Re: Anti-singularity spam.

From: Eliezer S. Yudkowsky (
Date: Tue May 02 2006 - 14:45:23 MDT

Bob Seidensticker wrote:
> Damien: You raise a good point. Specific predictions are often wrong
> ("We'll be using beams of light to send telephone signals wirelessly by
> 1990"), but the general idea (we'll have wireless telephony by the late 20th
> century) is often correct. Eliezer's observation that "Careful futurism is
> rarely popular, popular futurism is rarely careful" is excellent.
> My concern: where does this stop? Do you give a pass to the 1930s
> predictions of fleets of zeppelins because we *did* get fleets of airplanes?
> Or of 1950s predictions of huge nuclear-powered oil tanker submarines
> because we have oil tanker ships? If your criteria for success is wide
> enough, lots of things will pass. Perhaps the question is: if you take what
> eventually happened and show it to the original predictors, would they be
> satisfied or surprised? I simply want to hold futurists accountable rather
> than bending over backwards and saying that they were right in spirit even
> if laughably off in fact.

Hold futurists accountable? Sounds wonderful. *Which futurists*?

Futurism is commonly practiced as a form of entertainment, pure and
simple. It's storytelling about distant land called Future and the
bizarre customs of its inhabitants, just as one might tell stories of Oz
or Atlantis. The story obeys the laws of dramatic unity, and has plenty
of concrete detail, in accordance with the advice you will find in any
book on fiction writing. Like religion, popular futurism is "minimally
counterintuitive" in the sense that it violates one or two expectations,
just enough to be startling and memorable, while obeying all other

Of course they get it wrong. Why in the world would you expect popular
futurism to be right? It is like making a list of failed predictions
from the Lensman novels. Popular futurism has no connection to reality.
  It has rules, but only the rules of the storyteller. The dates
consist of people pulling random numbers out of their butt to support
the demands of the story. Futurism as commonly practiced is part of the
media-entertainment complex, not science. The fact that few people
bother to keep tabs on the accuracy of past predictions tells you this
instantly - imagine scientists publishing theories and no one bothering
to check on the experimental results.

I realize this point is not widely appreciated, but to draw an accurate
map of a city, you must walk through the city and make lines on paper
that correspond to what you see. Sitting in your apartment drawing
fascinating worlds isn't going to work. If you make stuff up at random
and it turns out not to correspond to reality, that is not very
surprising once you think about it.

There are people in this world who learn other rules than the rules that
storytellers know. (For an introduction to these rules, I recommend
Robyn Dawes's "Rational Choice in an Uncertain World". He has some
choice words to say about overestimating predictability.) There are
people who focus as much on common fallacies and reproducible biases as
storytellers focus on vivid details. When a careful rationalist dares
to say something about the future - and the daring part is not the
amazing startling prediction, but claiming that whatever-it-is is
predictable in the first place - then that is discussed and criticized
under different rules.

If someone makes a startling prediction in a newspaper, and there's no
carefully developed supporting argument for why whatever-it-is is even
predictable in the first place - then why is it *interesting* that
they're wrong? What else would you expect? Next you'll be telling us
that astrology doesn't work at better than the chance level.

Eliezer S. Yudkowsky                
Research Fellow, Singularity Institute for Artificial Intelligence

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT