Re: Event horizon

From: Eliezer S. Yudkowsky (
Date: Wed Jan 24 2001 - 09:12:47 MST

I'm kind of reluctant to get into this at all, since (a) it's one of the
basics (outside our charter) and (b) debates about it perennially pop up
in the Extropian mailing list, and they never do any good there. There
really should be a Singularity FAQ one of these days, and this really
should be one of the questions in it.

As I understand the "perceptions accelerate in time to keep up with the
Singularity, so there's never an event horizon" concept, the idea seems to
be that in the year 2000, you can see to neurohacking-based intelligence
enhancement in 2010, and to human-machine interfaces in 2020, but you
can't see beyond 2020, so that's the Event Horizon. And then, in 2010,
the neurohackers can see human-machine interfaces in 2020 and uploading in
2030, but they can't imagine what happens after that, so the Event Horizon
has moved ahead ten years in ten years.

It's a pretty little picture, but totally false-to-fact for two major
reasons. One, if neurohacks show up in 2010, then you can't see beyond
2010 because you can't take into account the effect of neurohacks.
Singularity::Event Horizon should not be confused with
Singularity::Transcendence; the Event Horizon shows up in the *first
instant* that enhanced intelligence starts altering the picture,
regardless of how long after that it takes for all the progress indicators
to warp to OVERLOAD.

Two, the "moving Event Horizon" theory is one of those slow-takeoff
scenarios that futurists just can't seem to get away from, where
everything happens neatly spaced-out over the course of human decades.
The heart and soul of the Singularity is the change that comes with
increased *quality* of thinking - but increased speed of thinking also
plays an interesting part, from a practical perspective, anyway. The
human timescale of 200hz neurons is not the only one in the Universe.

What actually happens is that genetic engineering and neurohacking and
human-computer interfaces don't show up, because they'd show up in 2020,
and a hard takeoff occurs in SIAI's or Webmind's basement sometime in the
next ten years or so. Even if the hardware for nanotechnology takes
another couple of weeks to manufacture, and even if you're asking the
newborn SI questions that whole time, no amount of explanation is going to
be equivalent to the real thing. There still comes a point when the SI
says that the Introdus wavefront is on the way, and you sit there waiting
for the totally unknowable future to hit you in the next five seconds.

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

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:35 MDT