[extropy-chat] Moving on (was CF research)
russell.wallace at gmail.com
Tue Feb 13 14:04:55 UTC 2007
On 2/13/07, Samantha Atkins <sjatkins at mac.com> wrote:
> Building something generally enough to have a sufficient market for a
> decent ROI and yet sufficiently intuitive, efficient and useful to
> precisely the project you are working on now without taking too long
> out of the schedule to learn the tool is a real trick. How many
> potential users willing to pay how much for what with what kind of
> long term maintenance, support and enhancement costs? It might be
> better to do it with a consortium of interested parties under open
> source. More of a Cathedral than a Bazaar though.
I've been thinking along similar lines, and my reasoning is as follows:
Suppose you tell the program "sketch a model of a teapot". For this to work,
it needs to know what a teapot actually is.
The sum and total of the required knowledge base will be far beyond the
resources of a single team to assemble. It'll need to be shared; in rough
analogy to Google and Wikipedia, the system's effective smarts, its ability
to solve problems, will be a function of what it knows by virtue of the
input of a great many minds distributed around the world.
As a matter of technical principle, a free/shared knowledge base isn't
incompatible with a proprietary program, but as a matter of practical fact,
it's far more likely to work if the program is also open source.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the extropy-chat