Sep

15

… according to a new book by Ian Ayres, an econometrician and law professor at Yale, this is a microcosm of a powerful trend that will shape the economy for years to come: the replacement of expertise and intuition by objective, data-based decision making, made possible by a virtually inexhaustible supply of inexpensive information. Those who control and manipulate this data will be the masters of the new economic universe. Ayres calls them "Super Crunchers," which is also the title of his book, the latest attempt to siphon off a bit of the buzz that surrounds the hugely successful Freakonomics.[Abstract from Newsweek]

Intuition replaced by statistics. Should one therefore learn to ignore one's intuition or at least ascribe less value to it? And what about quick heuristics, rules of thumb, 'blink'-like judgments… and millions of years of instinctual bias?

Roger Arnold asks:

Are there computer systems that are being designed to handle macro issues as well? I would think that would be highly complex and beyond the scope of computers today.

Nigel Davies replies:

MasadaI can answer that for you - they don't have a hope. And I can tell you exactly why:

Despite huge resources' having been pumped into "solving" a tiny, limited game called chess, computers are just rubbish at the kind of creative synthesis of ideas at which the human brain excels. And it shows. They are totally unable to balance factors such as doubled pawns against the initiative. They just don't "think" like that.

Sure, they've made "progress," they can now beat human chess players by employing huge processing power, crunching a zillion variations a second and never getting tired. The Romans had a much better win when they took Masada.

Vincent Andres remarks:

You don't have to commit suicide! Computers are our allies — we just have to use them. Of course we have to learn to speak to them.

"I hate computers: they always do what I tell them, never what I want!"

Nigel Davies explains:

Deep Blue vs Kasparov 1997 Game 6There have been attempts to run tournaments with the players having assistance from computers — they call it "Advanced Chess." But in activities which enhance our experience of life, computers have no part. It's kind of like having computerized yoga.

The suicide has been the chess world's insistence on pitting man versus machine, which brought the computer manufacturers their Phyrric victory and allowed claims that computers were now showing "intelligence." But to me this is like claiming a Porsche is an athlete if it can beat a human in a marathon. Computers are still just number crunchers as far as I'm concerned.

I like computers; they are nice obedient slaves. But the claims they are showing any kind of intelligence is just bunk. All that has happened is that they're crunching faster.

Does fast crunching lead to consciousness and the human ability to reason? I don't think so. Humans crunch very slowly but are nonetheless able to deal with problems in which crunching is less effective. One demonstration is the miserable failure of computers in Go, which is still a closed game but "bigger" than chess. As for non-closed systems they will, therefore, be utterly hopeless.

There is another issues arising from the way that many humans are now assuming that computers have intelligence and assuming that computerized models are going to work. This viewpoint is not only wrong, the reliance on computerized models can lead to people's suspending their own intelligence or subjugating it to the computer's ideas. This is one of the main problems when human Grandmasters try to look at a chess position with a computer running in the background — they end up letting the computer take the lead.

These are very complex issues which the world will be addressing over the coming decades. But there are great dangers here, and I believe this effect was behind both LTCM and the current banking crisis.

Vincent Andres responds:

Brute force algorithms are used in chess. But they are so many other ways! Computers are doing many other things in so many and so far different ways than brute force. Chess was a challenge for number-crunching, but please don't reduce computer science to that.

Do not confound newspapers and computer scientists. We knew for years what the end of the chess story had to be. Nobody is surprised, nobody is overproud. It was a tedious job — but it had to be done. 10^30 computations had to be done, 10^31 were achieved.

In the computer science community there were few remaining people interested by the human/machine chess battle. Even the finances for the projects were questioned. A complete battle of the past, as far back as 1995 for many of us.


Comments

Name

Email

Website

Speak your mind

4 Comments so far

  1. Market Student on September 14, 2007 4:39 pm

    Maybe those Mayan or Hopi calanders are right. Human world, as we know will come to end in 2012.
    Whenever a group of humans starting to think their logic based mind could outdo unexplainable wisdom of nature, horrible destruction & miserable end followed throughout history. Jews, Egyptions, Mongolians, Persians, Aztecs, Mayans.
    We must be nearer that we think, to putting in the last layer of bricks on our tower.

  2. Steven Buss on September 15, 2007 1:51 pm

    Of course, the distinguished inventor and writer Ray Kurzweil is a leading proponent of the view that some alive today will witness a revolution in information processing that we cannot imagine.

    Indeed, many of Kurzweil’s predictions seem farfetched and there are other computer scientists who sharply differ with Kurzweil’s vision. Nevertheless, it seems to me that this topic is incomplete without reference and links to Kurzweil and other proponents of the view that The Singularity is Near.

    Kurzweil’s website: http://www.kurzweilai.net/index.html?flash=1
    A link to his book: http://www.amazon.com/Singularity-Near-Humans-Transcend-Biology/dp/0143037889/ref=pd_bbs_sr_1/105-6183111-3176429?ie=UTF8&s=books&qid=1189878270&sr=8-1
    A summary of his essential argument: http://www.kurzweilai.net/meme/frame.html?main=/articles/art0610.html?m%3D1

    Lest it be thought that Kurzweil is some lone loon, other folks just as smart (or just as crazy) as he is have begun to hold yearly conferences to discuss whether “within this century, science may move humanity beyond its boundary of intelligence.”

    http://www.singinst.org/summit2007/index.html

  3. Robert McAdams on September 15, 2007 2:50 pm

    How do computers “win” at chess? It seems to me they win because the rules and playing area are fixed, information is perfectly known to all participants and they seek to “not lose”. That wont work in the real world where cycles change, knowledge is incomplete and people risk losing to win. Nigel says they arent creative. I agree. Creativity to me means remembering things learned from different areas of interest and blending them together to make something new. That is one reasons I enjoy DailySpec so much. It gives me things to think about. Some are saying that computers will soon dominate no limit holdem poker. I doubt it. Again you are dealing with imperfect knowledge, cycles in playing behavior and “going for broke”…things computers wont be able to handle.

  4. Matthew Chlapowski on September 15, 2007 2:57 pm

    One flaw with with Ayres’s argument is it fails to take into account the ability of humans to look at the algorithm from the outside. To beat the computer trading system one must only consider what it’s most likely decision will be and act in the way that takes maximum advantage of it. The recent quant fund implosion was a lucrative trading opportunity because it was known (despite what the quants wanted others to believe) how the quants were making their money in the market. Eventually they had to unwind their positions and create a profit opportunity for those who saw it coming. It doesn’t matter who or what is making the decisions, only that they are being made.

    The future will be similar to this because the computer systems will all have access to much the same data and will act in similar ways. This leads me to conclude the markets will be more efficient and it will be harder for groups and large funds to beat the market, but savvy individuals will actually do even better than they do now because of the added flexibility their individual perspectives will provide. The key will be to think outside the box at all times. Computerized systems will be the new box, and the gulf between the most talented intuitive participants and the average ones will be even wider than it is now.

Archives

Resources & Links

Search