I don't think there is necessarily only one model for intelligence possible, maybe AGI can work on principles quite different from the human brain (even by your own definition of AGI). We do not need to replicate, emulate or just understand in detail the workings of human/biological intelligence to build AGI - at least there is not much evidence (I know of) this is the case. But that does not mean we reach AGI just by scaling within the current approaches either. I just think the lack of understanding of the human brain was a rather weak argument, you may still be right on the larger point of AGI not being reachable using current approaches.
You are viewing a single comment's thread from: