10 years of progress is a flash in the pan of human progress. The first deep learning models that worked appeared in 2012. That was like yesterday. You are completely underestimating the rate of change we are witnessing. Compute scaling is not at all similar to biological scaling.