🌹Meet #ROSIE, our new AI that predicts spatial multi-protein expression straight from routine H&E images. Built on a large co‑staining dataset (>1k samples, 16M+ cells). It enables cell phenotyping & tissue structure insights w/o extra assays.

Try Sonic MAX in Cursor. It's free, faster than Gemini, and writes better code than Opus 4.1. Seems as good as Grok 4. 🫨

If you measure and interpret RNA levels in single cells, read this: "Our results suggest that instantaneous mRNA abundance of canonical markers is tenuously linked to the chondrogenic phenotype at the single-cell level." 1/2

17 Equations that changed the world by Ian Stewart.

As long as I can remember, I have been stunned by mathematics. Unfortunately, it wasn’t always because of its beauty: I struggled with math during my early years. (My elementary school math teacher still can’t believe that I managed to obtain a PhD in the field.) I am convinced

Check out our new paper where we use in vivo ovarian cancer model and 10x Genomics single cell RNA assay to show that pegylated liposomal doxorubicin supports immune maturation in high-grade serous ovarian cancer mdpi.com/2072-6694/17/1…

Biotech VC Translator

Select your cell: I'm prototyping an interactive sc RNA-seq portal where the UMAP is dynamically linked to morphing 3D artistic renders of each cell type. You can play with the initial build below.

Spatial-DMT is now published in nature ! For the first time, we can simultaneously map DNA methylation and RNA transcription in tissue context. Great collab with Wanding Zhou lab. Big congrats to Chin Chin , Hongxiang Hongxiang Fu and team! nature.com/articles/s4158…

SDR-seq is a droplet-based single-cell DNA-RNA sequencing platform for studying gene expression profiles linked to both noncoding and coding variants. Dominik Lindenhofer Lars Steinmetz EMBL nature.com/articles/s4159…

To anyone who claims AI is slowing down, hitting a wall, or getting more expensive, show them this Aider benchmark. Check out the difference between GPT-5 Medium and o3-Pro: better results at almost ten times lower cost! Nothing in human history has gotten so much cheaper while

By popular request: you can now branch conversations in ChatGPT, letting you more easily explore different directions without losing your original thread. Available now to logged-in users on web.

In a medical first, researchers report that they have implanted CRISPR-edited pancreas cells into a person with type 1 diabetes go.nature.com/3UYR0gD

Gemini 2.5 Pro is still first in LMArena Text, despite being rather old (6 months)

Can you rejuvenate an old brain by giving it young immune cells? 🧠 My lab Calico put it to the test. In our new study, we replaced the brain's immune cells in old mice with young ones. The result? The old brain environment forced the young cells to age RAPIDLY. A 🧵👇

Latest genomic AI models report near-perfect prediction of pathogenic variants (e.g. AUROC>0.97 for Evo2). We ran extensive independent evals and found these figures are true, but very misleading. A breakdown of our new preprint: 🧵

I'm really glad this re-evaluation was done but I just want folks to understand that when they see auROCs > 0.9 in any biological task, it is very rarely means they model is spectacular. It usually means the benchmark is severely flawed in some way. 1/

Cancer is an evolutionary disease, but does knowing a cancer’s evolutionary past help predict its future? Out today in nature, we learnt the evolution of 2000 lymphoid cancers and found it was highly correlated with clinical outcomes! (1/7, link below)

My favorite local coffee shop has a bot farm setup in their backroom to boost IG engagement, this is some next-level local marketing 😂

> be me > Larry Ellison > own a database empire, a sailboat, and a disdain for poor people > OpenAI wants to build AGI > needs compute > *a lot* of compute > like “$300B GPU cluster in a volcano” levels of compute > "Hello. I own Oracle Cloud. Also, I’m rich." > sign $300B GPU