2/ there are 2 schools of thought: 1) compute is the only real bottleneck to AI progress. the more we spend, the closer we get to AGI 2) we are hitting a data wall which will slow progress regardless of how much compute we have

→ View original post on X — @alexandr_wang