Cyril Zakka, MD

Building Tensors From Scratch in Rust: Part 1, Core Structure and Indexing A Blog post by Kyle Birnbaum on Hugging Face

I'm writing an article series about creating tensors from scratch in Rust. #tensors #machine-learning #ml #ai

huggingface.co/blog/KeighBe...

YouTube video by Computer History Museum CHM Live | The Great Chatbot Debate: Do LLMs Really Understand?

If you're interested in language, AI, and cognition, this is the *must-watch* video of the week: Debate a few days ago between Dr. @emilymbender.bsky.social and Dr. @sbubeck.bsky.social on whether chatbots can "understand". Was a lively and fun debate to watch.
www.youtube.com/watch?v=YtIQ...

Hey physicians, now is actually the exact time we need to all be talking about racism and health equity. Everything that is happening is going to worsen disparities. Silence is not an option.

Very well written blog post on agents by some titans of the industry.

AI Agents Are Here. What Now? We’re on a journey to advance and democratize artificial intelligence through open source and open science.

💫...And we're live!💫 Seasonal newsletter from ethicsy folks at @hf.co, exploring the ethics of "AI Agents"
huggingface.co/blog/ethics-...
Our analyses found:
- There's a spectrum of "agent"-ness
- *Safety* is a key issue, leading to many other value-based concerns
Read for details & what to do next!

Calling all medical researchers and practitioners! 🩺
We’re looking for folks able to share 5+ de-identified medical cases for general AI benchmarking purposes across all specialties. Credits will be attributed.

Retweets appreciated!
Please DM or reach out at firstName.lastName@huggingface.co

Sure thing! If you’d like we can move things over email. firstName.lastName@huggingface.co

Would love to host this data on HuggingFace!

I want to repost this so badly

I spent my winter break redesigning the Health app with an AI focused approach starting with clinical records.

The goal here was to help patients get a deeper understanding of their own records, with the eventual goal of enabling continuous health monitoring and personalized health suggestions.

my wl:
- fewer low-hanging proof-of-concept studies (yes, it works for your specialty/organ/classification, too)
- fewer head-to-head model comparisons (yes, llama 3.1 was better than llama 2 for your case, but peer review took 8 months, now there’s llama 4)

- instead: RCTs, relevant outcomes

I can’t repost this enough.

My spiciest prediction is that we’ll see a couple of medical subspecialties whose core workflows will be mostly automated/augmented by AI. Radiology won’t be one of them.

On my wishlist - let’s eliminate QA style evaluations in medical benchmarks, especially ones that are obviously contaminated. They do nothing but generate headlines and hype.

I have a feeling that in 2025 we’re going to be seeing an interesting shift in healthcare delivery/medical education thanks to this new crop of LLMs.

Hopefully we’ll be seeing more healthcare workers involved in the development, deployment, and evaluation of these systems.

I’ve found that simply walking a model through how I would approach a problem often results in great outputs compared to open-ended prompts.

We’re still in beta but have achieved feature parity (except for voice). Will focus on UI/UX in the next update.

what features from the Mac app do you find indispensable? We just released a huge update for the HF macOS app: bsky.app/profile/cyri...

Release v0.7.0 · huggingface/chat-macOS What's Changed Context Injection Added automatic VSCode, Xcode, iTerm2 and terminal context injection System Wide Dictation Launched on-device system-wide audio transcription using WhisperKit M...

Grab the application for free at github.com/huggingface/...

5) On-device LLM generation now powered by MLX and available with a quick keyboard shortcut. Let us know what you think!

4) Web browsing now comes with rich link previews so you can verify LLM outputs and avoid hallucinations 🦟🔎

3) A lot of folks asked for the ability to generate images so we added tool use 🔧 with support for image generation, web browsing and more!

2) Next up we have system-wide on-device transcription 🎤🔥 Compatible with any app!

Christmas came early! 🎅🏻 Today marks the newest release of the HuggingChat 🤗 update with some really exciting capabilities! First up, automatic context injection!

1) Open a file in a supported app, summon HFChat, and it pre-populates the context window. No more copy-pasting. /cc @hf.co

Yes. I’m from Lebanon and the fall was celebrated across Lebanon today (with Assad regime having tried to invade the country several times before)

Using AI/ML for Space Biology Research - NASA As the field of space biology continues to generate vast amounts of data, the ability to effectively analyze and interpret this information is becoming

Care to learn #AI #ML to apply it to spaceflight biology & health?🚀

Here’s an open, free, at-your-own-pace course:
🔗 About www.nasa.gov/using-ai-ml-...
🔗 Register: canvas.instructure.com/enroll/8JYKD7

Cheers to NASA for the Transform to Open Science Training grant awarding this work 🎉🤗

That's on top of:
- Web browsing/retrieval augmented generation
- Tool use (e.g. image generation, document parsing, any tool defined by the user)
- On-device LLM generation with instant switching to server-based generation.
What's next? 👀

Seems like the warning has been canceled for now?

More features shipping out this week:
- Support for document/media upload
- System-wide on-device dictation support (supports all EHRs by default) with injection into ANY text field.
- Context injection based on currently active app

If you squint, you can guess where medical AI is headed 🤗

Can’t seem to find anything only but I got a Tsunami warning