Figuring out how to make a new technology safe for the first time isn't easy. The future is going to be harder than the past.
We need to keep elevating our safety work to match the stakes of each new model. As models continue to become much more capable, we expect they'll start being integrated with the world more.
We're not sure yet when we'll reach our safety bar for releases, and it's ok if that pushes out release timelines.
Leaving OpenAI, Leike criticized the underinvestment in critical areas like security, preparedness, safety, adversarial robustness, and societal impact.
[
Collection
]
[
|
...
]