That's the most depressing graph I've seen. Via The New York Times

This post sometimes gets misunderstood (most recently on hackernews after my ICLR keynote :)). The idea behind having your entire life on your calendar isn't so much to say that your entire life needs to be prescheduled or you need to account for how you're spending every minute

"Why is exercise good for you?" (nature news & views) - Humans, as hunter-gatherers, evolved to 'always move' - Exercise alters gene expression throughout the body - With exercise, muscles contribute to anti-inflammation through IL-6 production nature.com/articles/d4158…

So happy to see our SimulateGPT paper peer-reviewed and published in CBM. Grateful to all the great co-authors, especially Stephan Reichl,QuantifiedScientist, Matthias Samwald and Christoph Bock Lab @ CeMM & MedUni Vienna sciencedirect.com/science/articl… Simulation-based prompting is becoming a powerful method to query LLMs

'The primitive endoderm (PrE) supports lineage plasticity to enable regulative development' - PrE is extra-embryonic tissue (at day 4) but can still recreate a full blastocyst (i.e. also embryo) - JAK/STAT supports this plasticity - cell.com/cell/abstract/…

"Transformers need glasses! Information over-squashing in language tasks" - In typical LLMs every token only 'sees' previous tokens (causal attention) - This leads to information inbalance across tokens - Thereby, later tokens 'vanish' in long sequences arxiv.org/pdf/2406.04267

"What is a cell type?" - science.org/doi/10.1126/sc… - Cellular identity is insufficiently described using current 'cell type' labels - Perturbations can lighten the landscape around given cells - Diverse starting points and multimodal measurments show lead to "phenoscape" of cells

"Contextual AI models for single-cell protein biology" - Protein embeddings are powerful. - But: they are usually cell-type agnostic - This paper: Embedding cell-type-aware protein representations via GNNs on scRNA-seq-pruned protein interaction graphs nature.com/articles/s4159…

Had an amazing time with the Bock lab at this year's lab retreat! Grateful for the opportunity to help with the planning and coordination—was really fun and rewarding to see the results.

🗨️ WANNA TALK TO YOUR CELLS? Try out CellWhisperer – our new multimodal AI that turns single-cell RNA-seq analysis into a conversation. No coding needed, just chat in plain English. Short walkthrough below. Web app & bioRxiv preprint linked in the thread. Let's dive in! (1/9)

🚩So happy to finally be able to share this with all of you!🚩 A huge 'thank you!' goes to the whole team and especially Peter Peneder. It has been such a pleasure to grow the initial idea into this mature stage together.

Excited to share our new multimodal AI for chatting with scRNA-seq data! 🧬🤖💬 Big thanks to all involved, especially Moritz Schäfer , who drove much of the project. It was a great pleasure to collaborate closely, growing the initial concept into this mature stage together! 🙌🎉

Had a fantastic time last week at Cold Spring Harbor Laboratory's #biodata24, with amazing discussions around AI and bioinformatics and presenting my work on CellWhisperer. Thanks Michael Schatz and all organizers, looking forward to being back in 2026!

Thrilled to see our work with the Christoph Bock Lab @ CeMM & MedUni Vienna featured in The Scientist! It's a really great read and builds on an interview I had with the journalist Kamal Nahas. the-scientist.com/an-ai-lab-part…

Super fun project with my colleagues at the AI Institute of MedUni Vienna.

Just presented our new multimodal histopathology method "SpotWhisperer" at ICML, one of the largest AI conference. SpotWhisperer enables spatially resolved annotation of histopathology images using natural language by "transferring" annotations from transcriptomic data.

🚀 I’m excited to announce Stanford Graph Learning Workshop 2025, happening Tuesday, October 14, 2025 at Stanford University University (with online livestream). Free registration! Submit a talk/poster. 📍 This year’s workshop will spotlight three fast-moving frontiers in AI & data

Transformers are great for sequences, but most business-critical predictions (e.g. product sales, customer churn, ad CTR, in-hospital mortality) rely on highly-structured relational data where signal is scattered across rows, columns, linked tables and time. Excited to finally

📝Finally out: Chat with your cells in English language - right in the browser. ✨ Try it yourself: cellwhisperer.bocklab.org Huge shout-out to my co-first-author Peter Peneder and to Christoph Bock Lab @ CeMM & MedUni Vienna and to all the other contributors for the amazing teamwork!

Love the work, great job Kuan Pang and Yanay Rosen!