Re: [agi] Failure scenarios

[Nathan Cook]

Ben, I take it you're using the word hypergraph in the strict mathematical sense. What do you gain from a hypergraph over an ordinary graph, in terms of representability, say?

To return to the topic, didn't Minsky say that 'the trick is that there is no trick'? I doubt there's any single point of failure in a complex enough system - if it does enough different things that correspond to human intelligence it's very doubtful they'll fail to gel, in my opinion. When you come to a problem that you can't solve, you don't just stare at it, you try another strategy.

People have been working on AIs that can model and  other agents' internal states and predict their behaviour, on vast KBs, on various inductive strategies for learning and problem solving, but they're all considered cutting edge in themselves - it would be seen as gauche to start a mega-project combining all of them at this juncture, so you're left with the people who hope it will fall out of a nice tidy initial design with the theoretical capacity to learn how to do all those things. Even AM/EURISKO was thought to have a small probability of doing that.

---

This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/[EMAIL PROTECTED]