The problem with LLMs reaching true AGI is it's basically "static" intelligence. Changing code, context, prompts and even fine tuning can improve output, but is still far from realtime learning.
The "weights" in our brains are constantly evolving.
The "weights" in our brains are constantly evolving.
It is categorically wrong that non static learning is a requirement of agi. The biggest problem we face is hallucinations and this isn’t caused by the fact that agi can’t learn on the fly.
I take your point about the non-necessity of dynamic learning for AGI.
But in any event, drawing conclusions about the real world from a fictional story seems fraught.
In fact the amount of limited agency Leonard had is similar to the limits in agency we want for an agi that serves humanity.
Anyway my main point is Leonard has intelligence well past what we expect from an agi. So because people like Leonard are documented to actually exist one can say live learning is not really needed to have Leonard levels ( aka agi levels) of intelligence