nec speraveris sine desperatione nec desperaveris sine spe
Texas
Joined December 2024
- Tweets4.3K
- Following2.8K
- Followers849
- Likes14.3K
Colin Son retweeted
Today, we are excited to introduce Ladon, a state-of-the-art biological foundation model that predicts disease trajectory and treatment response in patients. There is immense complexity in the individual biology of every patient. Two patients can have the same diagnosis and receive the same treatment, yet have very different outcomes: one responds well and lives for years, while the other’s disease continues to progress. Understanding and resolving this heterogeneity to predict outcomes is how we accelerate our delivery of medicines for patients and unlock clinical trial abundance.
Ladon is our first step towards this future: one where we can learn more from every patient enrolled so that we can run smaller, faster trials with higher success rates. We built it to give drug developers quantitative support on billion-dollar clinical trial decisions. Today we're excited to share Ladon's capabilities in our blog and technical report.
Ladon is a 3B parameter model trained on millions of RNA-seq samples spanning decades of technical and biological variation across tissues and species. Ladon beats gene expression baselines, established biomarkers, and existing RNA foundation models at predicting patient prognosis and therapeutic response in 140+ oncology tasks, including external validation on past clinical trials. This is significant because clinical trials are fundamentally generalization problems. We show that Ladon-derived predictions are robust across indications, treatments, and sequencing platforms.
Reach out if you’d like to get started with the model today!
Read more about Ladon in the blog: blank.bio/introducing-ladon
Find more technical details in the white paper: blank.bio/introducing-ladon
Colin Son retweeted
Jensen can still be CEO if Mike Pence only has the courage to do the right thing.
This is really terrible public health take. I say as a dem voter but also surgeon dealing with a lot of trauma. There is already no doubt autonomous self driving, while imperfect as all systems, is significantly safer than human drivers. Full stop.
Next week, Tesla plans to launch robotaxis in Austin— before Texas’ new AV safety law takes effect. We’re urging a delay until those safety standards are in place.
Public trust comes from safety and transparency. We look forward to working with Tesla to achieve both. #txlege
1
3
109
Colin Son retweeted
Did you know that @SenRickScott was forced out as CEO in 1997 (of what is today HCA) amid a federal investigation of Medicare and Medicaid fraud?!
Did you know that HCA ultimately paid approximately $1.7 billion in fines, penalties, and settlements arising from allegations of Medicare and Medicaid fraud?!
Did you know that HCA pleaded guilty to 14 federal felonies arising from the massive Medicare fraud investigation that began while @SenRickScott was its chairman and CEO?!
Did you know that @SenRickScott invoked the Fifth Amendment 75 times during a 2000 deposition in a separate civil lawsuit, not the federal criminal investigation?!
Did you know that @SenRickScott latest financial disclosure, covering 2025 and filed August 13, 2026, establishes a minimum of approximately $248.6 million, with disclosed asset ranges extending above $776 million?! (The exact amount cannot be determined because congressional disclosures use valuation brackets)
Colin Son retweeted
Doctors are sprinting to build solutions to frictions they’ve tolerated for years. But there’s a chasm between a vibe-coded prototype and a solution that can be deployed safely in a healthcare system.
Great piece by @mnedoszytko in @NEJM_AI
ai.nejm.org/doi/full/10.1056…
Colin Son retweeted
Can practice help AI agents get better at surgeries? 👀
We let GPT-6.1 Sol perform 3 successive brain surgeries, learning from its previous procedures how to locate anatomical landmarks, specifically bregma and lambda. After each attempt, an expert surgeon corrected the positions, and the agent saved images and measurements of both its estimates and the corrections.
This represents in-context learning from real-world experience!!!
By the third one:
- Landmark search: 169 → 86 sec
- Corrective moves: 4 → 2
- Position error: 0.6 → 0.3 mm
N=3, but exciting early results!
Colin Son retweeted
forbes.com/sites/richardmeng…
My colleague Dr John Cowan- first neurosurgeon elected to Congress.
Colin Son retweeted
Today's results are better understood as a left-center wave, not a woke one. Trump lost big but so did a lot of far left ideas.
- Billionaire tax failed in CA
- Transgender school sports restrictions passed in WA
- Increased penalties for fentanyl dealing and even possession passed in CO
- Wiener won over Chan in SF
Nationwide the electoral mood was normie lib wine mom, not far left. A victory against the horseshoe.
Colin Son retweeted
🚀 Introducing #PathCowork: Learning from real-world pathology practice to build AI copilots!
Most pathology AI learns what pathologists conclude, but rarely how they actually practice.
Diagnosis is more than recognizing patterns. It requires knowing where to look, when to zoom in, how to connect evidence, and what to examine next.
Using our #TissueLab (tissuelab.org) platform, we captured synchronized eye gaze, voice narration, and slide navigation from 16 pathologists across 48 sessions and over 1,000 minutes of real-world diagnostic examinations at @PennPathLabMed, under an active IRB.
We explored whether AI could learn directly from this clinical diagnostic expertise.
🔍 Learning where to look: Visual grounding improved from 5.6% to 52.2%, enabling AI to better locate diagnostic features in tissue images.
🧭 Learning how to examine: Our agents outperformed Claude Code and Codex across 5 pathology tasks, running ~5× faster at ~10× lower cost.
💻 We are releasing PathCowork-1T and PathCowork-27B, along with code and model weights!
Our vision is to turn everyday clinical practice into a continuous source of learning for medical AI.
The next breakthrough may come not just from scaling models and datasets, but from fundamentally rethinking what AI learns and how it learns.
📄 Paper: arxiv.org/abs/2610.39441
💻 Code & weights: github.com/zhihuanglab/PathC…
#DigitalPathology #MedicalAI #AgenticAI
@PennPathLabMed @PennMedicine
For what Tesla did popularizing electric it is incredible FSD/Robotaxi are going to be their most important contributions to humanity
I cannot stress enough how important self driving cars are for public health
In 10 years Karpathy might deserve a Noble in Medicine
3
277
Colin Son retweeted
Today in @Nature, we’re sharing the breakthrough research behind the Personal Health Agent (PHA)—the foundational science that was fundamental to how we built the Google Health Coach in the Google Health App.
Colin Son retweeted
A new randomized controlled trial evaluated whether AI-generated, layperson-friendly summaries of medical findings provided before outpatient consultations improve physician–patient communication. Learn more: nejm.ai/3W3z9Zy
ALT Illustration of an AI-generated, layperson-friendly summary of medical findings provided before outpatient consultations. On the left, a physician in a white coat and stethoscope speaks and gestures, set against a faded table of clinical trial statistics and medical terms. On the right, a patient in a hospital gown smiles and gestures back. A blue line connects a glowing pen icon with sparkles to a document icon labeled “Discharge Report Summary."
No one should downplay her experience. To be an amateur (or fuck even a professional) and riff like this against one of the biggest comedians? (Also bravo Conan just letting her go and playing along)
Impressive skill and a great Subway takes segment. She is awesome.
1
2
313
Colin Son retweeted
Today we are thrilled to announce Clinical System One (CS1), a fast and efficient general-purpose decision model for medicine.
Read our preprint 📄: arxiv.org/abs/2610.27099
Watch the video 🎬👇
Blog: linkedin.com/pulse/reading-f…
Healthcare has a scale problem. Every health system holds decades of notes, reports, and timelines. They contain answers to questions we rarely get to ask of more than a few patients at a time. Which patients could join this trial? Who missed a follow-up? What happened to everyone who took this drug? Reading all of it, for every question, has always been too slow and too expensive. That was true even with LLMs.
Thats why our team built Clinical System One (CS1). Its a small, fast decision model tailored for clinical records. Trained on 12.4 million questions and 3.6 million patients, CS1-4B beats open decision models on reading real-world patient records across diverse healthcare systems, and it runs on a single desktop GPU. We show the power of composing CS1 with traditional System 2 LLMs into agentic systems that read dense patient timelines and assess trial eligibility from a pool of a thousand candidates. Imagine trials that find every eligible patient automatically, registries that fill themselves, and care gaps surfaced across a whole population. Thats the future were building toward.
