I dont get it. I fully understand that Gary Marcus brings in scepticism everywhere, and I think that's important! We need a dissenting voice that counters the hype. But in the meantime it seems that he really only wants to see something negative everywhere in order to somehow talk it down. Sam announced a roadmap for GPT 4.5 and GPT 5 and Gary sees it as an internal failure. Sorry, but that's silly.

🚨 introducing the world's most advanced medical reasoning agent for chest x-ray, built with @LangChainAI, right here in Toronto! check out MedRAX! 📜 Paper 🖥️ GitHub 🌐 Website @UHN @UofT @VectorInst @CohereForAI

Fun fact: Out of FAANG, Netflix hasn't released a major foundational model since ChatGPT came out. It's also the best performing FAANG stock on a ~3 year and 10-year timescale.

Goodnight sleep tight ☺️💤

Only those who dug deep into offline RL knows the importance of online RL

Yep, I checked

okay someone else should consider hosting another house too because i can't possibly accommodate all the people reaching out to me. this is fire though

if you’re coming to @socraticainfo symposium, i’m organizing a 20+ ppl villa for 2 nights for awesome ppl. night of march 19th to morning of march 21st. expect: jamming music sessions, communal cooking together, socratica pomodoros during the day, evening hike + more spontaneous adventures. $200 CAD per person (4 meals covered). DM me if you’re in. if enough cool ppl are interested i’ll make this happen

only in waterloo

Our first accepted paper in 2025: 🚀 Introducing MorphoDiff: Our new diffusion-based generative pipeline for high-resolution cellular morphology prediction, guided by perturbation signals, now accepted at ICLR 2025! ✨ Key Contributions: 1️⃣ Perturbation-guided generation: Predicts high-resolution cellular phenotypes in large fields of view. 2️⃣ Generalization: Works with genetic and chemical perturbations, even unseen interventions. 3️⃣ Validation: Tested on 3 public Cell Painting datasets, excelling in fidelity, biological interpretability, and visual quality. 🧬 MorphoDiff bridges the gap between transcriptomic-level tools and rich phenotypic insights from high-content microscopy, driving progress in drug discovery. 📄 Learn more: BioRxiv: Code: 🎓 Led by the amazing PhD student @NavidiZeinab, co-supervised with @bhaibeka. Big thanks to @JunMa_11, @emiglietta1, @Le__Liu, and collaborators @DrAnneCarpenter & @CiminiLab for their invaluable contributions!