A clinical co-pilot, an administrative co-pilot, and an autonomous clinical agent meet in a bar... actually, they met on stage at the J.P. Morgan Healthcare Leadership Conference (@ShivdevRao of @AbridgeHQ, Joan Butters of @Xsolis_Health, @munjalshah of @hippocraticai), to talk about: 🔵 what's been good/bad for their business about the high speed, nearly daily evolution of foundation models in the last couple of years (good = continuous product performance improvements/capabilities; bad = increased noise from so many "basic wrapper" entrants because the barrier to building a live (seemingly) working AI product has gone way down) 🔵 whether it's getting cheaper or more expensive to build AI products/companies (getting cheaper on the model side, but specialized talent and distribution have always been, and continue to be, expensive) 🔵 a reality check on the AI job displacement trope (not really a thing in an industry where job reqs sit open/unfilled for months because of the labor shortage) 🔵 how the EHR platforms are responding amidst all of this (seem to be playing nice so far...) 🔵 how AI apps companies should all be grateful to the AI scribe players for greasing the wheels of the buyers - "those AI scribes are working really well, let's do more AI stuff!" The app layer continues to be where value is most rapidly accruing, especially in healthcare domains where specialized, enterprise grade models are required for high-stakes use cases. And the rate of organic, market-driven adoption continues to grow at unprecedented levels (and, as I wrote about recently, without any government mandates or artificial incentives or pandemics!). Thanks to the @jpmorgan healthcare team for hosting.