Standards in Artificial Intelligence
nmm1 at cus.cam.ac.uk
Thu Nov 2 12:45:56 EST 2000
In article <handleym-0211000923240001 at handma2.apple.com>, handleym at ricochet.net (Maynard Handley) writes:
|> It took what, 150, 250 years to understand how life reproduces, depending
|> on where you start the clock. And at the end of it all, the solution did
|> not come from the smartest person in world thinking his/her way to the
|> problem, but from a million insights collected along the way, including
|> apparently irrelevant things like the chemistry of phosphorus, or what
|> happens when you shine X-rays on crystals.
|> At the very least the AI program has started that sort of work in the
|> domain of conginitive science. Most people have, for now, given up on the
|> idea of understanding "thinking" and are settling for things a whole lot
|> simpler (how can I get a robot to walk? how can I get three computer
|> life-forms to learn from each other?). This may not be very glamorous, and
|> it may take 150 years for the collected insights to get us to the goal.
|> But the history of this sort of approach is that is works pretty well.
Exactly. And often what you learn on the way makes you realise that
you were aiming for an inappropriate goal, so you end up going
somewhere else! Anyway, what you learn on the way is often more
valuable than the final objective.
Your example of life processes is a particularly good example of this
aspect. A scientist of 1800 would be completely boggled at where
biochemistry and biology are today.
University of Cambridge Computing Service,
New Museums Site, Pembroke Street, Cambridge CB2 3QG, England.
Email: nmm1 at cam.ac.uk
Tel.: +44 1223 334761 Fax: +44 1223 334679
More information about the Neur-sci