No, that's getting crucified. But that hasn't happened yet.
Trump, for example?
So what is the real world example of a person getting convexified?
Imagine that there were only two famous econ papers that used Lagrange multipliers to identify optimal solutions, AM and KG... of course, many economists would compare those two papers even though Lagrange multipliers were invented elsewhere. This is exactly what happens here, except that the solution technique is Legendre-Fenchel duality or something like that rather than Lagrange. If KG used Lagrange and if you attributed Lagrange to AM, then you could say that "KG can be expressed totally in terms of AM" just because you are mistaking Lagrange multipliers for AM's main contribution...
No, neither KG nor AM invented convex analysis. In both cases, the contribution lies elsewhere.
This exactly.
Imagine that there were only two famous econ papers that used Lagrange multipliers to identify optimal solutions, AM and KG... of course, many economists would compare those two papers even though Lagrange multipliers were invented elsewhere. This is exactly what happens here, except that the solution technique is Legendre-Fenchel duality or something like that rather than Lagrange. If KG used Lagrange and if you attributed Lagrange to AM, then you could say that "KG can be expressed totally in terms of AM" just because you are mistaking Lagrange multipliers for AM's main contribution...
No, neither KG nor AM invented convex analysis. In both cases, the contribution lies elsewhere.
Weird troll.
KG can be entirely expressed in terms of AM. That is a true fact.
However, KG have a way better story, that has launched a whole new literature.
By the way, any idea why it is KG and not GK? Is K just another whale?
Can you seriously imagine that G can do theory? The guy is a reg-monkey.
The result in AM is literally that, for the purposes of computing the limit value of the game, it is "as if" the informed player could simply choose an information structure for the uninformed player once and for all at the start of the game, and commit to this choice in that no new information is revealed during the course of the game. And the choice of this information structure just has to respect Bayes's rule, and the resulting value can be computed by reading it off of the concave envelope.
So if somebody (i.e., KG) then goes and studies the commitment case, and says that the main result is that the choice of the information structure just has to respect Bayes's rule, and the resulting value can be read off a concave envelope, then I am not sure in what sense it hadn't been done before?
Yes, they popularized it, sure. (But we can argue whether that was a good thing or not...)
KG copied AM's result but gave a "better" story. This is enough for AER these days.
Actually, the better story -- the Sender-Receiver game with commitment -- was in Kohlberg (1975), "The information revealed in infinitely-repeated games of incomplete information"
"Zero-sum two-person games with incomplete information on one side are considered. It is shown that the information revealed by the informed player in an infinitely-repeated game with simultaneous moves is essentially the same as the information revealed by him in a one-stage game in which he must move first."