Interesting. It would be much more inspectable and controllable and modular which would be good for alignment.
You’ve got some good ideas in here, have you ever brainstormed any alignment ideas?
Unfortunately, it’s a lot easier to come up with good, or at least interesting, capability ideas than alignment ideas; and on the rare occasion I’ve had worthwhile alignment ideas, they often turn out to be tied to capabilities anyway.
Interesting. It would be much more inspectable and controllable and modular which would be good for alignment.
You’ve got some good ideas in here, have you ever brainstormed any alignment ideas?
Unfortunately, it’s a lot easier to come up with good, or at least interesting, capability ideas than alignment ideas; and on the rare occasion I’ve had worthwhile alignment ideas, they often turn out to be tied to capabilities anyway.