abhisuri97

med student here: Reducing immunosuppressant not something to be taken lightly for kidney transplant patients. I was shocked when I read that sentence.

this is pubmed. Most papers that are funded by NIH research are available on pubmed if the main publisher gives access to the full text (after some set embargo period...usually around a year).

Med student here: oftentimes the attendings who are googling are usually doing it because the patient's symptoms don't fit with the most common illness "scripts" we develop in our mind and have ready for the 90% of patients who walk in the door. The google is a quick sanity check to see if these symptoms are within the range of "normal" for the most likely differential diagnoses (i.e. list of most likely diagnoses based on the patient's presentation).

That or those symptoms are exceptionally vague or uncommon enough that they warrant a quick refresher on google for leads on additional questions we should ask of patients (the most common offender here is rashes/skin lesions imo since they can literally be a manifestation of super simple "oh you just changed your shampoo" to "you have a rare autoimmune condition"...asking a comprehensive history from patients can help determine what tests to order).

I'm honestly surprised it managed to output hex and still be sensible. what part of the training corpus even has long form hex values that isn't just machine code?

Yes. Doctors in the US have to get a certain amount of Continuing Medical Education (CME) credits (usually done by attending conferences and lectures).

Med student here. My guess is the following: So with CF your pancreas can’t secrete many of the enzymes that are necessary to actually digest food. You can take medications to help with that digestion, but regardless, you aren’t actually getting all the calories in your food with CF because it isn’t making its way into your body (notably patients with CF have steatorrhea which is fat in their stool because they can’t absorb a lot of fat from their foods like a non-CF patient). The med helps the pancreas recover some of its ability to secrete digestive enzymes and so patients can now eat and get more from their food. The issue is that there needs to be a recalibration in terms of how much patients are eating. Previously a 2000 calorie diet may not have gotten them so far because they didn’t absorb much of it, but now they’re absorbing a lot more of it. Plus increased work of breathing with CF expends more calories compared to a non-CF patient (and patients on this drug).

as a med student, i'd 100% pay for this service if it can take my in house lecture ppts as input

I’m more trying to see what the utility of stable diffusion (or just the text to image problem) in the long term. Right now people can play around with making weird art pieces and maybe it will be integrated into design tools...but then what?

Eg with other AI problems out there I can see a potential application to medicine, self driving cars etc, but I just don’t see what the bigger goal of this is going to be.