Thanks for taking the time to dig this out, appreciated. I've been reviewing it this morning. The formulas presented in the "measuring success part" so far, though interesting, seem to be arbitrary. For example the question of whether an agent should research for efficiency or pick the low hanging fruits for short time benefit is answered through a simple sum formula. Another example is when the author(s) state that universal intelligence should favor simpler choices and interact with the environment to cause less complexity. Well that's obvious! Just use a binary inverse logarithmic distributive operator. With the comical response to criticisms part (starting from 5.2) I feel like I'm in a Douglas Adams movie.
reply