Post

Conversation

Sam Altman
@sama
I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks. Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon.
Quote
Dario Amodei
@DarioAmodei
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our
158.2K
Views
View quotes
Amitav Krishna
Post your reply

…
@axrxrxr100
The billionaires like Elon Musk are responsible for putting us in danger with AI. Nobody asked for Ai, humans are capable of doing anything, robots are just taking jobs and careers away and making it harder for humans to survive
2.9K
Liron Shapira
@liron
To government: Do not let them set a precedent that the cooperation is a voluntary act by them. You will also get worse short-term consequences if you allow this.
48
Magnus Rosenberg
@rosenforge
I volunteer as an independent evaluator. Can't be many qualified people around not already working at other labs. Might as well be a random builder like me. At least I know what alignment looks like
774
David Flagg
@DavidFlagg20
And, there it is. Employee-like access, independent evaluators. Who pays for these evaluations, Mr. Altman? Who designs what they should look like? Who watches the watchers? Let me guess, they will be almost all former Anthropic, OpenAi, Deepmind employees and so on, right?
1.1K
Rabbi Chud
@AFGenXbro
Trust is so broken. At this point I can't tell if a pause is legit for safety reasons or to buy time on the giant bubble.
24
ergophobian
@indefatigabile
employee like evaluators sounds VERY difficult to kinda just have? what technical specification would these evaluators have and why aren’t they working for a lab already? like, it seems like they gotta be kinda cracked what do u think Sam
Image
572
Joshhfm
@joshhfm
it's too late for this now. chinese labs have capable enough models now to advance into oblivion in no time.
1.7K