From: Jack Richardson (jrichard@empire.net)
Date: Sun Aug 05 2001 - 14:21:19 MDT
In reading Eli's recent email, it sounds as if he has developed a clearer idea of what the next steps would be to create a Seed AI. Flare is being developed to provide the specific functionality that will optimally support the self-enhancing processing of the AI.
Now, in my experience where new technology is being proposed to achieve certain goals, it was always necessary to provide a 'proof of concept' to convince the powers that be to keep the funds flowing. A specific situation that would demonstrate a workable solution had to be defined in such a way that everyone would be satisfied that the general approach was on the right track.
I'm interested to know if such a 'proof of concept' has been defined using Flare as the programming tool. It would seem to me that this group would be an ideal mechanism to propose the specific situations that would provide a good test of the tool and the methods.
The proof situation should be narrow enough so that it would be doable in a short period of time--like a year or so. But it also ought to be good enough so that it was sufficiently convincing to all of us. It also ought to be able to connect up to the next functional component.
This is a key way to be able to tell if a team is still on the vision stage or if it is far enough along in understanding the task to be able to specify in reasonably precise detail its proof of concept.
Any suggestions on what it would take to make you sit up and take notice?
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT