eoravkin

4 ms·

1. 1.▲
   by eoravkin 9mo ago

   You built a cool product. I'm actually one of the founders of https://medisearch.io which is similar to what you are building. I think the long-tail problem that you describe can be solved in other ways than with live APIs
2. 2.▲
   by eoravkin 9mo ago

   Out of curiosity, did you actually see any pharma ads on OpenEvidence?
3. 3.▲
   MediSearch (YC S23) Is Hiring a Founding Engineer in Central Europe(ycombinator.com)

   1 points by eoravkin 1y ago
4. 4.▲
   by eoravkin 3y ago

   Thanks a lot! We put a lot of energy into solving hallucination. The tldr is that we have a eval set to test hallucination and tweak stuff to optimize for performance on this eval set.
5. 5.▲
   by eoravkin 3y ago

   Thanks! Glad that you like it:) We haven't trained or fine-tuned on neither of USMLE or MedQA. We use several LLMs in our system, and honestly I don't think there is any single thing that I can point to which is different to how o
6. 6.▲
   by eoravkin 3y ago

   We'd be interested to hear examples or ideas of how this could be useful for you!
7. 7.▲
   by eoravkin 3y ago

   Your point about non-urban regions is interesting. Haven't thought about that. I wonder whether it will also be much harder to show a tool like this to people there though.
8. 8.▲
   by eoravkin 3y ago

   Thanks! Glad you liked it. Symptoms and indications --> diangosis is a bit more challenging. When we talked to some doctors, they always said that seeing the patient directly is very important. Perhaps a system like this could be a good
9. 9.▲
   Show HN: Search engine for trustworthy medical information(medisearch.io)

   5 points by eoravkin 3y ago|5 comments