Questioning myself again about how we can define that AI becomes real intelligent, aka AGI, and how we can define that.

There two ways: practical tests and ideological meaning. First one was pretty long known as Turing Test, but now seems little bit outdated because of mimicry bug. Here I want to talk the second - ideological meaning. By that I mean our identification of who we’re compared to other leaving things.

Let’s bring ancient analogy about the ship: if you replace every wood plank in the ship, would it still be the same ship? That’s a Ship of Theseus or Theseus’s Paradox.

Imagine same thing about human: replacing every part of body with mechanical/electronic analogue with similar output would change definition of human? Well, maybe to cyborg or robot, but the nutshell it will be generic human creature. With some differences.

So replacing some amount of neurones in brain with external gpu connected to it would change human identification too much? We should ask paralysed guy who using brain-computer-interface to manipulate mechanical hands instead his own.

Getting back to replacements - imagine building neural network that big enough to duplicate literally same amount of neurones as in the human brain, or even more, with continuous growing, replacing or even obsoleting, with good mechanical body that relies on similar principles like blood pressure, heartbeat, temperature, and with emotions, mood, dreams, and even complains about taxes. Wouldn’t it make conceptual identity for humans, like robot humans, or artificial humans? I mean, they are not the same ofc, but since abstract picture looks very similar, so why not?

From point of view of aliens that comes from far space imagine one day they see a planet with humanoid “meat” and “metal” creatures that looks similar and act similar, won’t it be almost the same for them? The only difference could be that those can’t make a kids together. Until someday.