[omedalus]

Marvin Minsky famously published a book in 1986 called "Society of Mind", in which he argued that people don't really have a "mind", but rather an enormous assortment of specialized subsystems that interoperate to produce the illusion of coherent consciousness. I think that this belief is what's needed to guide us forward towards AGI. IMHO, no further technological breakthroughs are necessary to achieve AGI; the remaining work is subsystem integration. I think that Daniel is hitting the nail on the head here with respect to the direction we should all be working in. After all, there already exists a machine capable of attaining human-level intelligence: the human brain. We know what the functions of many of its components are, and we know that those functions must have evolved for darn good reasons. Mimicking its architecture seems like a logical extension of Rosenblatt's original work of building artificial neural networks in the first place, sixty-odd years ago.

[danielcasper]

According to a debate I had with Gemini (rofl what an "authoritative statement"), it claimed that LLMs can't get to AGI for three reasons. First, that it has no desire for self-preservation, to which I responded with self-preservation / non-corruption of itself. Arguably enough. Second, that it has no intrinsic motivation, to which I said it can seek to minimize entropy / maximize information gain of the world around us (min/maxing some function built on Information theory) and it relented. But what I can't solve for...LLMs can't learn like people. It is not a 20W piece of wetware running in the real world able to integrate and learn in real-time. It's gotta be batched trained then reinforced. Maybe this is irrelevant if the machine can pretend to be us sufficiently well, but I feel that, in truth, it's not a true AGI until we find a better algo / hardware substrate.

Oh...wait a minute...organoids are hitting the market :D