And I think we’ve already worked on robustness, privacy. And with the language models, I think it’s like a magnification glass with some of the problems we have been seeing so far. And that’s something that’s keeping us busy right now.
j previous speech k next speech