Friday, March 04, 2011

Making better predictions

I was sure iPad 2 wouldn't need a computer companion.

...  the 2010 iPad is more than $500 - but by 2011 the device will sell for under $500 with 3G-equivalent capabilities. An additional $15 a month will provide basic VOIP phone services ....

... By 2011 the combination of a $400 iPad (and iTouch for less) and $15/month VOIP access will start to replace a number of devices that are costly to own and acquire, while providing basic net services ...

Wrong. [1]

I figured no moving part netbooks would be selling for under $100 by 2011, finally following the trajectory of calculators [2].

Wrong again.

Maybe I shouldn't take this so hard. After all, even the pros have a hard time predicting the near future. On the other hand, honesty compels me to review how a few of my past non-obvious expectations turned out ...

1970s [4]

  • Eight track tape would go away. [3]
  • Information technology would lead to mass middle class unemployment. Maybe, but it's taking a long time.


  • Email would replace the fax machine. It's still not dead yet. I think I will die before the #$@#$ fax.
  • The net would kill the post office. True, it's dying now. But it won't die completely for decades. See fax.
  • CD ROM would revolutionize worldwide knowledge access. Well, maybe it would have. It did cut the cost of sharing knowledge dramatically. Except a few years later we had the net ...


  • Palm type devices would show up in cereal boxes. Again, I was fooled by the calculator. I mean, these things were cheap to make...
  • IBM's OS/2 would crush Microsoft's crummy Windows 3. Pathetic. I liked GeoWorks too.
  • The web would destroy most universities. Instead tuition skyrocketed. How wrong can you be?
  • American health care would collapse within a decade. It survived long enough to be saved by ObamaCare.
  • Modems would be gone by 2000, fiber to the desktop. We really believed that. Vast businesses were based on this premise. It's 2011, and there are still a few modems around.
  • Phone calls would be too cheap to meter. My voice services still cost a fortune.
  • Free WiFi would be city wide everywhere: Technology issues and business issues killed this one.
  • Credit card security failures would force industry reform: Twenty years later credit card fraud is institutionalized.



  • China's bubble is going to burst before 2012. Given my track record, this is probably wrong.

That's a pretty long list for a few minutes of thought. I hope I'm mostly remembering when I was wrong, and of course I'm not including obviously correct predictions like "Gopher will change the world forever" [5].

I'm not going to give up predicting of course. That would be boring. Going forward though, I'll try to keep these lessons in mind ...

  1. Choose your models carefully. No technology has seen the price crash of the calculator [6]. Several of my mistakes came because of my early experience with calculators; that was probably a major anomaly.
  2. In period of rapid innovation "winners" (CD ROM, Gopher) can have very short lifespans. The shape of the winner is more predictable than the details.
  3. Big, integrated enterprises take a very long time to die. Years ago I called this Canopy Economics.
  4. We are embedded in a complex adaptive system with gobs of inertia. Our world has a strong tendency to return to its historic trend; and to delay big disruptions by hook or by crook.

I'm not bad, I think, at predicting the future. I'm just bad at predicting when the future will arrive ...

- fn -

[1] Charlie Stross says I was almost right, but the death of a talented data center architect set back Apple's MobileMe plans. It's also true that 2011 isn't over yet. But I'll take the hit anyway.
[2] My first four function calculator was a lot bigger than a modern laptop, required a plug, and cost more than $400 in today's money [8]. I am old. Before that I had a slide rule. Really old.

[3] Ok, so that was a gimme. I include it only because I wanted an example to show that "tools never die" ain't true.
[4] I'm sure I had a lot more, but that was a long time ago. 
[5] Gopher was revolutionary. It would have changed the world, but a few years later we had Mosaic ...
[6] Perhaps because IP enforcement was weaker then?
[7] A post 9/11 meme of mine. My premise was that technological progress was making it possible for small organizations to purchase large amounts of havoc. Bioweapons, dirty bombs and so on. The cost of anonymous attack was falling quickly, but the cost of defense was falling more slowly.
[8] I originally wrote $200. I later read that a 1974 $1 was the equivalent of about $4.30 in 2010. So I adjusted the number.

See also


Anonymous said...

I think you have been pretty spot-on when you compare equal quantities. But the problem with predicting that tablets or laptops as a whole would decrease in price is that they keep adding components to them even as those components drop in price.

RAM per-unit prices drop by half and they stuff double the RAM into the laptop. Same with flash memory or a hard drive or CPU. They could only go so far in this direction with calculators by adding new functions before it stops propping up the market.

It is somewhat similar with voice communication. We decided we wanted mobility over price. You can get unlimited calls for a low flat rate from Skype as long as you are willing to tether yourself to your computer. And of course, the existence of a cell phone service cartel doesn't push prices down.

Your China bubble prediction still seems very plausible to me.

Zol said...

As far as the death of modems. I thought I heard that Egyptians dusted them off during their protests when the internet was shut down. They used long distance dialup services.

JGF said...

Calculator price drops were such an anomaly. In the first version of this page I actually understated the price drop. It today's dollars they went from about $400 (first one I bought) to free in about five-six years (marketing give aways).

I've never read an explanation of why their price cratered. It's true that it became hard to compete on features; for most users they were basically complete. It's also true that they didn't seem to have much patent protection.

Unknown said...

Au contraire, you're not wrong at all. All these things are very likely to happen ... eventually. (The one exception being the death of the fax. Technology never dies - people still ride horses and make pots on potters' wheels. Sorry.)

Your fundamental problem is that making predictions is a bit like Heisenberg's uncertainty principle - you can know the "what," the "when" and the "how," but probably not all three at the same time.

Moore's Law, for instance, let's you make surprisingly accurate predictions about when chips will reach x calculations per sec, but it would tell you nothing about what could be done with that computing power. Vannevar Bush could describe a memory storage and retrieval device with great insight, but he was off on how it would work (microfilm!) and didn't even attempt to predict a date. George Orwell was spot-on in his description of how a post-9/11 world would work, but he was off on when it would happen (1984).

Which is why predictions based on infinite extrapolations, like the technological singularity theory, are necessarily absurd.