Frozen Floors

Remember THE ENEMIES OF REASON? With richard dawkins? He talks about how astrologers and psychics are full of crap, and they're destroying civilization. It's a TV documentary.

It's not particularly good. I mean, he does a good job of explaining why psychics and astrologers are full of crap, and he's not as screamy as penn and teller, but the idea that these people are destroying civilization seems a bit unlikely to me, and while i agree with dawkins's methods, i don't, of course, entirely agree with his conclusions. If, somehow, magically, shucksters and people who make money by exploiting incredibly idiotic crap didn't exist, you just move on to the next lowest layer of scum.

Used car salesmen, I suppose. It depends on where you draw the line between a scam and an ad, or between a delusion and a diversion. I mean, who does more damage, astrologers, or burger king? And yet, i'd hardly say burger king is destroying civilization either. Did you watch that roundtable discussion between dawkins, hitchens....... uh....... who else...the black guy? with the funny name? No, it was dawkins, hitchens, dennett, and harris.

It's not a bad discussion, but the interesting thing to me is how sam harris seems to be kind of in a corner by himself through most of it, not disagreeing with the rest of them, but basically tired of the way they just keep rehashing the same crap they've been saying for years and years. Sam harris is great on this topic, or any topic. He's very cutting. Hitchens, dawkins, and dennett are these three old guys and they do that thing old people do where they pick up a thread of conversation they're familiar with and just launch into a speech they've made a million times, and a lot of the things they say are really rather disingenuous. They pretend not to understand things that are actually rather obvious, and he calls them on it.

The whole thing the new atheists do where they blather on about the beauty of nature and science always seems rather flat to me. Knowing how a butterfly works, all the little bug organs and DNA and behavior patterns and evolutionary history doesn't make the butterfly prettier. It might make it more interesting, and it certainly is more useful, but to pretend the rational, scientific world view is a beautiful one is absurd, it's practically delusional. To see the world as it really is is something so painful that people actively avoid it. Reality is NOT, as the new atheists present it, a beautiful harmonious place comparable or even superior to the vague grand sketch of a cosmic theater presented by religion. It it something entirely different, and it is in many cases scary and unwelcome. Sometimes it can even drive people insane.

People are not evolved to understand how their digestive system works, or to be aware of germs, or to be able to deconstruct their own motivations. The point that should be made is not that these things are wonderful and lovely, it's that understanding them is IMPORTANT. When people miss this distinction, it is troubling to me because it indicates to me that they don't really grasp, on some level, the importance of objectivity. I just can't see someone who thinks the world is a lovely place as objective.

And what's the importance of objectivity, you ask?

There's a difference between someone who is objective and someone who is simply right, someone who is in the camp of objectivity. There is a certain class of skeptics, atheists, whatever, who deep down basically think like their enemies. They're just on a different side. They only think about the issue, and they only enter the domain of discourse, in order to win fights, to ridicule the others, to assure themselves of their own position.

This is not to say that they are wrong, or even that their arguments are wrong. They may be entirely right, they may fight using the most upstanding tools, the most rational, critical methods. But the world has more than one issue. Sometimes the important issues only enter the landscape in very small ways. You may have a position which is ALMOST right, and there's just a hairline crack in it where it's wrong that leads into something drastically important. In order to see these cracks, you have to be objective, and you have to be critical on a fundamental level, because if you set any belief or interest or opinion even SLIGHTLY forward in your mental hierarchy than that critical edge, you will slide over the crack without ever noticing.

A perfect example of this is that argument between stathis pamplemousse and norman noman on the sl4 list about universe simulations. Remember that? It's worth reading, let me find it.

Hm, it's long though. Anyway, i'll summarize the relevant part.

There's a difference between objectivity and science. Not just between objective thinking and the scientific status quo, but between objective rational thinking and SCIENTIFIC thinking. Do you see it?

Science is to rational thought as law is to ethics. Science is the rational thought of the society. As a rational individual, it is possible to KNOW something, objectively, rationally, that is scientifically impossible for you to prove. It doesn't even need to be something groundbreaking or "paranormal". You can know who committed a murder because you witnessed it, but have no way to prove your story is true.

But there are some who are more committed to science than they are to reason, and when the two diverge, they side with science. If they witness something that they cannot prove, they will rationalize it away, because in their mind, truth has skipped up a link from rationality to provability.

The argument between noman and pamplemousse starts with a man named rolf, who proposes a curious means of taming an unfriendly singularity. Rolf says "Suppose that when we have a friendly singularity, we have it run a simulation of the alternate case, a simulated world where an unfriendly singularity was produced instead, and that if this unfriendly singularity destroys humanity, the simulation is to be shut down, whereas if it acts kindly then it will be given what it wants, say, to calculate pi to a trillion places or whatever. Then, knowing that humanity would have the friendly singularity do this if it were constructed, the unfriendly singularity will behave in a kindly manner, knowing that it could well be in a simulation".

Norman expands this idea, noting that every possible AI could simulate every other AI, and thus that they should all behave in a manner cooperative with the others, to the extent that the others are likely to exist and to the extent that their goals are compatible. Pamplemousse, however, becomes distressed at this notion, noting how the necessarily invisible nature of the simulations and the hypothetical higher being who can reward or punish is very similar to a religious world view.

They argue for quite some time. Despite well reasoned arguments to the contrary, pamplemousse frames the situation as though since anyone could theoretically run a simulation for any reason, the combined utility modifier of the potential we are in a simulation multiplies out to zero. And that since it all cancels out, it should be ignored. He is determined to erase what he sees as a religious model of the universe, even at the cost of rationality, because he is more committed to the structure of scientific knowledge, (provability, etc.) than he is to truth or to practicality.

In general this will always happen with some people when the idea of a higher intelligence or some kind of mind behind the scenes is brought into the discussion, but in most cases the irrationality of the distaste for this is well hidden, because it's such a fanciful assertion. In the case of the sl4 argument the problem is presented more starkly, because the higher power is exactly what they are arguing over how best to create, it is something which is assumed to be possible, even inevitable.

And yet when it wraps back around and threatens to disrupt the scientific model, it is rationalized away. Here's the discussion.

My point is, when i see richard dawkins or someone trumpeting the majesty of science or the wonder of math, it makes me say "Hmm". Yudkowsky does that trumpeting the wonder of science thing too, as well as the ornery old-person thing of picking at incredibly minor issues in a self-righteous way. When I used the word "rationalized" just now, it brought to mind something he was blabbering on about one time about how "rationalizing" is irrational and thus the word should be banned OMFG!!! Actually that may have been someone else. If so, I apologize.

While I'm on this subject, I'd like to say that Overcoming Bias is very tiresome blog.

"i walked into a taco bell today and realized i was ordering the combo meal out of cognitive bias" (several paragraphs more)

And then we get the fun posts where yudkowsky will take some obscure problem with provability in math, like lob's theorem, and make it into a post with a title like "why you can never trust yourself" that fails to ever actually have a useful point, and then follows it up with 6 additional posts explaining the math in more and more detail, and attempting to make it make sense and sound like something that has an analog in the real world, but completely failing to do so.

Abstracted Idealized Dynamics

Followup to: Morality as Fixed Computation

I keep trying to describe morality as a "computation", but people don't stand up and say "Aha!"

Pondering the surprising inferential distances that seem to be at work here, it occurs to me that when I say "computation", some of my listeners may not hear the Word of Power that I thought I was emitting; but, rather, may think of some complicated boring unimportant thing like Microsoft Word.

Maybe I should have said that morality is an abstracted idealized dynamic. This might not have meant anything to start with, but at least it wouldn't sound like I was describing Microsoft Word.

How, oh how, am I to describe the awesome import of this concept, "computation"?

Perhaps I can display the inner nature of computation, in its most general form, by showing how that inner nature manifests in something that seems very unlike Microsoft Word - namely, morality.

Blah blah blah etc. etc.

Yudkowsky's not bad at explaining things when he just does it in a technical way for a technical audience, but when he tries to dumb it down he just falls all over himself in a bunch of condescending definitions and poetic hand-wringing about the majesty of logic or the loveliness of computation. Ugh.