Pinned Tweet
EXCITED to share the release of two foundation models for electrocardiogram interpretation in @ehj_ed We built DeepECG-SL and DeepECG-SSL, two open-source ECG foundation models trained on >1M ECGs and validated across 11 external datasets (881K ECGs).
🔗 academic.oup.com/eurheartj/a…
In 662 angiograms reviewed by a core lab, the original clinical reports flagged only 47% of the lesions the core lab called obstructive.
A model that reads the whole angiogram got closer to the core lab. DeepCORO-CLIP is published in Nature Cardiovascular Research.
Code: github.com/HeartWise-AI/Deep…
Weights: huggingface.co/collections/h…
Figures: Harrabi et al., Nature Cardiovascular Research (2026). Fig. 1 adapted; Fig. 2 unchanged. CC BY 4.0: creativecommons.org/licenses…
Robert Avram retweeted
Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead.
You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement.
I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible.
But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier.
Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product-liability exposure if your products enable a truly damaging cyberattack. The market already punishes models that behave in unpredictable or unauthorized ways. After the Hugging Face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want.
Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders’ “shut it all down.” China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well.
So go ahead and pace the frontier. You are the ones setting it. The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system. So just do it.
If you do, you’ll buy goodwill for the next conversation. If you don’t, we’ll know this was just another bid for regulatory capture — or an election-season psyop.
Robert Avram retweeted
How does an AI-ECG model trained at a hospital perform in a community setting detecting structural heart disease? We answered this question using the PREVUE-VALVE data, out now in @JACCJournals with @timpotsMD @djc795 @PierreEliasMD!
Robert Avram retweeted
introducing ocr in anydoc
now your agents can read scanned docs for free
→ sub-5ms for non ocr
→ 190ms median per ocr page
→ layout detection, tables, formulas
→ ocr hosted option via @firecrawl
→ free, no api key needed
github.com/firecrawl/anydoc
Robert Avram retweeted
OpenMHC brings together 10+ years of My Heart Counts data, to our knowledge the largest openly accessible HealthKit dataset to date:
60M+ hours of sensor data
11,894 consenting participants
19 minute-level channels
169 linked health, lifestyle, mood & behavior variables
Robert Avram retweeted
No more LogiOptions crap... openlogi.org 🙌
Robert Avram retweeted
[6/7] @ICMtl cardiologists and residents reviewed reports without knowing who wrote them.
Rated individually: model 32% better, 35% same, clinician 33% better.
Side by side: model chosen 31%, 42% same, clinician chosen 28%.
Model and clinician reports were rated similarly.
Robert Avram retweeted
[3/7] How do you give an LLM an ECG signal?
1. Turn a 12-lead ECG into discrete tokens using residual vector quantization.
2. Q-Former connects the ECG tokens with text.
3. Instruct-tune MedGemma-4B to answer ECG related questions.
All training details in suppl. material.
Proud of you Rohan. Two years of tokenizer debugging and QA pipelines, and now it is out in the open. One model, one signal, from free text interpretation to LVEF and AF risk. Go celebrate, the reviewers can wait until tomorro
Can one model take a raw 12-lead ECG signal and answer questions ranging from free-text interpretation to LVEF, structural heart disease and AF risk, without a separate model for every task?
We explore this question in our new work DeepECG-Tok!
medrxiv.org/content/10.64898…











