I don't know but I hope that Chomsky is less wrong. Because if statistical methods reach an asymptote, we will have no choice but to try to better understand the principles and foundations.

If statistical methods do not reach an asymptote, I don't think we will have the incentive to reach a deeper understanding