Noam Shazeer

organizing and creating new information from, you know, some guidance you give it.

Can you help me write a letter to my veterinarian about my dog?

It's got these symptoms and I'll draft that.

Or can you feed in this video and, you know, can you produce a summary of like what's happening in the video every few minutes?

And, you know, I think

Our sort of multimodal capabilities are showing that it's more than just text.

It's about understanding the world in all the different kind of modalities that information exists in, both kind of human ones but also kind of non-human oriented ones like –

weird LIDAR sensors on autonomous vehicles or, you know, genomic information or health information.

And then how do you extract and transform that into useful insights for people and make use of that in helping them do all kinds of things they want to do?

And that's, you know, sometimes it's, I want to be entertained by chatting with a chatbot.

Sometimes it's, I want answers to this really complicated question.

There is no single answer.

source to retrieve from.

You need to pull information from 100 web pages and figure out what's going on and make an organized, synthesized version of that data and then dealing with multimodal things or coding-related problems.

I think it's super exciting what these models are capable of and they're improving fast.

I'm excited to see where we go.

Yeah, I guess I'm pretty excited about kind of a lot of fundamental research questions that sort of come about because you see

Something that we're doing could be substantially improved if we tried this approach or things in this rough direction, and maybe that'll work, maybe it won't.

But I also think there's value in seeing what we could achieve for end users and then how can we work backwards from that to actually build systems that are able to do that.

So as one example, organizing information, that should mean