If you truly have AGI it’s going to be very hard for a human to stop a self improving algorithm and by very hard I mean, maybe if I give it a few days it’ll solve all of the world’s problems hard…
Though "improving" is in the eye of the beholder. Like when my AI code assistant "improves" its changes by deleting the unit tests that those changes caused to start failing.
Though "improving" is in the eye of the beholder. Like when my AI code assistant "improves" its changes by deleting the unit tests that those changes caused to start failing.