 A lot of people have seen models like chat GPT, which have really captured the public imagination. And they think that foundation models are mostly about creating chatbots. Another common misconception about foundation models is that they're only about text. Certainly some of the most powerful and exciting foundation models in the first ones to come on the scene have been mostly focused on text data. But the principle underlying foundation models, this idea that we can leverage a large corpus of unannotated data, applies to other domains as well. At IBM, we're building foundation models for a wide variety of different kinds of data. Everything from geospatial data for predicting of, you know, flooding and weather and different sort of geospatial quantities to molecules to data from the operation of cloud systems. So really any kind of data you can imagine is potentially fair game for a foundation model.