Autocorrecting your spreadsheet is bad enough, imagine HAL9000 in charge of autocorrecting your spreadsheet?
Paul Krugman, whose column I read at the NY Times, was recently in conversation with SF author Charlie Stross in Montreal.
It’s a unique discussion, not least because it involves a dismal scientist trying to bridge the gap with a fictional scientist.
There’s a lot of interesting questions raised, like why the rate of technological change hasn’t been able to match the predictions of SF classics like Arthur C. Clarke’s 2001, William Gibson’s Neuromancer, and Greg Bear’s Blood Music.
What you came out believing if you went to the New York’s World Fair in 1964 was that we were going to have this enormously enhanced mastery of the physical universe. That we were going to have undersea cities and supersonic transports everywhere.
And there hasn’t been that kind of dramatic change.
My favorite test, which shows something about me, is the kitchen. If you walked into a kitchen from the 1950’s it would look a little pokey, but you’d know what to do. It wouldn’t be that difficult. If someone from the 1950’s walked into a kitchen from 1909 they’d be pretty unhappy – they might just be able to manage. If someone from 1909 went to one from 1859, you would actually be hopeless.
The big change was really between 1840 and the 1920’s, in terms of what the physical nature of modern life is like. There’s been nothing like that since.
And Stross on Genomics:
They have sequenced quite a few mammalian and other genomes and it’s getting cheaper all the time.
Craig Venter came up with an interesting project a couple of years ago to sequence the Pacific Ocean. If you have a bucket of seawater, it contains probably on the order of a billion organisms most of which are viruses, probably single virus particles in that bucket from a number of species. It turns out when they did shotgun sequencing on a bucket of seawater 98% of the genes they discovered were hitherto unknown. About 90% of those unknown genes were from viruses and we have no idea what the host organisms of them were…basically, viral soup.
There’s a lot of stuff we don’t know about how the genome works. It’s not, as was widely thought in the 50’s and 60’s, a blueprint. It’s more like a very very messy snapshot of a running computer program.
I wonder if they got a few floating genes from Moby-Dick in that bucket? That would explain why the sequencing went haywire.
And on my favourite hobbyhorse, AI, augmented intelligence, and general crackpot conspiracies:
PK: We’ve gone for augmented intelligence, not artificial intelligence.
PK: And it’s the weirdest thing – by finding the eigenvector with the largest eigenvalue you end up in effect doing a computer meld of many peoples’ intelligence without knowing it.
CS: Actually, Amazon is very big on human intelligence emulating AI. They have a system called the Mechanical Turk where they pay people piecework to do basic tasks and farm them out using the network and if you want to throw money at a problem, you can find a hundred thousand pairs of eyes to work on it if you can divide it up suitably.
PK: Whatever the algorithm that Amazon uses to make recommendations…
CS: That scares me.
Scares me too. If you want to read on, here’s the full transcript.