Re: Artificial life likely in 3 to 10 years


"DK" <dk@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:fbfbko$2f8t$1@xxxxxxxxxxxxxxxxxxxxxx

[...]
Whoever are people producing these graphs, they are quite
obviously people who have next to zero knowledge of biology.
99% of biochemists/cell biologists/physiologists are likely to just
laugh at it and never take it seriously.

Most of my career has been in the area of pattern recognition. I do not take
these graphs, or Moravec's claims any more seriously than you do. I guess I
find them less amusing and more embarrassing than you do, since these are
supposedly experts in my field. Perhaps they should be compared to the
wildest claims made by biologists during the human genome project.

The problem is not the hardware of course. The problem is the software,
which will need to be several orders of magnitude more complex than anything
produced so far, and as reliable as the very most reliable software we have
ever managed. This:

"Sorry, this program must shut down. Any data you were working with may be
lost."

is not an option. Even if we already had hardware that worked a squillion
times faster than at present, and knew how to write a program displaying
human-level intelligence, and had written it, I reckon the test-and-debug
cycle would take a few hundred years.

Graham Jones





.