Shawn Wang: belief

Transcript context

…Yeah. You guys had a very inspiring model spec. I think Joanne worked on that. Where you said, like, yeah, we don't want to overly refuse all the time. In fact, like, even if, like, not safe for work, like, in some occasions, it's okay. How, is there an API that we can say, not safe for work, okay? I think we'll get there. I think we'll get there. The mobile spec, like, nailed it, like, you know. It nailed it! It's so good! Yeah, we are not in the business of, like, policing, you know, if you can say, like, vulgar words or whatever. You know, there are some use cases, like, you know, I'm writing, like, a Hollywood, like, script I want to say, like, will go on, and it's perfectly fine, you know? And so I think the direction where we'll go here is that basically There will always be like, you know, a set of behavior that we will, you know, just like forbid, frankly, because they're illegal against our terms of services. But then there will be like, you know, some more like risky, like themes, which are completely legal, like, you know, vulgar words or, you know, not safe for work stuff. Where basically we'll expose like a controllable, like safety, like knobs in the API to basically allow you to say, hey, that theme okay, that theme not okay. How sensitive do you want the threshold to be on safety refusals? I think that's the Dijkstra. So a…

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