Well, go back to my metaphor of those three onion rings. This is really about onion ring number two, which I call the hardness. It's Menlo Ventures did a very interesting study last summer showing that only 11% of companies were actually changing model providers when they were using Entropic because precisely when you put a model into production in a given workflow or in a series of workflows, the way that model runs in that loop to get the tools to actually deliver and outcome matters a lot. 

Not just the intelligence, but how the whole intelligence works to give you your result as an enterprise user. 

And Entropic, why did I say orchestration in the moat last May? Because that is what Entropic has been playing. And then you have coding as the wedge, so you had orchestration as the moat and coding as the wedge. 

They built their configuration and their stickiness in the enterprise AI stack By going beyond, we just have a great model, but by allowing that model to actually deliver concrete results in the enterprise.