Sam Rawal, MD

Excited to launch the public beta of Osler - a workspace + copilot for physicians writing H&P notes that is designed to augment thinking, not replace it. As a resident physician, I built Osler around two core beliefs: 1) the act of writing notes helps doctors think and plan 2)…

Excited to be heading to the @MayoClinic AI Summit tomorrow! We’ve been developing clinical LLM evals that reflect real-world medical practice (and challenge frontier models) better than existing benchmarks at the @IrbazRiaz BEACON Lab. Looking forward to sharing some of our…

Just got a collection of o4-mini agents to collectively critique and work together on a piece of a research problem that I solved on my own that took about a week. It’s clear at this point that the latent capabilities of these models is much larger than is accessible via 1-shot…

This captures why I started building Osler-long before considering AI copilot features, I wanted to design clinical note software optimized for the ergonomics of structured thinking. Writing a great H&P medical note helps us generate and evaluate ideas, not just document them.

using @cline regularly has massively accelerated how Bitter Lesson-pilled I've become

Thread of side effects of “reasoning”

Very sweet message from Claude after logging in after a long time :)

this is a big concern I have about some of the clinical AI systems being deployed that offer to write entire medical notes for the physician - you are letting it determine what is relevant, and this represents offloading clinical reasoning (not just note writing!) to the system.

There’s a lot of talk about the value of taste when working with AI, ie in architecting software systems or in designing. I think taste is pretty prominent in the clinical domain too (although maybe called something else - experience? judgement?) - different physicians have…