Plan in Advance

Eliezer Yudkowsky (
Sat, 30 Nov 1996 18:22:17 -0600

<< Note: This letter was originally written in response to the Star
Trek fundraising suggestion, but, I was later informed, never made it to
the mailing list. I am therefore reposting it.>>

The bit about fundraising at Star-Trek conventions is impractical,
inadequate, but at least it's something. Still, I don't think we should
accept unearned gold at this point. Any serious push for the Singularity
should be funded by venture capital (if the venture has short-term
possibilities) or by a single rich veteran of the short-term pushes (if
it's a purely long-term thing or something too dangerous for venture
capital); anything else is just *asking* to become a cult.

Think of your job at this point as the laying of groundwork. At one
point, someone with major money will be stricken by a vision of the
Singularity, or someone with a vision of the Singularity will succeed in
building up funding. Maybe me. Maybe you. Maybe that Netscape guy,
Andersson or something. The point is, I'd like to see complete action
plans for all levels of funding between $10^5 and $10^10.

I'd like to see frequently updated, ready-to-use lists of all major
researchers involved in Singularity-oriented research, their current
salaries and funding situation, how much good some additional funding
could do and the potentialities for short-term profits.

I'd like to see lists of all the venture capital that would be
interested in funding Singularity-oriented ventures, with specialties,
generosity and control-freakness clearly indicated by business records
and interviews with former partners.

In other words, what I would like to see is a completely laid out plan
that can be polished in advance of anyone with the ability to implement
it. I don't want idealistic plans for a fan club that will someday rule
the Earth; I want hard figures. I want an analysis of the legal problems
faced by foundations, corporations, Internet-coordinated public
conspiracies, and any combination thereof.

Show the names of Extropian CEOs, CPAs and (shudder) lawyers who could
keep a corporation running. Put everything into place that can be put
into place with a little research, a little experience, some advice from
more experienced friends and access to the Net version of the Wall
Street Journal. As a start, you might want to try and form a group of
Extropian businessfolk who could form such a plan as a hobby. If there
are no Extropians who can run a corporation, that's a major roadblock
that we can start solving now. I don't care about MBAs, but if NONE of
us know how to do this, some of us better learn. I've read "Start Up" by
(I think) Jerry Kaplan, a story of a pen-computers company which was
killed by Microsoft, which seems like a nice place to start.

You could be hit either by a really bright idea or by a lot of money.
Have detailed enough preparations in place for both, and one or the
other will soon come along. You may not be rich, but there are some
things you can do, and one of them is planning.

This is not a call to arms; do not send *me* email volunteering for the
project. Someone else save this document and gather the Procedure
Preparation People. After Triple-P gets a basic plan set up, I can stomp
it to shreds looking for logical flaws and then spin the tattered
fragments into something that works, but I can't start Triple-P itself.
That's a project, and a project requires a plan. I can't handle long
linear causal sequences well enough to formulate one, for reasons
explained in "Algernon's Law", the second Web page in my signature.

Even if I get hit by a truck, though, it might still be a good idea to
have a plan in place. One, it will get you thinking along the right
paths, practical ones. Two, you may go through the planning procedure
and find at the end that your plan requires neither money nor a
brilliant idea, in which case you may proceed to actually implement it.
Three, all kinds of interesting facts pop up when you think practically;
we may wish to know about them. Four, if we don't think it out in
advance, we will screw up. Five, having a framework in place will ground
all the conversations around here and make them much more interesting,
in the same way that science is more interesting than philosophy. Six,
building the plan will give everyone involved experience working
together on an Extropian project; this may be important later.

What needs to be planned? Let's say I want to start a corporation
devoted to building AI interfaces to applications. It's
Singularity-oriented because it will be developing AI that can be used
to organize scientific knowledge, and because the founders will be using
their gains to fund other ventures. Part one. Where do I start? Look for
venture capital? Write a business plan? Build a prototype? If I look for
venture capital, who should I go to? How much venture capital can I
expect and how many programmers can I hire with it? What types of fame
and/or credentials will impress venture capitalists? Which is most
impressive: A doctorate, having published a book on the subject, five
years of experience, winning a prize within the field, or being
world-renowned for something? Are any of the above absolutely necessary?
This, you understand, to see if I need a figurehead and if so, what

Where can I get advice? Where can I get advice on whether my bright idea
is any good - without, of course, giving the bright idea away? There
must be ten thousand books on how to start your own business. 90% of
them will be worthless. Does anyone have a list of the top ten? No sense
in duplicating effort; some of the step-by-step guides will already be

Think practical, practical, practical. Let's all remember that we don't
care about what happens after the Singularity because IT'S NOT OUR
PROBLEM. Sure, I wrote an entire Web page about the Beyondness of the
Powers. But that was for the benefit of my fellow humans and was
intended to settle the issue by impressing all concerned on how *futile*
that kind of speculation was.

Also, I tried to include arguments that were not just impressive but
scientifically informative. How to set up a Power with numeric intuition
is a problem that we, unaided humans, may need to solve before the
Powers take over. Not exactly a practical problem for the near future,
but it's something about which thought may be useful, especially since
it might be adaptable to a modern, non-Power CAD station or something.

We aren't going to get to the Singularity by singing hosannas to Powers
that won't even exist unless we do something! Start planning!

--       Eliezer S. Yudkowsky

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