Here’s an interesting essay by Cory Doctorow: “The Internet Will Always Suck”.

His point is that “we always use our vital technologies at the edge of their capabilities.” The Internet sucked in 1995 because it could barely handle images; it sucks today because it can’t reliably deliver high-res movies to moving cellphones in remote areas.

1500 year old Roman comb

1500 year old Roman comb

It’s an excellent point— as technology gets better and more ubiquitous, it’s stretched, and it’ll be used in non-optimal ways, with attendant errors and frustrations. Doctorow pointedly reminds designers to plan for those error conditions… don’t succumb to the engineer’s perennial optimism that things will work as they’re supposed to, or as they do in optimal conditions at the engineer’s desk.

It’s the always that goes too far, though. We’re living in a time of hair-raisingly fast technological development, but that is almost certainly just dumb luck and won’t continue in the same way. Will the Internet still be advancing in leaps and bounds in fifty years? Maybe. In five hundred years? Almost certainly not.

Technologies do mature, and settle down in usable, predictable forms. There’s probably an example in front of you: the QWERTY keyboard, first produced by Remington in 1878, still going strong a century and a half later despite its original purpose (preventing jams on the typewriter) being entirely moot today. It’s not the best design, but it’s stable, and thus allows people to transfer their knowledge between machines and even between technologies.

Automobiles have changed in all sorts of ways in a hundred years, but the user interface of the automobile is nearly unchanged in the last half century. Your grandfather could drive your car, with maybe 30 seconds’ instruction in how to use the automatic transmission (mainly learning not to use the nonexistent clutch). As Bill McKibben points out, you’d have trouble understanding how to make a meal in the typical kitchen of 1900— but that of 1950 would be no trouble.

Many tools have had roughly the same shape and function for hundreds of years or more. The illustration is exactly what it looks like, a comb— the idea of dragging an array of hard spikes through the hair has never been surpassed.

The obvious objection is that there is improvement in automobiles, guns, hammers, pianos, coffee makers, whatever. We don’t make combs out of antler anymore. Well, of course. But we change the user-facing portions the least, and every field doesn’t see the spectacular rate of change of electronics.

Computers are still in rapid development… though I’d note that programming hasn’t advanced anywhere near as fast as computers. You can write a program in Javascript today that’s remarkably similar to the Pascal of 1970. And even if computers stop changing, business hasn’t finished thinking up all the possibilities for transforming services and production.

It’s easy enough to imagine the process continuing for another fifty years. But five hundred? Five thousand? Even sf writers can’t make that convincing; they just mutter about “weakly godlike entities” and talk about something else instead.

I’ll venture a prediction: as soon as you can have sex with robots, we’ll be done. Less provocatively phrased: we’re now trying to stream megapixel movies on demand. Imagine a few more iterations of that: moving hi-res holograms; involvement of other senses; responses to the user’s position and movement. When we have a sensorium that mimics real life… where else is there to go? You essentially have Star Trek’s holodeck… or the capabilities of Second Life in real life. Once you can near-perfectly fool the human senses, that’s all you need to do; there’s little point in a fourfold increase in speed beyond that.  All the engineers will move over to genomics, in order to make furries a reality.

(Well, there’s one more requirement: your 3-d printer needs to be able to create a pizza. Then we’re done.)