7 years of NKS--and its first killer app | Hacker News Reader

7 years of NKS--and its first killer app

blog.wolfram.com

10 pointsph0rque11 comments

Not convincing.

The most interesting part is this: "As a practical matter, many algorithms in Wolfram|Alpha were found by NKS methods—by searching the computational universe for programs that achieve particular purposes."

But without giving details on what was done and discovered, it's not convincing. Moreover, the method used would need to be compared with related work such as genetic programming.

I've observed you are totally negative about their attempt (you post similar negative comments here on HN). At least they are trying and I won't judge if I don't test it first. Also, from what his says it's obvious that they do realize it's a long term project.

Stephen Wolfram is a very smart guy. He obtained his PhD in particle physics at Caltech at a very, very tender age. He created Mathematica. He has an impressive track record. However, his obsession with that NKS baloney puzzles me. I don't understand why such an intelligent man is so enchanted with cellular automata and the like, which have relatively little value other than the recreational one. Here's a perfect example of a powerful mind which was hijacked by a silly idea.

If you think his book is about CAs, you haven't read the book.

I read the book. I would call it a New Kinds of Recreational Puzzles (except that they're not THAT new). It's interesting, but it's not Science. And Wolfram could try to be humble once in a while. And he could also try to give credit where credit is due. He's a very smart man, but he's not the only one. And if he's gonna stand on the shoulders of giants, the least he could do is acknowledge the contributions these giants gave.

"A Collection of Reviews of ANKOS and Links to Related Work"

http://shell.cas.usf.edu/~eclark/ANKOS_reviews.html

I have read the book, right when it was first published. I don't deem myself competent to judge Wolfram's arguments or his critics' arguments, but I did think the lack of scholarly citation of prior literature was more than a little off-putting.

I have to disagree, the field complexity (I don't call it NKS) since I was introduced to it by reading stuff from the sante fe institute is a unifying force across disciplines. The large amount of rules that explain the world today are manifestations of simple rules acting in complex ways so that the findings of today describe the emergent behavior of these simple laws. Cellar automata is just one way of tackling the problem.

Complexity is a wonderful thing and there are many complex phenomena which arise from simple rules. But unless you can make predictions that agree with observation and experimentation... it's not Science. The nexus between Physics and Complexity is very interesting and beautiful, but it's not necessarily very useful. If Physics were "digital", many symmetries would likely be violated, and that can't happen. The discrete world and the continuous world are entirely different beasts, and though I appreciate Wolfram's efforts and his recreational NKS stuff, it's a few millenia away from being a new kind of Science.

"Predictions are nice if you can make them. But the essence of science lies in explanation, laying the bare fundamental mechanisms of nature." -- M. Mitchell Waldrop

I agree. However, how can the fundamental mechanisms of Nature be explained if we do not confront theories with observation and experimentation? Physics is a fortunate field because the beauty of the mathematics is a good indicator of whether a certain theory is correct or not (excluding the Standard Model ;-) but other fields of the Natural Sciences cannot afford that luxury. I don't want to start discussing the Philosophy of Science because I am not qualified for that...

Actually, Quantum Theory states that the world, at the lowest level we currently conceive of, is discrete and 'digital'. Time, space, energy - everything - come in little chunks called quanta which represent the smallest possible subdivision that the universe permits. For example a photon is a quantum of energy - the smallest possible unit of energy. If you had a perfect dimmer switch, that would be the lowest setting - turn it even a fraction more and it's off completely. The universe isn't continuous at all - it just looks like it from waaaay up here where we live. If you zoom in far enough, you can see the pixels of reality. All properties of the universe behave this way, according to our best theories and best abilities to test them. Given that this applies to space and time too - look up 'plank length' - you can see why this would make people make the connection to cellular automata. Zoom in far enough and quantum theory tells you that space is divided up into minimum sized chunks, and quantum events can only happen within this framework.

Reply on news.ycombinator.com