Dario Amodei: belief

Transcript context

…But the possibility of use cases here, just the range is incredible. So how much to make it work really well in the future? How much do you have to specially kind of go beyond what the pre-trained model is doing, do more post-training, RLHF or supervised fine-tuning or synthetic data just for the agentive stuff? Yeah. I think speaking at a high level, it’s our intention to keep investing a lot in making the model better. I think we look at some of the benchmarks where previous models were like, “Oh, could do it 6% of the time,” and now our model would do it 14 or 22% of the time. And yeah, we want to get up to the human level reliability of 80, 90% just like anywhere else. We’re on the same curve that we were on with SWE-bench where I think I would guess a year from now, the models can do this very, very reliably. But you got to start somewhere. So you think it’s possible to get to the human level 90% basically doing the same thing you’re doing now or it has to be special for computer use?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.