> Go/Weiqi is also a computation problem given this argument.
I agree with the assessment that it is a computation problem.
As a go player for 25 years, I never thought that I would see a computer consistently beat a strong amateur in my lifetime (with many decades still theoretically left to live). Not only did I see it much sooner than I expected, but AI beats the best pros at an almost 100% win rate.
Within the Go community, there is consternation that a computer can make some moves that humans probably shouldn’t make due to the ability of the computer to follow up properly where humans cannot (yet).
This may change over time, but right now the computation part of the go world is much closer to expressing perfect play in a complete information game than humans are.
> What precludes asymmetric info games like poker from being "computation" problems, if one takes a wide enough view (to include the unknowns as the modeled state)?
Also as an avid poker player, I think many aspects of poker are rapidly leaning this direction, even though poker does not have complete information.
I’m not sure how much you know about the current poker scene, but GTO solvers have opened the poker world’s collective mind about what “good” play looks like, and the players are riding the tails of the computers in terms of strategic and tactical evolution.
In some formats, like heads up limit hold em, the game has been effectively solved.
Bringing it back to the original topic, I would love to see a GTO version of life. Even if it is hard to implement, it could provide a guide that informs decision making.
I personally don’t think I will see it in my lifetime. That said, like Go AI, I hope I’m wrong.