This is an excellent take and I agree with most of it. Current AI is still narrow AI but in the domain of pattern matching, which gives the illusion that it's understanding and processing things like a human. Despite its shortcomings it's already changing how we access information and I do think it will change interfaces across a lot of domains as suggested.
However I'm not so sure AGI will need emotions.
> I'm by no means a psychologist or philosopher, but I think we would need some sort of genuine consciousness or emotions like knowledge. Not just simulated responses, but actual felt experience.
I don't think you need emotions for intelligence. I think emotions are heuristics that evolution has created BEFORE there was human level intelligence. All emotions are tied to our instincts, to our animal side. We just haven't discovered the algorithm for intelligence in humans. LLMs will show impressive amounts of knowledge but then it gets confused or refuses to follow a simple detail in the prompt in a way that an intelligent entity would never do.
However I'm not so sure AGI will need emotions.
> I'm by no means a psychologist or philosopher, but I think we would need some sort of genuine consciousness or emotions like knowledge. Not just simulated responses, but actual felt experience.
I don't think you need emotions for intelligence. I think emotions are heuristics that evolution has created BEFORE there was human level intelligence. All emotions are tied to our instincts, to our animal side. We just haven't discovered the algorithm for intelligence in humans. LLMs will show impressive amounts of knowledge but then it gets confused or refuses to follow a simple detail in the prompt in a way that an intelligent entity would never do.