From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Wed Mar 28 2001 - 04:21:09 MST
Eugene hasn't turned strange, Eugene has been strange for a looooong time.
On Tue, 27 Mar 2001, Eliezer S. Yudkowsky wrote:
> Eugene, are you feeling all right? This post would have made sense two
So la la. Could be better, but thanks for asking.
> years ago. Insulting, and condescending, but it would have made sense. I
Insulting, yes. I'm a bastard that way. Condescending, no. I think you're
making a mistake, and need some advice. You might heed it or not, but I'll
say my piece one more time.
> should now point out two things:
>
> 1) I already have the perfect job. I would trade salaries with Bill
Good, the qustions is: is this the perfect job on the long run, if you
want to produce results? There's a glut of terrific projects out there to
throw yourself onto, but you might at a later stage discover you've landed
in a cul de sac, which is terribly hard to backtrack. You need to work
with other people, and these people have expectations, resulting in a
sequence of windows of opportunity. This is for real. Plus, there's mental
hardware deterioration (which we can't fix yet, and may not be able to for
a long period of time). You *can* still get results at a later age, but
you'll have to labour a lot harder than while you're young.
> Gates, but not jobs. I continue to work on the design, and when SIAI gets
> the funding to hire another four or five programmers we can start coding.
I've learned that much from Brian. Interesting project, good luck.
> Why on Earth would I want to drop that so I can go to a university and not
> start coding until I'm 25?
You might discover that you can't code the fishes from the deep or the
Moon from the sky all on your own, however easy it may appear initially.
You will need to refer to work done by others, which you need to be aware
of. It will be worth of your time to study prior art, such as what happens
inside of our heads, at all levels of abstraction. You need to understand
the architecture of your target platform, as your task will involve
dealing with all layers, including hardware design. You'll need to
understand supercomputing, and parallelism. And whatever else might come
handy, which will be plenty.
You don't learn such things by learning to code like a demon before you're
25. You'll get plenty of time to code before your're 25, just don't do it
exclusively. Read the journals and literature for at least a couple of
hours, but daily. This will be worth much more than coding (you can always
hire a good coder if your project is important enough for a lavish
budget, and eventually this will become a very large programme with a
rather intimidating budget).
> 2) I'm 21 and already past your deadline. Are you experiencing a
You're past my deadline, that was my point. You're starting to lag,
though not badly yet.
> flashback to 1998 or something?
Oh, that would be great, since giving me a couple extra years.
> Eugene, you now know that I'm not going to go to a university. Period. I
No, I did not knew that. Now I know, and I think it's a mistake you may
later come to regret. Making stuff that thinks is not at all like a
carreer in plumbing, where you can learn it on the job. There's value in
doing it madwand way at first, but sooner or later you'll need to check
out on all the stuff you've missed when discovering on your own. Unless
it's a real new field (say, one you've just invented), you'll discover
that surprising numbers of people have been down similiar pathways before,
and some of their stuff even works.
> will notify you if that changes. If you make further posts along these
> lines, I will consider it not friendly advice, but a snide attack on my
> reputation, and respond accordingly.
Do what you want. I think I owe you some advice, because I had none when I
was your age. You need to listen to other people (of course being
selective about it) if you want to go places.
> I'm serious about this, Eugene. We've discussed this before, and you've
> been warned before. I tolerated this when I was eighteen because it could
> have been well-meant. The plausibility of that hypothesis is now
> dwindling towards zero. You've been one of my steadfast (albeit
> repetitive) critics where AI is concerned, and I nearly respect that, but
> I can't respect this. I don't know why you would suddenly decide your
> goal in life is to trash my reputation, and it would be paranoid of me to
> attribute to malice what I can attribute to well-meant condescension, but
> it would be equally careless of me not to consider that I may be under
> deliberate attack. If you continue making these remarks, I will assume
> that they are entirely malicious in nature.
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:06:44 MST