[m_fayer]

There’s something phenomenally powerful, uncanny, and potentially deeply corrosive about current AI. Dismissing it as “evil” is pat, and prevents any full encounter with something that now irrevocably exists and deserves and demands the consideration of thinking people.

[mlsu]

What would you call it?

Anti human, corrosive, powerful, uncanny... but not evil?

I am fully encountering this, believe me. My conclusion is that it's evil.

[m_fayer]

Potentially corrosive. Emphasis on potentially.

Powerful and uncanny are morally neutral traits.

At root an LLM is a different kind of mind that thinks different kinds of thoughts. (I’m making no claims about consciousness here.)

It’s a novel object for us humans, and we’ll need decades to come to grips with it. In the meantime I think a bit of humility is due.