<D <M <Y
M> D>

[Comments] (2) Hundred Dollar Brain: I just finished Len Deighton's 1966 computer-age thriller Billion Dollar Brain and unfortunately must report that it's much less computery than I'd hoped. Deighton wrote an excellent alt-history, SS-GB, so I'd been hoping for some retro SF or at least sciency fiction, but in this novel the titular Brain is naught but a minor piece of set dressing, to the extent that I kind of want to write the spy novel that seemed to be taking shape and which would have been really groundbreaking had Deighton gone there.

Basically, if you're using a computer with a telephonic voice interface to run a privately-funded spy ring in 1966, there's no guarantee the individual actions of your agents add up to what you're trying to do. You're incredibly vulnerable to the ELIZA effect. Someone else could be using your computer and your agents to run their own spy ring! (Again, this is not what happens in Billion Dollar Brain.)

I will reproduce the most technically sophisticated paragraph in the book, since it's clear Deighton at least talked to someone who knows computers and I like to see that rewarded:

"I don't want to bore you," Harvey said, "but you should understand that these heaps of wire can practically think — linear programming — which means that instead of going through all the alternatives they have a hunch which is the right one. What's more, almost none of them work by binary notation — the normal method for computers — because that's just yes/no stuff. (If you can only store yeses and noes it takes seven punch holes to record the number ninety-nine.) These machines use tiny chips of ceramic which store electricity. They store any amount from one to nine. That's why — for what it does — this whole set-up is so small.

No, please, bore me!


[Main]

Unless otherwise noted, all content licensed by Leonard Richardson
under a Creative Commons License.