Why it’s hard to make machines think original thoughts | Hacker News Reader

Why it’s hard to make machines think original thoughts

thinkartificial.org

14 pointshhm11 comments

Yet a predictable system doesn't always lead to expected output.

I once heard a story about a student, probably graduate, who wrote a very complex genetic algorithm, complete with a physics engine, to build (out of blocks representing legs, arms, etc) a creature that could best walk to a finish line.

But what it actually did, after months of work and weeks of running, is stack all the blocks on top of each other and have the creature fall over the finish line, because it only measured how fast the creature reached the finish line, not whether the entire creature crossed it.

And don't tell me that isn't a creative (albeit rules-lawyering) solution to the problem.

There was another genetic solution to the problem. They put energy using dampers on all the joints of a walking figure, except one, the neck. After running the simulation, they found the figure, diving head first into the ground, balancing on it's head, tipping, and then repeating, in this bizarre sort of half cartwheel.

Sounds like a model that reinforced his/her preconceived ideas of racing to a finish line! After all, what else are computer models :)

And yet there are at least "15 instances where genetic programming has created an entity that either infringes or duplicates the functionality of a previously patented 20th-century invention, 6 instances where genetic programming has done the same with respect to a 21st-centry invention, and 2 instances where genetic programming has created a patentable new invention."

See http://www.genetic-programming.org/

There are also neural networks that have created patentable inventions. One of them is mentioned in the article.

creativity seems to be a word with no referent. what we really seem to mean by creativity is non-obvious solutions to complex problems arrived at by unknown means.

This article tends to define creativity as 'something that entertains humans' which is in my view a seriously flawed perspective.

A bird call would entertain another bird more than the greatest music ever written. Human attempts at imitating the bird call are probably at best derivative and usually nonsensical. So from a bird's perspective, humans could be regarded as less creative than birds. Similarly, judging an AI system's creativity by something inherently human as art or literature is to use an unfair measuring stick.

P.S. Actually I don't really know what the state of the art is in bird call simulation is but you get the point right.

You are wrong about the article. It explicitly mentions that human activities are "nowhere near the whole story" and that even primitive organisms can be considered creative. The statement "creating something new or being original is an essential part of creativity" is the only one used to explain what creativity is, and to explain this particular problem.

The problem in question affects all systems, whether they are human-like or not.

> It explicitly mentions that human activities are "nowhere near the whole story"

I read this as being relating to artists (specifically graphical) rather than the entire class of human activities.

> even primitive organisms can be considered creative

Must admit I missed this as the sentence was in what I thought was a plug for a book.

> "creating something new or being original is an essential part of creativity"

This didn't really mean anything to me.

Most the article was explained using specific examples which were human specific activities and for me that was the main thrust. Perhaps it wasn't the authors intention but that is the impression it left on me even re-reading it.

Do you think that at least some part of the problem is the difficulty involved in getting people to think original thoughts?

I thought of this too. I suspect that this question is somehow "obvious."

Reply on news.ycombinator.com