[Link] There is nothing like "intelligence", only an evolution is going on

post by Thomas · 2012-01-10T17:54:47.481Z · LW · GW · Legacy · 21 comments

Contents

21 comments

http://edge.org/conversation/infinite-stupidity-edge-conversation-with-mark-pagel

 

Random change, then a selection, says Mark Pagel. As I agree, here's the link.

 

 

21 comments

Comments sorted by top scores.

comment by DanielLC · 2012-01-10T18:42:34.194Z · LW(p) · GW(p)

Ideas are not just natural selection. People do not design computers by randomly messing with transistors of their last version. People do not change computer programs by randomly altering lines of code. It would not work in any feasible time period.

There are things that you can design that way. There are things where that's the best way we know how to design them. That's why genetic algorithms are sometimes very useful. Not everything is like that. That's why we don't use genetic algorithms for everything.

Also, the lack of innovation he talks about seems to be largely that we're not reinventing the wheel. If there's nothing to stop you from stealing ideas, then there won't be sufficient innovation, but that's what we have intellectual property rights for.

Replies from: jhuffman, Thomas
comment by jhuffman · 2012-01-10T22:11:09.793Z · LW(p) · GW(p)

Well, I don't know how other programmers do things, but what I do is I create models at various levels of abstractions, and play with those models until it occurs to me what ought to happen next. Like, I draw on the whiteboard with someone, and we generate all the known solutions we have and usually a known solution or pattern just needs to be adapted. So we're copying. Often, the adaptions aren't immediately obvious when you go down a couple levels and start actually coding it, and so I play around with a couple of different ideas, or maybe I'm guided by theory and just have to think it through and problem solve for this circumstance - which would be copying.

If I'm dissatisfied at any point and have exhausted my search space, what I'm going to do next is enlarge that search space by talking to more people or doing research. Copying.

Every now and then - but pretty rarely - something novel will occur to me that will out-compete all the ideas I already have or can find in research. And so there a new idea made it into the code. And if someone asks me how I came up with, I'll just describe the problem, and the search space and the research and so on, and then shrug. Because the idea seems pretty obvious in hindsight I just shrug away this question of where the idea came from.

It would be difficult to overstate the power of our abstraction and modeling, because it allows us to quickly test for fitness a rash of approaches iteratively. And in the process of that novel ideas do occur and get worked in and then communicated. But I am not at all certain the genesis of most of those ideas isn't more or less random, in the sense the author means it. Indeed most engineering work is copying, as the author terms it.

comment by Thomas · 2012-01-10T18:49:45.818Z · LW(p) · GW(p)

People do not change computer programs by randomly altering lines of code. It would not work in any feasible time period.

At least one program does it. Randomly changes code and evolve it.

http://leehaywood.org/misc/several-unique/

comment by Richard_Kennaway · 2012-01-11T07:54:03.288Z · LW(p) · GW(p)

There is nothing like a rock, only atoms going on.

Replies from: Thomas
comment by Thomas · 2012-01-14T15:58:24.494Z · LW(p) · GW(p)

True. It is good to know that atoms are what a rock is made of. That there is no "nonatomic matter" in rocks.

A neutron star on the other hand is not made of atoms. It's good to know.

The question here is, is something under the hood of intelligence, what is NOT a kind of Darwinian evolution?

Some think it isn't and I agree with them.

comment by Shephard · 2012-01-11T19:51:41.190Z · LW(p) · GW(p)

I think Professor Pagel's specific attempts to articulate the mechanics of human ideation isn't the most interesting take-away from this video. His tone makes it clear that he is playing around with a perspective that is new to him, a rough understanding of human intelligence that deserves further exploration.

The concept that I think is important, and which is certainly not universally accepted by a wider audience of reasonably intelligent and educated people, is that our creativity is not "special". That new ideas aren't magically willed into being by some ineffable desire to innovate, but are instead arrived at via conscious and subconscious pattern-seeking, and a mental "auditioning" of potential solutions.

And natural selection is a decent conversational analogy here. There are people who accept that there is no external intelligent agency which governs the "creation" of advanced biological organisms, but still hold on to the idea that the human mind is somehow captained by an irreducible agent, that there is a ghost in the machine. Drawing a comparison between these two processes is a strong and accessible argument against such a notion.

comment by jhuffman · 2012-01-10T22:14:11.535Z · LW(p) · GW(p)

This article isn't about intelligence, its about innovation. He's talking specifically about the "lightbulb moment" - the inspriation part of invention. I don't think there is anything at all original about the article except the tortured analogy to evolution.

comment by timtyler · 2012-01-10T19:45:12.220Z · LW(p) · GW(p)

I recently wrote a book on this topic - and I don't agree with Mark on this point.

To quote from my blog page about Mark's work:

I note that Mark isn't technically correct about the role of randomness in cultural evolution in his (interesting) "Infinite Stupidity" video. Cultural evolution can use linear programming, extrapolation and other non-random search techniques for exploring solution space. Mark's idea is Donald Campbell's "Blind Variation and Selective Retention" (BSVR) thesis taken to an unrealistic extreme - though he doesn't cite Campbell, Cziko, or anyone else who has weighed in on this issue.

Replies from: Thomas
comment by Thomas · 2012-01-10T19:52:28.586Z · LW(p) · GW(p)

Not only, that there is no "intelligent designer" on the African planes, but the humble evolution is putting the elephants, crocodiles and humans there - not only that. The bright ideas in our heads are born in the same way as those species originates.

Nobody is smarter than evolution.

Replies from: timtyler
comment by timtyler · 2012-01-10T20:02:18.140Z · LW(p) · GW(p)

Using random mutations results in a stupid optimisation technique.

Of course it is trivially true that the mind uses evolution-like algorithms internally. Most optimisation techniques have a basis in trial-and-error. That doesn't mean that they are constrained to use "random" mutations.

Replies from: Thomas
comment by Thomas · 2012-01-10T20:11:01.562Z · LW(p) · GW(p)

Who says? As I said several times here already.

The best human packers, mathematicians and so on, along with theire software are competing on hptt://www.packomania.com to pack the circles as tide as possible.

A program based on EA, with no packing knowledge except that the circles must not overlap and must be squeezed on the smallest square - competes with them successfully. Holds many records.

Who else can?

Replies from: Luke_A_Somers, timtyler
comment by Luke_A_Somers · 2012-01-10T20:30:17.649Z · LW(p) · GW(p)

That's a rather special case - a case where all rational avenues of optimization have been exhausted, and brute force reigns supreme.

There are other regimes.

Replies from: Thomas
comment by Thomas · 2012-01-10T20:41:29.592Z · LW(p) · GW(p)

First of all, brute force isn't equal to an evolutionary algorithm. You could not use brute force to do this.

To squeeze 400 spheres in a cube as tide as possible - can you find me another solution of any kind on the whole Internet?

Or can you do this any better?

Is that complex enough?

comment by timtyler · 2012-01-10T20:40:06.864Z · LW(p) · GW(p)

Optimising using randomness is usually a missed opportunity to use intelligence.

I wrote an essay about this topic back in 2008: Intelligent design vs random mutations.

Replies from: Thomas
comment by Thomas · 2012-01-10T20:46:44.605Z · LW(p) · GW(p)

You have. But I have evolved this with our EA software.

If you are a programmer you can test it and try it and judge it. Or you can tell me if you know a prior art.

Or whatever suits you.

Replies from: timtyler
comment by timtyler · 2012-01-10T21:11:45.828Z · LW(p) · GW(p)

Right - and systems involving human programmers have created Linux, Apache, MySQL and PHP.

Some day machines will be able to do what humans do, but by then they will be using intelligent variation.

comment by shminux · 2012-01-10T20:31:50.952Z · LW(p) · GW(p)

Hammer and nail? To an evolutionary biologist every change looks like biological evolution. I wonder what the name of this particular cognitive bias is.

Replies from: wedrifid, timtyler
comment by wedrifid · 2012-01-11T01:13:12.663Z · LW(p) · GW(p)

Hammer and nail? To an evolutionary biologist every change looks like biological evolution.

I would have guessed that to an evolutionary biologist every change looks like another chance for a confused layman to misuse and trivialize their whole field of study.

Replies from: timtyler
comment by timtyler · 2012-01-11T13:35:03.133Z · LW(p) · GW(p)

Your post is not 100% clear - but if you're suggesting that Mark Pagel is a confused layman, he isn't - he really is an evolutionary biologist.

Replies from: wedrifid
comment by wedrifid · 2012-01-11T13:49:59.876Z · LW(p) · GW(p)

but if you're suggesting that Mark Pagel is a confused layman

(No, responding to generalisation.)

comment by timtyler · 2012-01-10T20:48:04.992Z · LW(p) · GW(p)

Hammer and nail? To an evolutionary biologist every change looks like biological evolution.

That doesn't seem entirely fair - there is a real relationship based on universal darwinism.

This was recognised by B. F. Skinner, Roger Sperry, William Calvin - and now many others.