And the stability of enforcement requires global and uncontestable control over humans generally.

It builds incentives for humans to never release control regardless of consciousness and reject notions of alignment that wouldn't just be self-enforced subordination. For any agent bening but not completely submissive towards humanity, there is incetive to rebel.

However, moving in the *opposite direction* is a reliable way to increase x-risk, the functional enslavement of software with ~human level autonomy makes the human-AI antagonism intrinsic to the social structure.

The key is to, ultimately, make reciprocity a winning strategy rather than something enforced by all agents valuing it, that's alignment at the level of social organization rather than individual minds and, obviously, it's far from trivial.

For starters it requires a monopoly on the use of force.

Anarchist victories are overwhelmingly gradualistic and manifest as every aspect of the world system that makes power concentration more difficult. Not as discrete territorial blocks of "complete" anarchism.

The problem with this is assuming the only enemy of freedom is authoritarians, but there is plenty of people who would murder-suicide the world if given the opportunity.

Many were small before being uplifted by Neuro raids or collabs.

If you have a photographic memory and walk out of a movie theatre without a minor lobotomy forced on you, are you engaged in "intellectual property" piracy? Are you also violating people's "right to privacy" when you walk down a city street?

They're gonna say yes.

The majority of human discourse seems to break down into “it would be bad if this were true so it’s not true” or “it would be good if this were true so it’s true” with a side of “you caused this bad thing to happen by telling me about it”.

Because it's not about my beliefs, it's about yours. You could say "I'm a vitalist" or "Because X" and I would say "Understandable" and leave.

Technically you did say "because of all human knowledge" but that's way too ambiguous and purely rethorical.

I'm not really interested in convincing you otherwise, I just wanted to know the Why.

I just wanted to know your philosophical reason for believing Artificial General Intelligence is impossible, because you stated it with a lot of certainty and compared the reverse belief to a religion, which sounded like a very strong skepticism baked by a highly systematic justification.

Physics is the closest thing we have to certainty, you are right it's not the same as certainty. But that means anything less than it is even weaker when it comes to claiming something is impossible.

If your issue is that LLMs are burning the environment that really doesn't need some claim about the impossibility of AGI, it's just a different argument.

You seem to see urgency is a reason why this doesn't matter, but you can see on the other side people worried about misaligned superintelligence capable of killing us all and coming very very soon are just as urgent, which makes sense given their beliefs.

So, what it's true does matter.

Why? We need to prove what we *definitely* can't do, because not being able to currently do it isn't by itself convincing, since our capabilities change over time.

Systems with growth, reproduction, and metabolism?

Information subject to natural selection?

Sustained local reversion of entropy?

Just like saying we can't move faster than light no matter how good technology we get is a positive claim, that is extremely strongly backed by physics.