Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Ghatanathoah comments on In Praise of Boredom - Less Wrong

23 Post author: Eliezer_Yudkowsky 18 January 2009 09:03AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (100)

Sort By: Old

You are viewing a single comment's thread. Show more comments above.

Comment author: Ghatanathoah 15 June 2012 05:51:20AM *  3 points [-]

Still think it is "absurd"?

Wow, just wow. I'm extremely disappointed with Schneider and Sagan. Not because of their actual research, which looks like some interesting and useful stuff on thermodynamics. No, what's disappointing and embarrassing is the deceitful way they pretend that they've discovered life's "purpose." Like many words, the word "purpose" has multiple referents, sometimes it refers to profound concepts, others times to trivial ones. Schneider and Sagan have discovered some insights into one of the more trivial concepts the word "purpose" can refer to, but are using verbal sleight of hand to pretend they've found the answer to one of the word's more profound referents.

When someone says they are looking for "life's purpose" what they mean is that they are looking for values and ideals to live their life around. A very profound concept. When Schneider and Sagan say they have found life's purpose what they are saying is, "We pretended that the laws of physics were a person with a utility function and then deduced what that make-believe utility function was based on how the laws of physics caused life to develop."

Now, doing that has it's place, it's easier for human brains to model other people than it is for them to model physics, so sometimes it is useful to personify physics. But the "purpose" you discover from that is ultimately trivial. It doesn't give you values and ideals to live your life around. It just describes forces of nature in an inaccurate, but memorable way.

I'm not saying it's absurd that to say that entropy tends to increase, that's basic physics. But it's absurd to pretend that entropy is the deep, meaningful purpose of human life. Purpose is something humans give themselves, not something that mindless physical laws bestow upon them. Schneider and Sagan may be onto something when they suggest that life has a tendency to destroy gradients. But if they claim that is the "purpose" of human life in any meaningful sense they are dead wrong.

Comment author: pragmatist 15 June 2012 06:39:18AM *  1 point [-]

I read Into the Cool a while ago, and it's a bad book. Schneider and Sagan posit a law of nonequilibrium thermodynamics: "nature abhors a gradient". They go on to explain pretty much everything in the universe -- from fluid dynamics to abiogenesis to evolution to human aging to the economy to the purpose of life to... -- as a consequence of this law. The thing is, all of this is done in a very hand-wavey fashion, without any math.

Now, there is definitely something interesting about the fact that when there are gradients in thermodynamic parameters we often see the emergence of stable, complex structures that can be seen as directed towards driving the system to equilibrium. But when the authors start claiming that this is basically the origin of all macroscopic structure, even when the "gradient" involved isn't really a thermodynamic gradient, things start getting crazy. Benard convection occurs when there is a temperature gradient in a fluid; arbitrage occurs when there is a price gradient in an economy. These are both, according to the authors, consequences of the same universal law: nature abhors a gradient.

Perhaps Schneider has worked his ideas out with greater rigor elsewhere (if he has, I would like to see it), but Into the Cool is in the same category as Per Bak's How Nature Works and Mark Buchanan's Ubiquity, a popular book that extrapolates useful insights to such an absurd extent that it ventures into mild crackpot territory.

Comment author: timtyler 15 June 2012 10:11:31AM -1 points [-]

But when the authors start claiming that this is basically the origin of all macroscopic structure, even when the "gradient" involved isn't really a thermodynamic gradient, things start getting crazy. Benard convection occurs when there is a temperature gradient in a fluid; arbitrage occurs when there is a price gradient in an economy. These are both, according to the authors, consequences of the same universal law: nature abhors a gradient.

That's right - MEP is a statistical characterisation of universal Darwinism, which explains a lot about CAS - including why water flows downhill, turbulence, crack propagation, crystal formation, and lots more.

Schneider's original work on the topic is Life as a manifestation of the second law of thermodynamics.

Comment author: Ghatanathoah 15 June 2012 05:11:28PM -1 points [-]

Of course, while this work has some scientific interest (a fact I never denied), it is worthless for determining what the purpose of intelligent life and civilization should be. All it does is explain where life came from, it has no value in determining what we want to do now and what we should do next.

Your original statement that started this discussion was a claim that our civilization maximizes entropy. That claim was based on a trivial map-territory confusion, confounding two different referents of the word "maximize," Referent 1 being :"Is purposefully designed to greatly increase something by intelligent beings" and Referent 2 being: "Has a statistical tendency to greatly increase something."

When Eliezer claimed that intelligent creatures and their civilization would only be interesting if they purposefully acted to maximize novelty, you attempted to refute his claim by saying that our civilization is not purposefully acting to maximize novelty because it has a statistical tendency to greatly increase entropy. In other words, you essentially said "Our civilization does not maximize(1) novelty because it maximizes(2) entropy." You entire argument is based on map-territory confusion.

Comment author: timtyler 15 June 2012 10:04:27PM *  -1 points [-]

Your comment is a blatant distortion of the facts. Eliezer's only references to maximizing are to an "expected paperclip maximizer". He never talks about "purposeful" maximisation. Nor did I attempt the refutation you are attribting to me. You've been reduced to making things up :-(

Comment author: Ghatanathoah 18 June 2012 04:23:05AM *  0 points [-]

Eliezer's only references to maximizing are to an "expected paperclip maximizer".

Eliezer never literally referred to the word "maximize," but the thrust of his essay is that a society that purposefully maximizes, or at least greatly increases novelty, is far more interesting than one that doesn't. He claimed that, for this reason, a paperclip maximizing civilization would be valueless, because paperclips are all the same.

Nor did I attempt the refutation you are attribting to me.

You said:

Our civilisation maximises entropy - not paperclips - which hardly seems much more interesting.

In this instance you are using "maximize" to mean "Has a statistical tendency to increase something." You are claiming that everything humans do is uninteresting because it has a statistical tendency to increase entropy and destroy entropy gradients, and entropy is uninteresting. You're ignoring the fact that when humans create, we create art, socialization, science, literature, architecture, history, and all sorts of wonderful things. Paperclip maximizers just create the same paperclip, over and over again. It doesn't matter how much entropy gets made in the process, humans are a quadrillion times more interesting because there is so much diversity in what we do.

Claiming that all the wonderful, varied, and diverse things humans do is no more interesting than paperclipping, just because you could describe it as "entropy maximization" is ridiculous. You might as well say that all events are equally uninteresting because you can describe all of them as "stuff happening."

So yes, Eliezer never used the word "maximize" but he definitely claimed that creatures that didn't value novelty would be boring. And you did attempt to refute his claim by claiming that our civilization's statistical tendency to increase entropy means that creating art, conversation, science, etc. is no different from paperclipping. I think my objection stands.

Comment author: timtyler 15 June 2012 10:13:16PM 0 points [-]

When Schneider and Sagan say they have found life's purpose what they are saying is, "We pretended that the laws of physics were a person with a utility function and then deduced what that make-believe utility function was based on how the laws of physics caused life to develop."

When biologists say "the purpose of a nose is smelling things" you don't have to personify mother naure to make sense of what they mean. Personifying the organism is often enough. Since the organism may not be so very different from a person, this is often an easier step.

Comment author: Ghatanathoah 18 June 2012 03:50:42AM 0 points [-]

When biologists say "the purpose of a nose is smelling things" you don't have to personify mother naure to make sense of what they mean. Personifying the organism is often enough.

That doesn't change the fact that personification is a way to help people think about reality more easily at the expense of accurately describing it. Noses don't literally have a purpose. It's just that organisms that are good at smelling things tend to reproduce more.

The problem with Schneider and Sagan is that they confound this metaphorical meaning of the word purpose (the utility function of a personified entity) with a different meaning (ideals to live your life around). Hence their second book makes the absurd statement* that, when you strip the word "purpose" from it basically says "knowing that decreasing entropy gradients is a major reason life arose will give you ideals to live your life around." That's ridiculous.

*To be fair that statement was a cover blurb, so it's possible that it was written by the publisher, not Schneider and Sagan.