Re: How hard a Singularity?

From: Eugen Leitl (
Date: Sun Jun 23 2002 - 05:51:17 MDT

On Sat, 22 Jun 2002, James Higgins wrote:

> I don't think the Singularity could, even possibly, be slowed down to
> take a decade. If for no other reason that some other group (probably

Singularity is contiguous. If we can put a hard limit on it, is where we
hit a/the plateau. Barring new physics and unconventional uses of known
physics there is enough matter in the local system to last 10^2..10^3
years to the plateau state.

I can imagine a lot of Singularity flavours that take several decades, or
longer. This is plenty in terms of subjective monkey time.

> several, actually) will also be working on it and at least one of them
> will go for the hard-takeoff option (or will botch the slowdown).

It depends on the threshold and the payoff factors. I don't expect a
desktop box suddenly exploding into a growth of nanofilaments overnight,
devouring the world, and then going up with taking a snack on the cosmos.

> Somewhat slower, however, should be strongly considered. The more we
> (humanity) can keep and eye on the progression and guide it the
> better.

> It is, unfortunately, most likely that we won't be able to greatly
> influence the shape of the Singularity. Scares the heck out of me,

Total shape, no. Outcome, no. But early stages, possibly. Seen from the
mountaintops, whether we monkeys make it or go extinct is completely
irrelevant. It is not going to make an observable difference in the end
state. However, we monkeys do see it differently. It is completely in our
paws to make a difference.

> but I still don't see it otherwise. Unless we could somehow manage to
> continually upgrade a handful of HIGHLY TRUSTED humans to at least
> somewhat keep pace with the Singularity I don't see how its possible.
> And I haven't heard of any real options for doing that...

Life is uncertain. It doesn't mean we're doomed/shouldn't try.
> Well, Eliezer has a problem with people dying. The longer your wait

Many people have problems with people dying. Though not nearly enough,
orelse we would start doing something about it some 30-40 years ago.

> to get the Singularity in full force the more people die (that's my
> interpretation of his vision, at least). Personally, I'd rather let a
> few hundred thousand people die while making certain that the
> Singularity won't just wipe everyone out. I mean, what's the point in

I agree with this assessment.

> rushing to save lives if everyone gets converted into computronium
> anyway? My best guess is that the final result will be somewhere in

Everyone but those who have died before the technology came online. You
certainly can't resurrect a person if given an urn full of cremation
ashes, or even mummified remains encased in permafrost.

> between (but probably closer to Eliezer's wishes than I'd like).

I'm expecting less control, not more of it. Lots more diversity, and a
much grander stage for the play called life.
> Here I'm split. Personally, I think if we try and wait that long
> we're making a major mistake. There are STILL people who aren't
> comfortable with computers today. To try and get humanity "on-board"

There are still people who'll resist receiving modern medical treatment,
but I'm noticing there are not too many of them. It requires remarkable
testicular fortitude to remain true to one's convictions if confronted
with one's immediate personal demise.

> the Singularity would take centuries, which we don't have. The most

I think most people will be able to make an informed choice between being
left at the mercy of uncaring gods, or becoming a part of the pantheon.
Especially, if there are members of the tribe who went ahead, and are
obviously liking it in there.

> moral (in my opinion) option is to create the Singularity as fast as
> possible while maintaining a strong measure of safety (ie:
> non-extinction, slavery, indifference, etc). And yes, I know actual

Fair enough.

> slavery would be pointless (we would be useless junk to an SI). But

No, you're a static part of the landscape to an SI. Food, in other words.

> we could be "slaves" by loosing much of our freedom & choice, which is
> too likely. But, as mentioned earlier, I don't think we'll actually
> get this option (due to the Eliezers out there).

We don't have that many Eliezers out there. In fact I would think there
are about some 10 of them world-wide right now.
> While I doubt we'll get the chance for a moderate takeoff, we should be ALL
> MEANS try. It is never pointless to try and protect the entire human race!


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:39 MDT