Will Ryan
banner
fieldtheory.bsky.social
Will Ryan
@fieldtheory.bsky.social
Flight Paramedic | MechE
Average at paramedicine, wildly proficient in Excel
Here to learn, help improve out-of-hospital medicine in some small way.
When ‘about 75 mcg’ triggers a mathematical crisis…
One thing I hate about talking stats on the internet is that there is always a pedant who lives farther up on the precision-usefulness tradeoff than I do and thinks I should be there with him.
December 28, 2024 at 7:27 PM
Reposted by Will Ryan
You will be visited by three ghosts
December 25, 2024 at 3:00 PM
This is interesting because GPT-4 alone outperformed a group of doctors who could consult with GPT-4. This contradicts the usual assumption about how humans will work with AI. While it’s a small study, it may offer insights into where we are headed.

jamanetwork.com/journals/jam...
Large Language Model Influence on Diagnostic Reasoning
This randomized clinical trial evaluates the diagnostic performance of physicians with use of a large language model compared with conventional resources.
jamanetwork.com
November 25, 2024 at 3:49 AM
Reposted by Will Ryan
You've probably heard that Bicarbonate has to turn into CO2 to raise the pH.

“Don’t give bicarb if you can’t increase ventilation.” But why?!

How much CO2 is there in an amp of sodium bicarbonate? Hos much does it raise pH? And how long does it last?

A 🧵 all about our ICU frenemy NaHCO3

1/
November 19, 2024 at 1:46 PM
Reposted by Will Ryan
A frustrating component of many scientists believing “we just need to teach people facts to change their behavior” is an inability to recognize most people don’t think or operate the way scientists do. (Scientists, in theory, are carefully trained to change our minds/views in response to new data).
In brief, this is the idea that if people just KNEW more about the problem, they’d think and act differently. Knowledge ->attitudes -> behavior. This leads to a lot of cries that we need to just “educate people”…
November 17, 2024 at 11:28 PM