Bryan Hanson list of free and open source software for spectroscopy is a good indicator of the poor quality of commerical software for science equipment And the opportunity for disruption by startups

Very interesting observation from Charles Yang that VC funding just for "autonomous labs" (robotic labs that can produce and do real-world tests of AI-generated candidates in chemistry, biology and material science) now exceed total NSF funding in those areas.

One weird quirk of the AI bubble is that VC investment in AI for Science, specifically autonomous labs, now rivals the annual NSF budget for materials and chemistry research: ml4sci.substack.com/p/venture-capi…

Who is the most cracked "under the radar" AI for science person you know? Getting a team together for something big.

if you want a piece of the great race for lab automation, think about electron microscopy. it sounds like a medical procedure, but it’s probably how we’ll unlock the secrets of the universe. one of the most powerful observational tools humankind has ever built. hack to start

The partnership between Deepmind and the U.S. Hurricane Center is a good example of how private sector compute innovations can support the national interest, and depends on public institutions ability to adapt and integrate new technologies

great paper. These are exactly the kind of tools you need to enable autonomous matls discovery in the TEM

A 50 year story of scaling compute + diffusion: Left: advances in numerical weather prediction models + scaling compute, 1955-1985 Right: Google Deepmind's performance at hurricane prediction out to 120 hours outperforms ensemble computational models and human experts, 2025

Don't make me tap the sign

This is an excellent ethnography of how researchers across a number of field use AI today Reading through the examples though, it is worth flagging: these are problems across a number of fields of science, but all of them are posed in computational/mathematical terms

It's funny that some of the best analysis of U.S. scientific infrastructure is from Chinese researchers Despite all the angst from progress studies/metascience, U.S. science funding/infra is still the gold standard everyone looks to

New piece from me where I argue science advances through new instrumentation e.g. - TEM for understanding superalloys - cyclotrons repurposed for uranium enrichment - cryo-EM for protein data bank w/ implications for how we think about AI for Science: substack.com/home/post/p-17…

A few reactions on the Genesis AI for Science EO today: > notable to see DOE and its 17 national labs charged with taking the lead, as opposed to NSF > while not quite as organized as other frameworks, one can definitely trace the throughlines of leveraging federal data +

read this take not openai's