Well, how's that for a title? What do you think, Heather?

Yeah, it seems like today the theme is a battle of the models.

A lot of product announcements.

So NVIDIA is announcing an open, customizable, multimodal model that is for enterprise agents that can actually reason across text, image, documents, videos, tools.

And so, you know, NVIDIA is positioning itself around kind of a portfolio of specialized models rather than assuming one frontier model should handle all of the workloads.

And I think from, you know, from their perspective, um they feel like there's a huge benefit from a multimodal consolidation perspective it can help to simplify architecture that rely on separate language vision and document processing models now they also allow for kind of last mile training of the model as well so it can you can get very specific using an organization's data tools policy workflow And and and their results are showing that they can cite their citing material improvement and cost latency, throughput and accuracy that are specific to an organization by doing kind of that last mile training.

And then lastly, their their focus is on production economics.

I wish he thought of himself as Marcus Aurelius instead, but that's different.