Applied Compute CEO Yash Patil argues custom AI models are the natural next step once general models prove feasibility

Custom models aren't needed for everything, but most of the value will accrue to them for the reasons you mention

The right to capture margin only exists for as long as you have a unique (~custom) value proposition

And general models compete on data and competition rich frontiers

Custom models have been tempting for years, but they are finally becoming worth the effort.

We are doing everything we can to lower that barrier to entry.

@PossibltyResult 🎯

@ypatil125 @ypatil125 just posted this one treating agent(not model) as the entity to improve and building systems of learning that make enterprise adoption more platform centric over time.