Slate Magazine has been looking at the issue of job automation: just as automotive production line workers have seen themselves replaced by robots, so too could people in many other occupations, from pharmacists to journalists. A peculiarly relevant part of this project has been an attempt to create an automatic version of Jason Kottke, proprietor of kottke.org (which is, not coincidentally, the place where I came across this story). The result, Robottke, is here. The idea of Robottke is not that it writes a blog, but that it trawls the web for the same kind of interesting link that Kottke himself does. Because what is supposedly unique about kottke.org is that Kottke has excellent taste in stories: he has a gift for selecting the very most piquant stuff from amidst the vast mediocrity of the web.
Robottke works by scanning many of the websites and Twitter feeds that Kottke himself follows and searching within them for words that match Kottke’s post tags (you know, the words you use to label your blog posts. You can see mine in a cloud to the right). The result is reasonably impressive: probably only half jokingly, Kottke actually chose a link from Robottke for the very next story after he linked to Robottke itself. It’s like this just happened:
Of course, just choosing similar links isn’t really enough to recreate Kottke: the charm of the blog is how it makes you interested in the things it’s chosen. Robottke just links; Kottke thinks of a nice headline, introduces the subject, selects a relevant paragraph of text. The big question is how long it would take to program a more advanced Robottke that could do all those things itself. Slate’s article looks at the work of a couple of companies that might one day be able to do just that.
At the moment their best efforts appear to be writing brief summaries of baseball matches, essentially selecting and inserting sportswriting cliches between the match stats in order to lend them some colour. But you can see how this could quickly develop into something more sophisticated, given a larger number of input variables, a bigger database of linking clauses and rules for their insertion.
In fact I disagree with the Slate writer Farhad Manjoo when he tries to claim that computers in their present state can’t replicate human creativity. He seems to fall into the familiar trap of assuming that we are able – somehow – to create something out of nothing:
While I recognize that there are algorithms that power my work, my columns are also peppered with stuff computers can’t master… Two of my most popular Slate articles this year—a rant about people who erroneously use two spaces after a period, and a rant about the snobs who write letters to NPR—weren’t inspired by my rigorous daily Web searching. They just jumped into my mind out of the blue. (This happens to me at least once a week in the shower.)
Could a computer have come upon these topics? Not today. If we define human creativity as a kind of invention meant to please other human beings, machines would seem ill-suited for the task—computers are good at copying, not at coming up with wholly new things.
The key phrases here are “jumped into my mind out of the blue” and “wholly new things,” both of which are obviously erroneous in the sense that the writer uses them – and could be replicated, in effect, by a computer. The guy behind Everything is a Remix has summed all this up in a very beautiful way so you should really go watch his videos, but the argument is, essentially, that when we talk about creativity we’re really talking about the recombination, recontextualisation or reinterpretation of existing things. I think it’s a dangerous arrogance to assume that we won’t be able to teach computers to do all that before too long.