Journal Club: Performance of an LLM on the Reasoning Tasks of a Physician
Journal club slides on Brodeur et al. (2026), Performance of a large language model on the reasoning tasks of a physician (Science).
cfu288
Blog covering topics spanning computer science and medicine.
Latest posts
Large Language Models and Agents: A Hands-On Companion
The interactive companion to Part 2 of the AI primer for clinicians. Run a small language model in your browser and watch memory, grounding, and agents behave the way the lecture describes.
Large Language Models and Agents
Part 2 of the AI primer for clinicians: what large language models and agents actually do, how they fail, and how to use them safely.
Machine Learning Foundations
The first lecture in a primer series introducing artificial intelligence concepts for clinicians.
Exploring CQL Support in SDC Questionnaires and Potential for Executing CQL in the Browser
Exploring the feasibility of adding support for CQL in FHIR Questionnaires (and eventually to lforms), enabling the execution of user-provided CQL expressions and libraries within questionnaires in the browser. Enabling this would allow for the creation of dynamic forms that can calculate the values of items in a form based on user-provided CQL expressions and libraries.
A Visual Guide to the Hierarchical Navigable Small World Algorithm
Explore the Hierarchical Navigable Small World (HNSW) algorithm through interactive visualizations and demos, providing an in-depth understanding of its mechanics and applications in similarity search.
Introduction to Python
Learn the basics of Python programming with this beginner-friendly, interactive tutorial.
A Visual Guide to Vector Search and K-Nearest Neighbors
Explore the K-Nearest Neighbors (KNN) algorithm and how it helps in similarity search.
Securely Creating and Storing Cryptographic Keys in the Browser
Doing cool things by combining WebCrypto and IndexedDB
Container Orchestration, Object-Oriented Programming, the Human Body, and Resiliency
What do these things have in common?
Creating Per-Pull Request Environments on Digital Ocean's App Platform With Github Actions
Automate creating staging environments to easily see your changes before deploying
Python has a type system, but where are the interfaces???
AKA how I took too long to discover Python's Protocol classes
Rendering in React - the devil is in the details...
Just a few key points