A good step would be looking at Lenat’s published work around that time to get a feel for how he thought about AI and what strategies he was likely to implement. Lenat was a firm believer in symbolics meaning he worked almost entirely in Lisp. Another thing is that his systems were interactive—they relied to a large degree on user input, possibly in ‘selecting’ the right strategies which the machine would then investigate in further depth and possibly try to optimize. For an example of this, rule 2 of the AM system is, “If the user has recently referred to X, then boost the priority of any tasks involving X.”
Lenat himself said that the Traveller TCS solution was 60% himself and 40% Eurisko, and given that he was trying to promote his software, this is probably an overestimate. It’s likely that Lenat won Traveller TCS, not Eurisko, and that Eurisko was simply a tool he used to carry out various simulations and computations that would have been tedious to do by hand. That doesn’t mean that the system is useless though. I’m personally extremely interested in interactive AI systems that use the ‘best of both worlds’, relying on a combination of human intuition and raw machine computational power to perform tasks that neither could have done separately.
A good step would be looking at Lenat’s published work around that time to get a feel for how he thought about AI and what strategies he was likely to implement. Lenat was a firm believer in symbolics meaning he worked almost entirely in Lisp. Another thing is that his systems were interactive—they relied to a large degree on user input, possibly in ‘selecting’ the right strategies which the machine would then investigate in further depth and possibly try to optimize. For an example of this, rule 2 of the AM system is, “If the user has recently referred to X, then boost the priority of any tasks involving X.”
Lenat himself said that the Traveller TCS solution was 60% himself and 40% Eurisko, and given that he was trying to promote his software, this is probably an overestimate. It’s likely that Lenat won Traveller TCS, not Eurisko, and that Eurisko was simply a tool he used to carry out various simulations and computations that would have been tedious to do by hand. That doesn’t mean that the system is useless though. I’m personally extremely interested in interactive AI systems that use the ‘best of both worlds’, relying on a combination of human intuition and raw machine computational power to perform tasks that neither could have done separately.