Click here to support the Daily Grail for as little as $US1 per month on Patreon

Tuesday Roundup 29-08-2006

A strange assortment to get you through the week…

Enjoy!

Editor
  1. Artificial Intelligence
    Way back, years ago, in the 60’s, I programmed a computer to work out for itself how to add two numbers together. The point of the exercise was the beginning of research into artificial intelligence to see if it was possible to program a computer to randomly attempt to make sense out of a simple request. In those days I believed that it was possible for a computer to become conscious.

    No way.

    I found out then, and I have been convinced ever since, that a computer will never be able to achieve the things that a human achieves. Yes they can remember more things, yes they can search quicker. But come on folks – since when is “intelligence” equal to remembering things quickly? In order to be intelligent you have to work out how to tackle an entirely NEW problem. That is also what is wrong with all so-called “intelligence tests”. They only test your ability to learn from all the previous intelligence tests you took. They never test your ability to “think outside the box” as is the fashionable way to call this these days.

    How do you program intuition? You dont. How do you program clairvoyance (assuming for the moment that it exists). You dont. How do you program suspicion? reading-between-the-lines? What-that-woman-thinks-of-you? Do not tell me we don’t need that – that is confining and constraining what we mean by intelligence.

    A computer basically does one-thing-at-a-time. A human processes thousands – maybe millions of signals simultaneously, matches them and mixes them, and comes up with a response. Does your wife not like coffee today, but she always drank it before? So maybe she’s pregnant, or having a period, or is feeling sick after last night’s meal, or just plain sick in the stomach because she has to meet her boss today who might fire her for that job she happened to be responsible for and someone else messed it up. Can a computer tell you why she doesn’t want coffee?

    There is no chance that computers can come to that level. There are of course areas where they will be of great value, such as economics (but not yet), space-shots, engineering, even strategies, things like that. But humans are irrational-and-still-understandable. How are you going to program your fancy artificially intelligent computer to tell you what your enemy is most likely to know about you? His/her weakness? His/her strength? His/her intelligence?

    People talk about uploading their brains to a computer. Better them than me I say. But what would be the result? It will all be facts and figures from their experiences. No feelings, attitudes, ideals, philosophy. No HUMAN values at all. No fancy personal ways to kick a football or say just the right thing to make your boss appreciate you.

    Yes, we will get “artificial intelligence”. But it wont be what they say. It will be machinery, electronics – and electronics can ALWAYS be fooled.

    Neanderthal
    ‘Whatever you do, comes back to you’

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Mobile menu - fractal