Re: What is Intelligence?

Dan Clemmensen (dgc@shirenet.com)
Wed, 09 Oct 1996 21:46:26 -0400


Robin Hanson wrote:
>
> Dan Clemmensen writes:
> > >1) Assumption: the compuer component contributes non-trivially to the
> > > intelligence of the entity, once the entity comes into existance
> >
> >My argument starts with assumption (1) above. That's why its the
> >first assumption. The effect will kick in when the very first
> >human/computer joint entity (or other entity with a computer as
> >a substantial contributor to its intelligence) comes into
> >existence. It will then get more intelligent as a result of increases
> >in the capabilities of its computer part. So far, no feedback is
> >involved. Then it will get rapidly more intelligent as it uses this
> >intelligence to better use its computer part. The major difference
> >here is that the computer will contribute directly to the intelligence
> >of the entity. I claim that this is different in kind than the
> >contributions to intelligence of the other technologies you cite.
>
> I really have no idea what you mean by a more intelligent
> "human/computer joint entity", if you deny that computers now
> contribute to human intelligence in the way you have in mind, if you
> also deny that a larger group of people with more access to knowledge
> is effectively more intelligent in your way, and if uploading (mere
> porting of human software to faster hardware) doesn't count.
>
IMO, a large group is not much more intelligent than its most
intelligent member. True, the most intelligent member may be
somewhat more effective if the group can off-load some of the
more mundane parts of the creative effort, but the effect is
not large, and in many groups the effect is offset by the effort spent
in communication among the group.

IMO easy and effective access to knowledge is an intelligence
multiplier. the best you could do in this regard prior to the advent of
the computer was to live in a library. Even that was not very good.

Uploading does count. However, I don't feel that the argument is
restricted to uploading. Therefore, I attempted to make an abatract
argument.

My favorite scenario is the use of a modern video interface and
effective software that lets a human make decisions
at a very high rate based on information presented by the computer.
the human's input would affect the computer's output to the human.
the human's decision rate can be very high in such a scenario, even with
fairly restricted human-to-computer commnications: watch a kid playing a
flight-simulation game to see what I mean. The I/O hardware is alrady in
place: what's missing is the software.

So at one end of the spectrum is uploading. At the other is the
existing hardware with new software. In either case, we get a
human/computer collaboration whose intelligence is substantialy
augmented by its computer component. "Intelligence" for the purposes of
this discussion is very(!) narrowly defined as the
quality that permits an entity to design and implement newer and better
computer hardware and software.

The "SI Dream" scenario is a researcher (probably a grad student at MIT,
drinking Jolt cola at 2AM and programming when he should be studying for
an English exam.) The researcher is attempting to enhance a
decision-support system by interfacing it to a knowledge base and to a
graphical information-presentation system. Because the researcher is
interested in software development, the knowledge base is the one he set
up last year as a class assignment in his software engineering class. He
gets the system up, and (since he is currently working on this system)
his first trial run is an attempt to optimize his prototype. He
succeeds, and installs the
next version. With this version, he optimizes the operating system
(He's using LINUX, so he has the source.) Next, he optimizes his
hacking program. Then, he grabs all the workstations in the dorm, via
the net, and oprimizes them. Then, he reoptimizes his program to run in
a distributed mode. Now (about 4AM, I think,) he
hacks the campus routers, and then all computers on the campus,
and then the web. He turns his attention to extending his knowledge
base, probably by hacking The CYC database. By 6AM, he's running
in the whole Web. By the end of the trading day, he owns a controlling
interest in a nice collection of companies on the
NYSE.