One of the most exciting (and challenging) applications of physical AI is thinking about how it can accelerate scientific discovery. Imagine a robot that can pipette, handle stem cells, operate existing lab machinery, 24/7 - a physical AI scientist. This is what the new episode of @GreylockVC Change Agents is all about, featuring @michellearning, founder and CEO of @medra_ai. We talk about building an AI-powered lab, how accelerating physical AI lab experimentation is the key to unlocking drug development, and how Medra is automating experimentation today inside pharma companies. 00:00 Teaser 00:56 Medra’s founding mission 04:17 Hypotheses are not enough 13:27 Medra’s differentiation 20:22 The experimentation bottleneck 27:06 Refining the agent harness 29:05 The role of the physical lab 33:23 Managing context across different customers 34:59 Working with models 40:26 Towards personalized medicine and drug discovery 43:15 AWS for life sciences Thank you for joining @michellearning

New @GreylockVC Change Agents is out today featuring @tuhinone from @baseten We discussed all things inference: - Moving from rented to owned intelligence - How the most advanced AI companies are post-training OSS models to give them more control over performance, runtime requirements, and cost - With agents, inference is evolving to a set of tools for continuous learning vs just "run this model" - Capacity constraints are worse than you think, and nobody is willing to take a hit on capability Full episode below

we're entering the era of AI for the physical sciences: @altaratech comes out of stealth today with a scientific intelligence platform that helps R&D teams go from data to insights in minutes. They're already working teams across semiconductor, battery, and advanced materials to accelerate the process from experimentation to commercialization. Their agents ingest and reason across the complex, multimodal data: semiconductor wafer maps, high-resolution inspection data like SEM images, large-scale instrument time series data, scattered spreadsheets, unstructured research documents, and domain-specific legacy systems – all without requiring years-long migrations to new systems of record. We at @GreylockVC led their $7M seed, alongside our friends at @neo, @BoxGroup, @Liquid2V and incredible angels like @JeffDean and execs at OpenAI and AMD. @catherinehyeo and @evatuecke are building an exceptional team (ex-NVIDIA, SpaceX, Applied Intuition, Jane Street, Microsoft, Warp, ++) in SF. If you want to work on one of the hardest and most consequential problems in technology, check out their careers page