The 90/10 rule of scientific discovery: 90% of your results come from 10% of your explorations, but it's not possible to know in advance what specific explorations those will be.

Just tried Google’s NotebookLM to turn my complex paper on Bayesian Illumination into a podcast. 🤯 It really outperformed my expectations! Better flow than some real life podcasts, easy to listen to, and mostly accurate. 👇

Our new paper in Nature Machine Intelligence tells a story about how, and why, ML methods for solving PDEs do not work as well as advertised. We find that two reproducibility issues are widespread. As a result, we conclude that ML-for-PDE solving has reached overly optimistic conclusions.

So here’s a rather painful 🧵 on how unpleasant it is when people jump to conclusions about inappropriate behaviour. It happened to me almost twenty years ago, but the fact it is still so fresh in my mind indicates quite how unpleasant I found the experience. 1/6

FDA approves first schizophrenia drug with a new mechanism of action since the 1950s tinyurl.com/34ceshw4 Karuna/Bristol Myers Squibb’s KarXT (Cobenfy), which acts on M1/M4 muscarinic acetylcholine receptors, expands the treatment landscape beyond dopamine-targeted drugs

Bayesian Illumination: Inference and Quality-Diversity Accelerate Generative Molecular Models 1. This paper introduces Bayesian Illumination, a novel generative model that integrates Bayesian optimization with quality-diversity methods to accelerate the discovery of small

Many congratulations to the The Baker Lab Podcast! Only one episode in and already winning a Nobel. ✨

Goodbye to my beloved Oslo Philharmonic ... perfect way to conclude 4 years in Oslo. 🇳🇴 ✨

The overall story, so far: Nothing here looks like a story of various campaign decisions or a VP choice or big set piece speeches. It's an electorate mad at the incumbent WH's record and expressing it pretty uniformly everywhere, blue and red states alike.

Word of the day is ‘recrudescence’ (17th century): the return of something terrible after a time of reprieve.

🚀 We are hiring full-time ML researchers and PhD-level interns! Join us for exciting projects in AI for Science (weather and climate models) and understanding LLMs. Full-time role: jobs.apple.com/en-us/details/… lists US but EU is also OK—apply regardless of location preference!

It is great time to be in protein design Rosetta Commons #rosettacon

Disordered protein binding by Joseph Rogers (ft. the fab four) at #rossetacon.

i don't think all academics are driven by trauma. but academia attracts ppl who experienced school as a central space of validation, maybe the only place they were ever praised/"seen." also, academic accomplishments are wielded to counter parental neglect & societal debasement.

The state of field at the moment ... writing the alphabet in proteins. #rosettacon 🤯

Our group is recruiting a postdoctoral researcher for a project on reactive machine learning potentials. Reposts and spreading the word to interested people in your network appreciated! jobs.ethz.ch/job/view/JOPG_…

If anyone here is in London and interested to have the best classical music experience of their life for free, DM me! Due to sickness, I'm giving away a ticket for Mahler's 9th conducted by the 97 (!) year old Herbert Blomstedt. Tonight, 7.30pm, Royal Festival Hall.

I got on this platform as a teenager to follow the arab spring. I'll never forget the first images to come out of Syria back then. I still hope to see Assad in a docket at The Hague one day, but for now I'll settle with his regime being brought down. #syria

Well, tweeps ... it took me a while but I think it is time to stop visiting this place. I'm off to bluer pastures ... 😉 🐦 See you there!