I think there is a small chance that you are right, but a far more vast chance that you are wrong. I don't think anything is going to "cap" any longer. No more AI winters ever again. Further, this is the reason I am fairly confident that a 'human unfriendly' (that is the computing and computing derived AI will be external from the human mind) "technological singularity" is going to occur about the year 2029, give or take two years.