= 27d567bc2d48d9f40e9ccc20ea370b15)
cousin_it comments on On the Power of Intelligence and Rationality - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (187)
This is really, really false. People use armed force all the time with their neighbors. Just like at the national level, armed force is very effective as long as no one bigger and badder will smack you down for using it, as they almost always will if you're a corporation operating on a large scale in an industrialized nation. The situation is virtually identical.
On a separate point, while the Nazis had some crazy beliefs, they still excelled in a number of important areas. A superintelligence trying to figure out how to take over a country would probably create a public enemy and invoke as much Nationalism as possible, and publish lots of propaganda, and indoctrinate the youth, and establish a cult of the leader, and so forth. Acting rationally does not mean winning your opponent over with well-reasoned blog posts, unless those are the most effective means you have of getting to him. Acting rationally means using the means that will accomplish your desired ends most effectively.
If the Nazis had had their crazy values but actually been rational and carried out their war differently (less aggression towards Russia, not dragging the US in, delaying the Holocaust until they could actually afford the military resources for it), they could have been a whole lot more successful than they actually were. So let's be glad they underrated rationality.
So intelligence == winning by definition? What a way to cut the argument short!
Saying "yelling at large crowds" is irrational, without giving any definition of "rational," is quite problematic. The author can't mean "rational argument", because the power of the West is not built on Modus Tolens; it's built on guns, smallpox, and better industry. Thus, he does not seem to be using a consistent definition of rational; furthermore, "You won't conquer the world by reasoning with people" is hardly an insight. Additionally, he makes no claim that the Nazis succeeded because of their irrationality; indeed, if they hadn't been quite so adamant about killing all the Jews and taking back the Caucuses, they likely would have been a whole lot more successful.
In short, he has no clear definition of rationality, nor an explanation of how having less of it was good, or having more of it was bad. In the absence of a consistent definition or an explicit claim about methods and results, it seems entirely appropriate for me to use the conventional (around here) definition of rationality as systematically making good decisions that advance your goals.
Now, if his whole argument were, "You can accomplish a lot without being consistently rational," I wouldn't dispute that. But his point is, "You can accomplish a lot without being consistently rational, therefore, rationality isn't all it's cracked up to be." Given that the Nazis probably could have accomplished more if they'd had slightly fewer crazy ideas, he does not appear to support his own claim.
But there is an interesting question, which is how many fewer crazy ideas would be required. I mean, arguably if the Nazis were completely reasonable, it's unlikely that they'd have been all that successful either. They had to 1) appeal to a certain contingent, while 2) scaring the crap out of everybody else. Arguably, having a certain amount of sheer batshit crazy ideas helped with both... whereas being merely mildly eccentric and power-mongering probably wouldn't have helped.
IOW, it's likely difficult to separate the Nazis' results from their irrationality... even if that just means they were being rationally irrational, so to speak. (Of course, making a commitment to being scary-crazy as a deterrent/threat strategy means you're likely to still be quite crazy when you actually have some power.)
"Saying "yelling at large crowds" is irrational, without giving any definition of "rational," is quite problematic. "
Yelling at large crowds is neither rational nor irrational, as it is not a belief and rationality is about beliefs. What I was saying was, effectively yelling at a large crowd doesn't depend on whether your beliefs are accurate, any more than the ability to roll your tongue does.
"The author can't mean "rational argument", because the power of the West is not built on Modus Tolens; it's built on guns, smallpox, and better industry. "
The power of the West isn't built on rational argument with other people, it's built (in large part) on rational argument internally, Westerners arguing more rationally with other Westerners. This caused the West to become better at science and business, which caused it to become more effective at creating wealth, which in turn generated guns and industry.
"In short, he has no clear definition of rationality, nor an explanation of how having less of it was good, or having more of it was bad."
That, as you said yourself, was never my thesis; my thesis was that, for attaining many goals on the individual level, rationality is not the most important factor.
Their "crazy" ideas included most of their top-level goals. Whatever they achieved or didn't achieve has to be measured in terms of their goals, so you can't just propose to throw away some goals to improve the chances of achieving others.
Take the example of killing Jews. This was a top-level Nazi goal. Yes, if they had achieved complete and lasting world domination, or even domination of Eurasia, they would have been able to kill Jews later. But given a significant chance of military failure, and an even more significant chance of partial failure - holding on to Western Europe, losing the rest of the conquered territories - diverting resources to killing Jews was a rational tradeoff.
Which is part of the reason for calling them irrational. Killing the jews was originally a means to the end of preventing jews from breeding wtih (other) germans, which itself was a means to the end of making Germany stronger. When the killing became a top level goal that was what Eliezer would call subgoal stomp. The means became a supergoal, and their pursuit of that goal was part of what defeated the goal that it originally should help achieve.
From the Nazis' POV, racial contamination was only one way in which the Jews maliciously influenced society. They also directly intervened in politics, the economy, culture, war efforts (like when they caused the German surrender in WW1), etc. The jews and their allies controlled states like the USA and used them to interfere with Germany, and they could not be allowed to threaten Germany directly from the inside.
The jews could not be reformed, their malicious tendencies could not be eradicated, and therefore they had to be killed. Tthat was a legitimate supergoal.
Which was just more reasons for having it as a subgoal to making themselves more powerful. If a subgoal defeats its supergoal you are irrational, no matter how important the subgoal is.
Besides all of that could have been achieved by taking away power from all jews and putting them in ghettos. Which they had already done before even starting on the "final solution".
By your argument, they should never have made any effort that did not lead directly to increasing their power.
But their rule wasn't (just) a goal in itself, but a tool used to achieve other goals. Given that they had a chance of failing to achieve lasting world domination, it made sense to pursue goals like killing Jews which would leave a lasting legacy even if they were ultimately defeated. In other words, the destruction of the Nazi regime is not the absolute possible min-utility.
If the ghetto solution had worked, they would not have needed the Final Solution, and would not have worked on it. The F.S. wasn't just for the fun of it - it was created to handle the problem of large jewish populations in newly-conquered Eastern territories. In some of these territories they never instituted ghettos because killing the jews right away was thought to be economically cheaper or more efficient (e.g. Ukraine). In others (Poland, etc), they thought that ghettos wouldn't work in the long term, and they had or feared local management problems. It's easier to kill a hostile conquered population than to keep it from revolting while it starves. (They also had plans to kill most of the Poles, but they never had the manpower to carry those out.)
Later, towards the end of the war, it acquired an additional (rational) reason: killing all the jews quickly before the relevant territories were lost to the advancing Red Army.
There are at least two 'increasing their power'. One is individuals gaining power in Germany. That was a subgoal. The other is increasing the power of their nation or race in the world. That is the supergoal that, according to their ideology, all nazis should be working for. Beyond that they could have personal selfish goals. Perhaps including survival, reproduction and a legacy of having achieved something constructive. All of which they failed at.
In the values they expressed they took an extremely long perspective. They would extablish a thousand year empire, when they designed a building the first thing they thought of was what it should look like as a ruin and they included Darwinian evolution in their methods and considerations, as though their plans covered at the very least tens of thousands of years. But in their actions they threw away every chance at creating something stable for some futile gestures and a bunch of murders.
And wasting their forces to conquer territories that they knew that they would lose immediately was in itself irrational.
They didn't need the final solution to protect themselves from jews. They "had to" do it because they had used the jew as their outside threat to make the germans fall in line at the start of their rule, and they had internalized their own propaganda to the point that killing jews became a supergoal.
Not every chance. A small reduction in probability. The key word in my argument that you're not addressing is that they made a tradeoff.
if you assume (without evidence) that they didn't believe in their stated anti-Jewish goals at first, and that it was only propaganda, then you're undermining the whole discussion. We started out by considering their success in terms of their own goals and under the assumption they were being rational.
By this subgoal stomp, do you mean that when their subgoal (Killing Jews) became a priority, that it diverted resources away from their supergoal (world domination, or at least the domination of Europe/Asia), and thus they stomped on their own supergoal by irrationally promoting the importance of the subgoal?
If that is what you mean, then I think that a strong case could be made for this, as a lot of manpower was sucked into that particular void.
Yup. Exactly.
No, rationality == winning by definition. Intelligence doesn't seem to be well defined.
Rationality = having accurate beliefs about the world, not winning. If Eliezer Yudkowsky were serving a life sentence in a prison in Pakistan, he would be much more rational than the average person, but much less capable of winning.
Eh, I'd just convince them to let me out.
Hahahaha... How long have you been saving that one?
I suspect that it's easier to talk your way out of an AI-box than a prison cell.
Perhaps, but I'm a bit stunned at the idea that being in a prison cell means you should just give up. Have I taught you all nothing?
Well, come on, the first ideas I thought of wouldn't work in that situation. What do you expect me to do, think about it for five minutes?
The full phrase is "think about it for five minutes by an actual physical clock".
Good point.
Or, if you like, you can take it from Amanda.
Rationality == winning, but you still have to work with what you're given, obviously. Given that you're serving a life sentence in a Pakistani prison, (perfect) rationality will let you make the best of it.
Given that you have to figure out how to win, not just decide to win, it's not that trivial!