Re: We need better tools to make better tools.
Sun, 1 Dec 1996 21:23:31 -5

> From: "Max M" <>
> I don't see why we need AI at all to reach the singularity!!!
> My idea for reaching the Singularity is then: If we keep building better
> tools for humans to use to make better tools to make better tools .... you
> know. Recursive stuff. Then we would finally have a software tool where all
> we have to input is new ideas and a vague notion about how it is done. The
> tool could then do the rest or we could experiment until we had it right.
> Does this sound like nonsens?

Not at all. It is, in fact, the direction that things are already
going. Code generators and high level languages do what you describe
to a limited extent right now. Modern operating systems handle more
and more of the details of utilising specific hardware so that
applications developers don't have to know all the details of every
printer, video card, communications device, etc. There is enourmous
room for improvement, yet there is also considerable progress
already. Each generation of tools helps to create the next. The same
phenomenon is already affecting the development of better hardware as
well. Better hardware provides the resources that are needed by more
sophisticated software. More sophisticated EDA software then
contributes to the development of better hardware. Languages, code
generators, EDA and CAD tools, better hardware, better manufacturing
technology. It all interacts to create better & better computers &
software at an accelerating rate. The magic moment when technology
begins to contribute to it's own development is not still waiting for
us in the future. It began years ago with the creation of the first
assemblers and crude debugging tools. Or maybe it was the creation
of the first electronic test instruments. Or mass production. Or
maybe it was the invention of the printing press. Or agriculture. Or
the first primitive human who discovered that he could use a rock as
a tool to create a blade, an even better tool. Take your pick. The
point is that technology has been contributing to its own development
for a long time. I think that the fatal assumption made by
"singularity" proponents is that the present rate of development is
not already feeding itself. It should be clear that this is not the
case. After all, a human with a circa 1996 engineering library, a
modern cad system, and a connection to the Internet is a better
designer than an otherwise equivalent human with the tools that were
prevalent in 1976. Then they created TRS 80's. Now look at what is
being created. Our intelligence is not unaided even now.

Lest someone think that I am arguing that we are not on the verge of
a very significant turning point, I must state that this is not the
case. I think that singularity proponents are correct in their
assessment that we are approaching a point in history in which things
will change in very profound and fundamental ways. But the
enthusiasm that this recognition generates can color perceptions and
cause some assumptions to go unquestioned because we really like
their implications.


Peace, William Kitchen