Post by Larry DeackHave you read Harry Harrison's The Turing Option (1992) (with Marvin
Minsky) They put it at 2023. I'll buy you a beer if we are both around
for that one :-)
Yeah, I did read that, and a bunch of other pie-in-the-sky stuff -- I
work in that field, I've met Harrison, grew up reading him and Minsky.
Smart fellows, but like other predictions of the future, it's purely
guesswork. If we don't blow ourselves up, yep, we'll have a future, but
nobody can predict what it'll look like. Without a working crystal
ball, it's all extrapolation from the known, and there is always an
X-factor that screws things up. Something that causes us to veer
slightly to the right instead of the left, and takes us down a path
nobody could have expected.
Back in the thirties a bunch of the smartest scientists got together,
the cutting edge guys, and predicted what 1969 would look like. Those
mile-high buildings, rocket packs, dirigibles, and hundreds of channels
of radio. They got the last one right, but missed television (already
invented) computers, passenger jets, even transistors. No LEDs, LCDs,
no EEGs, no CAT scans, such things never occurred to them.
Go back fifty years and look for somebody who predicted what things
look like today. Nobody did. People got little pieces of it, and some
of those were actually seeds -- Telstar, cheap computers -- but mostly
somebody who claims to be a futurist is blowing smoke and waving
mirrors. We can't predict the weather two weeks out, save in the most
general of ways; the chaos of that is beyond us. Trying to work out the
sociology, biology, and technology of civilization ten or twenty years
out? Waaay beyond our capabilities.
That's why we aren't on Mars, and why the cities don't look like Fritz
Lang's Metropolis -- thanks for that one, Steve -- we zigged when the
sci fi guys said we'd zag.
Certainly with a sophisticated enough program and the tools, Alan
Carruth or his like, could probably program a machine that would
duplicate much of what he can do by feel and experience. But having a
machine that can take apart one of his guitars, weigh and measure
everything, and then build an identical one from scratch -- one that
will sound the same? Not by 2023 CE. Alan can't make two exactly the
same and a computer that can thin a top exactly as the last one
probably won't be able to match the last one because wood isn't an
exact medium. The luthier AI might not be able to look at the wood and
realize it ought not to be exactly that thin because of some very hard
to pin down element even Alan can't quantify. An AI as complex as the
human brain? Not in twenty years, not gonna happen. The
suddenly-sentient internet is a great science fiction concept, but
CyberDyne Systems isn't about to give us SkyNet ...
Somebody could come up with a computer program that writes better than
I can. But one that beats John Locke? I'm not holding my breath. If I
live another twenty years, I don't think I'll see it.
Too many variables for our current technology to handle. The science
stuff is easier than the art stuff. Someday, maybe, but we won't live
to see it.
--
Steve