Mathematica man brews 'AI' Google Killer™ Stephen Wolfram - the lovable George Costanza of the mathematics community who developed the invaluable Mathematica suite and wrote the much talked about but quickly forgotten "A New Kind of Science” - is trying his hand at artificial intelligence. His new project, Wolfram Alpha, set to go live in May, combines natural …
Are we talking translation or AI? translate.google.com Just add machine to the drop down list?
All joking aside, how about a question like:
What is the best looking RPG for the PC?
To any human researcher, they would know to look for role playing games made in the last couple of years that got good reivews for it's graphics and give a short list of possible "best" RPG's for the PC. It should have in it's top 10 either answers or reviews of or someone else's list that includes: Fallout 3, Mass Effect, The Witcher, Oblivion etc.
Instead pretty much all the search engines give you crap. Maybe pages that have the Best game of 2008 etc. but not RPG's. Search engines could do some translation like "best looking" = "best looking" "looks awsome" "looks great" etc. on pages that include RPG and PC. But then you still end up with games that looked awsome in 1996. So it's translation plus AI to understand context. "What is the best looking RPG for the PC? " implies now, or at the moment. So only include the last 3 years or something..... Then again that only applys to the way things look in the PC world. "What is the best looking car" wouldn't necessarily have to be only made in the last 3 years. Who does your computer learn style from? Guess we're back to page rankings.
I continue to be astonished at the number of comments and remarks about NKS, all fashionably negative, by writers (very likely including Ted Dziuba), who have never in fact actually read the book! Instead we learn from Mr. Dziuba that it has been "quickly forgotten", employing a bit of passive voice to include some unknown community of readers.
I can assure you that anyone who has actually read the book with any attention has not forgotten it.
There is a reason for peer review, and NKS is the perfect demonstration. I am not a genius, but I found several places where Wolfram claimed supernatural properties for automata, but which were very ordinary artifacts of simple number theory.
Add in the unfettered "I'm such a stunningly clever bastard" subtext that not-so-subtly underlies every paragraph, and it quickly becomes a sickening experience. "If you focus your attention and listen closely, you can hear his ego approaching critical mass, preparing to implode on itself" - right on the nose.
It appears that Wolfram has not made too many friends due to a perception that he is an egomaniac. Well, egos happen. For an example in the tech world, consider: does anyone actually "like" Larry Ellison? The question seems irrelevant to me, especially when one considers that many important and/or influential thinkers have boasted notably unpleasant egos (e.g. Nietzsche, B. Russell, and almost any 20th century philosopher). One can forgive or hate, or better yet, laugh that aspect off, and then move on.
I appreciated Greg Fawcett's perceptive comments, but would like to suggest that Wolfram may have suggested metaphysical, or maybe even mystical, properties to the results of his testing, but "supernatural" is stretching a bit what I think he was up to. Rather, I think that he was trying to demonstrate through lots of visual examples the interesting philosophical implications of a primitive concept: tremendous complexity can be generated from a remarkably simple algorithm. This may of course be second nature for a mathematician (not me) to understand, but it's a concept that I think is little appreciated in typical thinking about our universe; that is, a teleological view is completely unnecessary, and that the idea of a grand goal or a complex design as the foundation of our everyday reality is not only unnecessary but superfluous. I'm willing to accept that maybe this is an obvious, banal observation, but if so, I would greatly appreciate that someone spread the word to the billions of people who seem to think that an omniscient sky-god has set everything in motion to fulfill his mysterious purpose, or that we're all part of some amazing fuzzy universal-consciousness grand plan that will be revealed in all of its crystal clarity one day.
Concluding thoughts, if Wolfram at least managed to introduce some interesting ways of looking at things from a different perspective, NKS was successful. A great thinker and guru for the ages? Probably not. Creepy guy? Arguably yes. But I still think he has made a worthwhile contribution to modern thought.
@Andy
Google and Wikipedia both rely on the question having been asked before
No. They rely on the words you use to describe your question being used in pages they index or have produced.
@Steven Jones
nobody yet has come remotely close to understanding how natural intelligence works
A nice apparently clear and unequivocal blanket statement. I'd say rather they don't like the answers. The human brain is not a singular processor of Intel or anyone else.
AC@10:18
tell you when my snake oil detector turns on, it is when people wish to use the human brain as a model for AI, and the detector goes off when someone tries to keep it to mathematics
What a brilliant bit of English. A sentence that can be parsed in 2 exactly opposite ways. 1 problem. Your snake oil alarm goes off when someone wants to use the human brain to model AI, If I've parsed it correctly.
You might like to read "Perceptrons" by Marvin Minsky. This killed neural network research for a couple of decades by demonstrating that Perceptrons (a simple class of neural networks) could not solve certain problems humans can -> cannot possibly work as an AI model. Fair enough. Funding competitor neatly killed off.
Its just a pity that human brains, some of which seem to exhibit intelligent behaviour are neural networks. Never let the facts get in the way of a grant application.
@bob
imagine you're re-naming files and your computer could see what you're doing, analyse the criteria you're using and offer to complete the task for you. Or you're sorting photos into folders, Imagine being able to tell your computer "I heard a new tune on the radio this morning - find out what it was and download it to my phone
I note most of what you want does not need much "intelligence" but considerable awareness of "context," Your drives, your radio, your phone. And if you want the pictures sorted by what's actually in them that will be a long way away.
You might like to look up an ancient DOS program called Automator MI.
Otherwise you'd better put your order for Windows 16 in now.
@bounty
"What is the best looking RPG for the PC?"
"So it's translation plus AI to understand context."
And a *lot* of context at that. RPG has at least 3 well recognised meanings. Then there's the aesthetic question. Its not the question you ask, its how much such a system has to know about *you* personally to answer it. And how much of your life you would want such a system to know about you.
You may find its performance a bit disappointing.
Chances of pulling this off. Practically zero. But I try to avoid absolute statements about AI.
Why oh why do the same as*holes come here to read Ted's articles and then diss him about them? If you know he sucks, don't come back to read new articles damn it. What a tiresome practice having to put up with all these sour-pus*es. Get a job or something! And leave your names, don't post anonymously you slimey pri**s.
Nice article Ted btw, enjoyed the witty comments, as ever.
Very funny to have someone who worked for the guy post here as well, gives a neat overall picture of this dude mentioned.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
