I’m back from vacation, where I got a bunch of reading done. So I’ll have several posts related to stuff I read last week in the next day or two. But first I want to note some interesting stuff on the Internet from last week. My friend Reihan Salam has an interesting article in Slate about Futurism:
The best futurists take present-day trends in technology and extrapolate from them based on a few fundamentals: that large-scale institutions will keep being slow-witted, that small groups of people are good at learning and adapting to new circumstances, and that death and taxes will always be with us. Reynolds partisans can sit back and wait for “the comfy chair revolution” to come. Meanwhile, I’ll be stockpiling enough ammunition, Cipro, and NewsRadio DVDs to last me through the coming robot wars.
However, I’m inclined to concur with Matt Yglesias’s critique:
Broad social trends don’t seem to have real directionality. There was a time when the technical innovations of the railroad and the steamship strongly supported the concentration of human living patterns around railroad stations and ports (see Douglas Rae, City: Urbanism and Its Limits). Basically, you had a situation where transportation from port-to-port or station-to-station or port-to-station was quite cheap, but transportation away from the nodes was very expensive. This encouraged people to huddle together quite closely and urbanize, which wound up having implications for how goods were produced, etc., etc., etc.
Mass production of cars, of course, changed all that. And the recent telecommunications revolution seems to have further re-enforced a trend toward decentralization. But it’s not the case that “technology = empowerment = decentralization”–we got so centralized because of empowering technologies in the first place. Even very similar things can have different effects. The use of the radio spectrum to broadcast radio and television shows encouraged centralization of the media. The use of the radio spectrum for WiFi appears to encourage decentralization. And one simply doesn’t know what will or won’t come down the pike next.
I’m still slogging my way through Ray Kurzweil’s techno-utopian book and I plan to offer some more specific criticisms of his argument once I finish it. But I think Matt’s general point here is enough cast doubt on the whole futurist enterprise. I remember reading in grade school that we’d all soon be communicating with video phones. It didn’t happen, not because the technology to build video phones wasn’t feasible, but because it turned out there wasn’t a whole lot of demand for video-phone services. Indeed, the latest instant messenger applications have video-phone service, but most people don’t seem to be using them. We now have the capability for wide-scale, dirt cheap video phone services, but it turns out that most people just aren’t that excited about the idea.
On the other hand, many of today’s popular uses of the Internet were largely unanticipated in 1996 (to say nothing of 1986). Blogs, social networking sites, self-publishing sites like YouTube and Flickr, and social filtering sites like digg all had their antecedents in the 1990s, but I doubt anyone foresaw their rise to prominance. Indeed, the .com bubble was largely the result of people attempting a naive extrapolation of the trends of the 1994-1999 trend far into the future. Needless to say, their extrapolations turned out to be wrong.
By the same token, it’s a pretty sure bet that by 2016, we’ll have computers that are an order of magnitude more powerful than today’s machines, and they’ll be able to do a lot of cool stuff. But trying to predict the details would be foolish, because we have a lot of learning and experimentation to do in the interim. More importantly, the technologies of 2026 will be built with the knowledge gained from experimentation with the technologies of 2016. Since we don’t know exactly what those will be–to say nothing of how they’ll be used–it’s absolutely futile to try to predict in any detail what high-tech will look like in 2026. And confidently predicting what will happen in 2036 or 2046 (the timeframe in which Kurzweil predicts a “singularity”) is insane.
If future technologies were a linear extrapolation of current technologies, we wouldn’t need to wait until the future to build them. The reason progress doesn’t happen faster than it does is that we learn new things about the world with each new generation of technology. Opining about the state of technology in the 2020s is silly, because we won’t have the necessary knowledge until sometime in the 2010s.