Mentally circumambulating regarding three dimensional GO and Artificial Intelligence.
Ok so given how deep blue defeatede Kasporov in 1997, then Alpha go beat lee sedol while the internet watched in 2016... It almost seems as if we've reached the point at which humans have to create intelligence scaffolding to develop better AI's. Here's an idea. Continue to use the game of Go as an AI battle royale... However instead of a 19x19 two dimensional board. Make the game three dimensionsal. Now I know 3d chess for Star trek is pretty much bullshit... but the simplicity of go lends itself to three dimensions. Further more instead of 19x19, just increase the board size to 190x190 or 190millionx190million. Then you have real tournaments between the various competeiting AI's. Anaylze data, heurestics what have you. But ultimately you have an AI that can play ULTRA-GO better than any other machine. It is the best. Zeta Go. This type of intelligence can go on forever right? Or is there a single three dimensional go strategy for all board sizes so that it, past a certain point, doesnt matter how large the 3d board is, Zeta Go has found the best strategy for all board sizes out to infinity?
