Technolog

Here’s a thought I sometimes find depressing. I write about 5,000 words a month for the pages of PC Pro. Let’s briefly pretend that the Dickensian era of a job for life still exists and assume I worked for PC Pro for my entire professional career – approximately 45 years. At 5,000 words a month, that’s 60,000 words a year and around 2.7 million for my entire working life. This is just the copy that appears in the magazine, of course; in addition, I write about 500 words a day in emails and then often go home and write another thousand or so for my personal journal or one of the numerous half-started novels most journalists have kicking about the place. Say, two thousand extra words a week.

Totalling that lot up, we can say my output as a writer, if I did it every working day for 45 years, equals a moderately impressive figure of just over 13.2 million words in a lifetime. Going a bit further and calculating the average number of characters per word in my writing (by scientifically invoking the Word Count function on this column so far), you get four-and-a-half characters. Including spaces, that’s five-and-a-half characters per word. So, those 13.2 million words consist of around 72.8 million individual characters. Bear with me.

A character in a standard ASCII text file occupies one byte. And so this is the potentially depressing part: my entire lifetime’s work – even assuming I take no holidays and keep stoically churning out words while everyone else is trying to get their Christmas tree lights to work – will fit comfortably on a 128MB USB flash drive, leaving almost 50 per cent of its capacity spare. And you could probably get it down to about 16MB if you zipped it up. My whole life. You can’t even buy 16MB USB flash drives any more, they’re so outdated.

There’s been a fair amount of publicity recently surrounding Ray Kurzweil’s new book The Singularity is Near. Ray is an inventor and futurologist. I suppose you could call him a latter-day Arthur C Clarke, genetically spliced with Trevor Baylis by way of Nicholas Negroponte. His theory is that all the other futurologists constantly get it wrong by calculating their timescales based on the assumption that the pace of technological change is linear, whereas in fact it’s getting faster on an exponential basis. There’s good evidence for this in Gordon Moore’s assertion that transistor density and hence processor power double every 18 months. And so Kurzweil is confident that we’ll have machines able to exhibit human-level intelligence by a week next Tuesday, based purely on the observation that computers are getting really fast, really quickly and we’ll soon have one powerful enough to completely simulate the human brain. And since the exponential growth in technological pace will continue to increase, he predicts being able to fit human intelligence in a pinhead and achieve larger AI machines with millions of times that reasoning power in very short order. This is his ‘singularity’.

It sounds fantastic. Maybe it is, maybe it isn’t: it’s impossible to refute arguments based in the future, which is why futurology is great work if you can get it. I would say, though, that Kurzweil’s reasoning is an inappropriately statistical way of arguing for something that’s a far subtler problem than just getting hold of a bigger, faster box that goes bing.

Nonetheless, it’s true that one decade you can be dreaming of the day computers will be powerful enough to display real, moving video, and the next you can find yourself in the queue for the bus, twizzling the touch-dial on a tiny white box containing what used to be your impressively large CD collection and wondering if Rymans sells Memory Sticks big enough to hold an entire series of The League of Gentlemen.