I think core CS needs to be brought back into the picture. This Omega character is posited to be able to determine a Rice property about rational agents. As the

I think core CS needs to be brought back into the picture. This Omega character is posited to be able to determine a Rice property about rational agents. As these agents are presumably Turing-equivalent or BETTER, this supposition contradicts our established theory. (for people with less formal CS education: Omega can solve the halting problem) The author's other article on Newcomb's problem (http://lesswrong.com/lw/nc/newcombs_problem_and_regret_of_rationality/ http://lesswrong.com/lw/nc/newcombs_problem_and_regret_of_ra...) hints at this but doesn't get to it explicitly.