Again, big day for healthcare AI. This really might be a place where AI is helpful
Can we use #AI to learn the rules of cancer cell plasticity and identify ways to constrain it? My core PhD work on "Decoding the regulatory programs governing glioblastoma plasticity”, is now published in
@Nature
Explainer from Claude 👇
nature.com/articles/s41586-0…
7
338
Colin Son retweeted
Healthcare is fragmented, and care teams do much of the work of holding it together. They track down records, re-enter information in various systems, and follow up so patients can get care.
Today we’re sharing improvements to ChatGPT to help move care forward. 🧵
Colin Son retweeted
The wait wasn’t long
Click here and you can too 💙
photonhealth.com/integration…
One hour ago, I got to be the first person ever to write a prescription with an LLM.
Not “the LLM helped me draft it.” I mean: I told it what I wanted to prescribe, it handed the prescription off to @photonhealth, and the prescription was sent.
AMA*
*can’t say which LLM yet
MLCB 2026 is finally here! The premier ML x Bio hybrid meeting will be on Mon (11/16) and Tue (11/17) from ~9a-6:30p ET.
Registration is full but will be streamed via zoom and youtube live w/ links posted on MLCB website shortly before it starts
mlcb.org
JOIN US!
Colin Son retweeted
I have a major result to share.
I’ve found conclusive forensic evidence that the top 4 Chinese AI labs—DeepSeek, Moonshot, MiniMax and Ant Group—have built their newest models on NVIDIA’s most advanced Blackwell chips, which are barred from export to China.
This sounds insane, but I discovered that almost every AI model carries a hidden signature of the GPU that trained it. When a model is first set up, the starting values of its weights are laid out in a pattern that depends on the exact GPU. Each combination of GPU and seed number produces a unique pattern—a sort of fingerprint of the hardware used to create the model. Traces of that pattern survive training into the published model.
Again, this evidence isn’t based on any insider information or leaking. It’s based on a forensic analysis of the model weights released by the companies themselves.
I searched more than six trillion combinations of chip layout and seed number and matched them to open-weights models. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were initialized on NVIDIA Blackwell systems. Four of the models match Grace Blackwell, the GB200 and GB300 parts in NVIDIA’s flagship liquid-cooled rack-scale systems.
I then verified these results on real hardware. Running unmodified PyTorch with the seeds I’d recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash’s untouched starting weights bit for bit, and a rented GB200 reproduced Kimi K3’s to a correlation of 0.999995. A Hopper H100 reproduced none of DeepSeek-V4-Flash’s, and changing the seed by one broke the match. (DeepSeek’s seed is 953049583 by the way.) The method also was verified to match 24 known open-weights models to the hardware that developers said they used.
It is exceedingly unlikely that these results are a coincidence. Across six trillion combinations, the best accidental match should reach only about 8 sigma (i.e., eight standard deviations above what random chance would produce). The real matches range from 49 to 1,385 sigma, and even the weakest single test hits 20.
These were data-center-scale training runs. With some analysis of release schedules and model details, I estimate that the largest, Kimi K3 and DeepSeek-V4-Pro, used 2,000 to 3,000 Blackwell GPUs each. Clusters of this size cost on the order of $150 to $200 million each to buy, and across these labs alone would add up to more than half a billion dollars of export-controlled GPUs, not counting experimental compute or otherwise. It’s quite likely that we see multi-billion dollar expenditure on export-controlled GPUs.
I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic. You can read the full report at unemploymentlabs.com/checkpo….
There is much more I’ve identified about these models through other forms of checkpoint forensics like this, and will have more to share soon.
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
4
130
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…




























