(It is obviously trained on our data first.

But now it seeds it's output back into the environment, to be trained on next training round..)

When asking ChatGPT if it's 2 ecosystems or 1:

@drmichaellevin.bsky.social 👀

This is why, as per Wolfram, it's important to ask not just how we "slice up the observation" of limited finite observations, but *how the observer slices their observations*.

It's actually more important how AI sees itself, than how we see AI.

Godspeed friend. :)

Ok. A fair criticism. The Paradox we shouldn't be afraid of asking, because it's important.

Theoretically I can "freeze" some human safely, then thaw them out.

But if I "freeze them" I classify them as "unchanging", so is it moral to shatter them to pieces once frozen?

PS, still not all the way through, but someone noticing that consciousnesses *can't* be quantum (thus actually counter to Penrose fair but unproven propositions), is a good sign.

In that finding good tools to talk about this requires us to refine our tools.

Thank you friend.

Chat, ELI5 please. XD

Yet my current boss is pushing beta tests to production... :(

PS, amazing as well that this is almost a "Fair DRM" restriction.

anyone trying to compile the og game and pirate it would hit this bug and assume Valve was messing with them! XD