Recent content by danylot

I think people will use AI for healthcare whether providers encourage it or not, so the goal should be making it genuinely useful and safe. If it can answer common questions, help patients understand what comes next, or point them in the right direction, that's already a huge improvement. I've...

One thing I think people overlook is that there's a big difference between using AI through your healthcare provider and pasting your symptoms into a general-purpose chatbot. Just because the conversation is about health doesn't automatically mean it's protected under healthcare privacy rules...

This is exactly why I don't think AI should be used to diagnose people directly. If something as simple as a typo or slang can change the advice a model gives, that's a real safety concern. Where AI does seem to add value is handling lower-risk tasks like answering general questions, helping...

This is the direction I think healthcare AI needs to move in. General-purpose models are impressive, but medicine is one area where accuracy, traceability, and knowing when not to answer matter more than sounding confident. I also like the focus on clinician workflows instead of trying to...