It's nice that he's doing this and I wish him well. But thousands of very smart people have been working on AGI since 1958 and many of them initially thought they could crack the problem in a couple of years. They made a huge amount of progress, but the more progress we make toward AGI, the farther away the goal seems to recede.

AGI is very much a research problem. It's not going to be solved with a clever hack.