Thomas replied to me: > For my understanding, that rule about limiting the skill of a (robot) program > to the skill of the programmer should be ignored if we talk about > "commercial" software as for your battle suit. Of course the rule make sense > if you have a single coder doing the software, but for a professional team of > developers (perhaps even supported by a proper KI …) it should be possible to > create something that increases the skill(s) of all single members of the > team.
Hello Thomas, in my experience, software projects fail if none of the 'techies' understands what the customer wants. It doesn't have to be the best programmer in team, and it could be it is actually someone from the customer side who happens to 'think like a techie' and who translates the requirements, but some sort of cap based on subject matter experts makes sense. > In addition, a neural net is a "self-learning" piece of soft- and/or > hardware. As I understood "accumulated skill" in a character, this is > proportional to time (and effort) spent on study and training. As a "silicon > based mind" could learn faster and memorise better than a biological one (and > is easier to reproduce …) I would say that such a (robot) program could > outrun the skill level of its original creator by multitudes. There is the -- possibly apocryphal -- story of the team which tried to teach a neural net to tell pictures of a forest with tanks from pictures of a forest without tanks. The neural net got very good at telling pictures from one day (the day when tanks were on the range) from pictures of the other days. Say you have an AI trained to an impressive skill level and you want it to self-learn to become better. Like any other character in that situation, it would be learning without a teacher, without books, on the job. The GM should ask the players just what that job is ... Regards, Onno _______________________________________________ GurpsNet-L mailing list <[email protected]> http://mail.sjgames.com/mailman/listinfo/gurpsnet-l
