We're using 20 to 25 different models because creating the output that Gamma creates requires so many different layers of expertise.

Monday partner meeting for venture firms right now is how much value is ultimately going to accrue to the model companies versus how much will accrue to the application companies. The reality is is even Gamma doesn't use just one model. If the model companies could create a phenomenal presentation, we we would only have to use one model. Definitionally, we we're using I think 20 to 25 different models because creating the output that Gamma creates requires so many different layers of expertise that you have to source from multiple different APIs and infrastructure providers and image generation providers and foundation model companies.