Yeah, some people (who don't understand how LLMs work) think that LLMs can grow into AGI. The fact is that neurobiology is many, many orders of magnitude more complex than any AI we have. Even if we did try to model a new AI after biological neurons, I still don't think it would be a good solution for AGI, since it will probably get things wrong as often as a human would. But LLMs are just fundamentally not going to get us to AGI, in the same way that building a taller ladder will never get us to the moon.

But that's just because LLMs lack the complexity required to exhibit emergent behavior (beyond what it was specifically designed to do), and even if it did, or we found some other framework that could lead to AGI, I'm not sure we'd ever have the compute to facilitate such emergence.

But that also doesn't mean that emergent intelligence cannot exist. So no, I disagree that that is where my reasoning leads.

Humans are more than just biological computers

I never said anything about computers, you did. I'm just talking about the biochemical substrate from which our bodies work. I'm very curious, though, about what you mean by "more than biological computers". I'm going to assume it's consciousness, awareness, whatever. Is it some part of us that doesn't have some roots in our biochemistry?

I'd like to know which part of us you think that mind-altering drugs act on. Alcohol inhibits activity in the frontal lobe, and our decision-making capability is diminished as a result. Psilocybin, LSD, and other hallucinogens affect how we perceive the world. SSRIs and other antidepressants generally tend to improve mood. Xanax alleviates anxiety. Anesthetics cause a loss of consciousness (as do a lot of things).

I agree that humans are more than the "biological robots" I mentioned, but I am not sure we agree on why. I would say it's because of emergent complexity, but if you say it is something else, then I'd like to know how you explain the effects of the drugs I listed (majority of which we know the biochemical mechanism) on human cognition. Or how we can detect changes in brain activity for different scenarios or neurodegenerative disease in fMRI.

That’s a theory, without enough evidence to confirm

Yes, it is a theory, but it has mountains of evidence behind it. It's not "confirmed reality" but it matches and explains all of the data we have. It's the theory of evolution by natural selection. The only thing that matters is the gene. Everything else is to help the gene's existence persist. You eat energy-rich foods, that helps your survival, and your genes make sure you feel good for it. You make friends, that's forming a social bond, helping your survival, and your mental health improves. You have sex, that's propagating the gene, and so the gene makes sure your body is organized in a way that makes it feel good.

Framing it as “turns out actually that” ignores the facts. Presenting a hypothesis as if it’s some confirmed reality is anti-scientific.

And saying "humans are more than just biological computers" isn't presenting a hypothesis as confirmed reality?