it was judged by some to be potentially harmful and unsettling. The post was deleted, although it appears to be possible to find copies of it or at least the basic idea if one cares enough.
So presumably the containment protocol for Object Class: Roko is simply to destroy the offending information and maybe take steps to prevent a recurrence? I'm mostly guessing, anyone who actually knows this context firsthand want to comment on whether my guess is close?
This seems likely. The user in question was a top contributor and made a lot of creative and speculative posts along similar lines.
More precisely he made a post presenting a different clever game-theoretic solution which could, among other things, be used to counter the thing that became called a "basilisk".
It was a question of expectation, not preference.
Right, but I'm curious about your preference.
While it is clear that the Tasmanian aborigines did lose a lot of technological know-how, there's some dispute over whether they actually lost fire. Unfortunately, I don't have a great source for this. The claim is sourced in the relevant Wikipedia article, but the citation is to a dead-link.
As I pointed out to the person who brought that up in the discussion I linked, the dispute is pretty desperate pleading.
Hmm, reading your argument there I'm convinced. The tertiary nature of the sources claiming they had fire-making, combined with the well-documented preservation of fires are both pretty strong arguments.
This is the same instance, it's word for word the same as her previous nightmare, this chapter just continues it a little farther and shows that there are people all over the globe who are also having visions of bad things to come.
Unless my memory has totally failed me.
That's weird. So is the first instance supposed to be a massive flashforward (despite no one ever noticing this before because it was written as present tense), or is this second instance supposed to be a big flashback (despite being written as though it happens after Harry finishes his soliloquy)? Maybe Eliezer deliberately or accidentally just made it very similar.
No, this one is 11pm, the previous one was 2am.
Um, the accepted Outcome Pump explanation of prophecies says that only the right listener will discharge the time "pressure". (Possibly relevant.) The same prophecy could fail to erupt many times.
OT: In Ch. 25, Hold Off On Proposing Solutions, Harry considers only intelligent and evolutionary causes of optimization. I have no clue if an Outcome Pump could coherently explain all magic.
As the CS saying goes, things happen 0, 1, or indefinitely many times. Why does this Sybil failure happen only twice (as opposed to every night, the prophecy not having gone off on a vacation or anything), and why is it linked with additional characters who were not also linked to the previous incident?
I don't think we should assume that. The end of 85 reads to me like a flailing optimization process that can't 'find' a natural route to changing Harry's future and is pushing absurdly improbable routes.
I think it's safe to assume it's deliberate, although I do not think it is the same instance.
I've heard (I forget which of two sources it was so I can't cite) that per anthropological theory, the Tasmanians had taken not a retrograde, but an alternative, approach - that there are two branches humans have taken in regard of technology.
One is to have a maximal technology base, growing as new ideas are learned and maintained down the generations by apprenticeship and later by writing. Even at the flints-and-shells stage this requires specialism to get things done expertly.
The other is to have a minimal technology base, one kind of pot, one kind of weapon, windbreaks instead of fire, and all made out of things that can be expediently rustled up from common materials when needed and casually discarded when not, and which can be taught without effort and without specialism. It means that the species can be scattered down to the least grouping, and lose nothing. It means the individual is complete, alone and naked. They can drop everything and recreate it afresh at need.
The Tasmanians (and to a lesser extent, the aboriginal Australians) took that path. It wasn't some sort of massive technology fail. It was a different way to be successful.
That theory is possibly the most elaborate sour grapes I've ever seen.
I don't follow, care to explain?
FWIW the expedient technology route is the one taken by all other species that have any technology at all. A chimp drops his ant poking stick when he's done poking the ants. It's clearly capable of being an evolutionary success.
The Parlevar were wiped out entirely. Both species of chimp have an ICUN Red List status of Endangered. I would suggest that being wiped out or nearly so by competitive pressure brought to bear by close genetic relatives who took up a different strategy is not a marker of a strategy being an "evolutionary success".
Inability to cope with technology maximizing societies is kind of a special case. It applies to basically ALL animals, birds, fish, plants, and even to other humans who decided on being expedient technologists. If you can't call the Parlevar successful ("Before British colonisation in 1803, there were an estimated 3,000–15,000 Parlevar" -- Wikipedia) then you can't call any of the species successful that we wiped out or massively reduced.
That sounds kinda awesome in a "specialization is for insects" way, but at the end of the eon you're still dying of appendicitis.
Nick Szabo discusses similar ideas here with regard to Polynesians.