Quote:
but as far as i understand the matter the question to be answered would be if if they have found a method that approaches GTO and will land there about, given enough time or horse power.
damn-floating-point-numbers-never-add-up-to-one-ly
The problem here as far as I can tell is that they're not actually interested in clearing this up.
People who don't work with computers in general, or AI in particular, are prone to misunderstanding these issues pretty badly. At every stage in the evolution of AI, such as it is, people are constantly solving the easy half of a problem, then declaring that solving the rest is a trivial matter of applying some "horsepower" to it, and then never actually solving the problem. AI stuff are full of these non-linear problems where you come to a promising point early on and then fizzle out.
But let's be clear here, if they had a GTO solution for multiplayer poker, or even a real idea of how it could be feasibly produced, we'd be talking like nobel prize material. I don't think people appreciate how far-fetched the idea that they have near-GTO solutions for multiplayer full stack games is.
Like, the well known progress in poker GTO, imagine that's, like, a walnut. A GTO solution for full stack multiplayer NLHE would be like, our galaxy.