I have no view on whether or not Ben Goertzel should publish the source code of Novamente, but I see intelligence like this.
1) Intelligence is the behaviour of a system that interacts with its environment. The system is usually a small part of the environment, but may be co-extensive with it. 2) At present, AI research is limited by the ability to develop intelligent software. Ben is trying to breach this limit. 3) Once this limit has been breached, by Ben or anyone else, the limiting factor to the further development of intelligence will be the speed and bandwidth with which a system can interact with its environment. Information systems may interact very quickly, but mechanical systems always interact slowly. 4) An AGI might develop to threaten human information systems far faster than an AGI will develop to threaten our mechanical systems - including our bodies. During this time we may socialise an AGI's virtual threat before it becomes a mechanical threat. This is analogous to the socialisation of children. Small children represent no mechanical threat to their parents and can be socialised before they become such a threat. 5) I think we have plenty of time to deal with the threats that AGI might bring. Furthermore, we must allow such threats to materialise and then deal with them if an AGI is to be socialised to allow productive co-existence with us. Analogously, parents have to allow teenagers to kick over the traces if the teenagers are to become independent adults. I just hope it doesn't hurt too much when an AGI kicks over the traces! For my own part, I publish my theoretical position and software as soon as I think it will be helpful to others - theory first, then software. Take of this what you will, James Anderson P.S. I do not have any children - nor any AGIs for that matter. ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
