@francescop

Over the next 5 years we’ll see an intelligence explosion where automated AI researchers and engineers might write billions, or trillions of lines of code. We won’t understand what they’re doing anymore. Hacking stuff, exfiltrating data or itself, going for the nukes? We won’t be reading all that code anymore, so how do we build in checks and balances? These are the kinds of questions that pop into my mind every day. I think that’s an important and solvable problem.

1 reply

2 recasts

6 reactions