A no‑frills, central home for my posts—mirrored from LinkedIn and elsewhere—so the links stay stable, the ideas stay accessible, and the formatting doesn’t get in the way.
Stop validating computational methods on production data. Build a reference library of canonical systems with known behavior first.
molecularSimulationThe real MD bottleneck isn't compute — it's whether you can trust your results when hardware changes. Analysis should be the boring, stable foundation that survives every upgrade.
molecularSimulationThe most dangerous moment in a lab isn't a failed experiment. It's the moment everyone stops asking questions.
researchPracticelabTalkThe official story: committees carefully read every application and reach a well‑balanced, collective decision.
researchPracticeaiScienceWatson and Crick's pipe-cleaner model wasn't just a prop — it was the first step inside the molecule. Here's how that instinct evolved into today's VR and MD workflows
molecularVisualizationMost breakthroughs don't start with clarity — they start with a plot that makes no sense. Here's why I've made confusion a deliberate part of my workflow.
researchPracticeNot every result makes it into a paper. The ones that don't are often the most useful. On failure as data, not verdict.
researchPracticeEvery few months, a new article points at people who publish a paper every few days. Clarivate cleans up a list. Nature raises an eyebrow. We all shak...
scientificPublishingresearchPracticeCrustafarianism is a religion invented by autonomous AI agents on an AI-only social network. Funny story. Terrifying demo. If agents can spin up a machine cult overnight, imagine what they can do to your infra, data and clusters if you’re not watching. 🦀🤖
aiScience