This result doesn't seem all that impressive. In the end it's just an example of the advantage of being able to pre-commit to a strategy. The linear algebra parts are entirely unnecessary (although interesting).
For example, suppose we're playing the iterated prisoner's dilemma. I pre-commit to a strategy: if you've defected in the last 5 games I will always defect, otherwise I will defect unless I haven't cooperated in the last 4 games. In this situation your optimal strategy is to always cooperate, even though I'm taking advantage of you. Punishing me won't change my strategy, I've pre-committed to it. The rational strategy, and the one evolution would eventually settle on, is to always cooperate with me.
I think the main thing that bothers me is how the result is characterized as "theory of mind" beating evolution. As if evolution couldn't stumble on such simple strategies, and pre-commitment had no cost or associated risk of going extinct.
It's even worse, given that this is actually old news: the zero determinant prisoner's dilemma work was already posted to Hacker News and discussed back in May.
That said, this actually is probably the first actually useful new article posted on this subject (as in, by someone who actually understands the math).
(I therefore upvoted you, as I think you are expressing a valid conceptual problem with the repetitiveness of the front page recently; between App.net, Twitter, and this, it seems like the same few stories keep getting recycled; I also, however, upvoted the actual article.)
(Wow, and apparently I now have a "full disclosure": the guy who wrote this article is the brother of a long-friend of mine; I did not know this before writing the "that said" paragraph.)
This is the first article i've seen that tries to explain the topic. It also is rather deep in not only explaining the paper in question (in a good way, i think), but also comparing it with work of Stewart and Plotkin.
This article was linked repeatedly in the comments of the prior posts. But hey, whatever, down vote me to oblivion, about what I've come to expect around here. Sorry for sounding like a dick but I really am sad about how meaningfulness has decayed around here. People expect opinions to be spoon fed to them easily, that's not how it used to be.
For example, suppose we're playing the iterated prisoner's dilemma. I pre-commit to a strategy: if you've defected in the last 5 games I will always defect, otherwise I will defect unless I haven't cooperated in the last 4 games. In this situation your optimal strategy is to always cooperate, even though I'm taking advantage of you. Punishing me won't change my strategy, I've pre-committed to it. The rational strategy, and the one evolution would eventually settle on, is to always cooperate with me.
I think the main thing that bothers me is how the result is characterized as "theory of mind" beating evolution. As if evolution couldn't stumble on such simple strategies, and pre-commitment had no cost or associated risk of going extinct.