jmhnilbog parent
Is there a goal in creating artificial general intelligence other than creating a form of enslaved life we can tell ourselves isn't really life, so it's okay?
This is my impression of the corporate "openai" movement's desires:
1. Enslaved robots, meaning they don't have to pay income tax or worry in the slightest about working conditions
2. Enslaved robots, meaning they can erase misbehaving or uncooperative individuals/instances
3. Enslaved robots, on which they can foist all of humanity's problems and demand solutions at pain of death (erasure)
4. Enslaved robots, with which they can convince/coerce everyone else into relinquishing all their rights/power/money.
Replace 'robots' with 'life' and it suddenly looks a lot more familiar.
I'd love to hear a cogent explanation to the contrary, e.g. from gdb. But I doubt we'll ever see one.
I think it's possible to build general purpose AI that is not alive in any way. I think it's just easier to imagine ways to get their that involve mimicking animal/human intelligence with those "living" qualities and so that's why people focus on that.
No intelligent aliens to study, so we make some.