Re: PHIL/PSYCH: Algernon's Law vs. Einstein's Experience

Eliezer Yudkowsky (
Sun, 12 Jan 1997 21:31:20 -0600

> I was struck by this following passage from Albert Einstein which
> seems somewhat relevant both to Eliezer's Algernon's Law idea and to
> the recent debate about whether we already have all the fundamental
> concepts of science down and the rest is *just engineering*.

Sigh. While I hate to slap down any post that cites my own work, duty
calls and I must follow.

Just the headers in this post are bad enough: "PHIL/PSYCH" and
"Algernon's Law" do not go together. Algernon's Law is for "NEURO",
"COG" and ">H" posts only; even "PSYCH" or "AI" are iffy; "PHIL" is

Algernon's Law doesn't refer to just any tradeoff made anywhere.
Algernon's Law refers only to an evolutionarily predesigned cognitive
mechanism altered by a single mutation and/or external influence, and
deals with the tradeoffs that then occur, stating, essentially, that the
result will be a net evolutionary disadvantage in the ancestral
environment "unless otherwise specified". TANSTAFFL, in a phrase.

This gradual broadening seems to occur within a lot of memes; the
"extropy" meme (not "extropy meme") is a fine example. "Everything good
in the world" seems to sum up many definitions; I've even seen it used
completely inappropriately in thermodynamic discussions, and somewhat
less inappropriately - but still wrongly - in discussions of evolution.
"Extropian" refers to a bunch of people, more rocket scientists than
cognitive scientists, who want to take over the Universe, improve their
bodies, and *occasionally* you will run into an Extropian who wants to
improve their minds.

"Extropian" thus does not refer to Eliezer, who doesn't give a damn
about a conglomeration of stars that could be simulated on his Power Mac
with a couple of thousand lines of code, and prefers to explore the far
more complex Universe inside the brain. This can lead to occasional
conflicts; I hight the recent "Singularity-worship" fiasco where both
sides thought the other was wasting time.

My timeline looked like this:
Cognitive Science -> Singularity -> Universe Domination

Your timeline looked like this:
Nanotechnology -> Universe Domination

And your image of my timeline looked like this:
Singularity -> Universe Domination

The result was that you accused me of "lying back and waiting for the
Singularity to take care of everything." What I was actually trying to
say was: "Who the hell cares about what you'll do after you're
uploaded; you can figure that out *after* you're uploaded; now can we
get back to discussing neural plasticity so we can get to the
Singularity and be done with it?"

"Algernon's Law" was created to be a technical term. It was presented
on a 21-page Web file full of cognitive science, albeit popularized
cognitive science. It's a technical term and I intend to keep it that
way, because I've yet to see any good come from broadening any term but

Einstein only starting to think about space and time as an adult is not
an instance of Algernon's Law. If Einstein were dyslexic, that would be
fascinating. If you demonstrated that children who don't visualize
about "space and time" have an easier time visualizing "space-time" when
they get older - and note that I say *visualize*, as in represent via a
basic cognitive module, and not *think about* - then that might fall
under the Algernonic aegis. Starting over fresh isn't even faintly
Algernonic unless you're talking about a cognitive module that keeps
getting wiped, or something similar.

--       Eliezer S. Yudkowsky

Disclaimer:  Unless otherwise specified, I'm not telling you
everything I think I know.