When tested with a classic psychological assessment, advanced AI models experienced a total breakdown in focus. A new PNAS Nexus study suggests these systems lack the human-like executive control necessary to override automatic responses and maintain complex goals.
So true, and the things that LLM agents are good at, humans test very poorly by comparison, particularly on speed.
To be fair, run an LLM on a machine with an equivelent power requirement to the human brain and we might se some different results on that one.
While it’s true that a human brain only uses ~20W of power, it’s a really specific kind of organically delivered power with all sorts of environmental requirements that we, being humans, take for granted, but in the bigger picture it’s really a rare location in this universe that doesn’t kill us nearly instantly - much less provide that 20W of power in a form a brain can use.