Ooof. This sounds like the worst idea I've heard of for a while.
Uncategorized
10
Posts
8
Posters
5
Views
-
Ooof. This sounds like the worst idea I've heard of for a while. I don't feel like this should need to be said, but don't trust an LLM for medical advice. Seriously.
-
Ooof. This sounds like the worst idea I've heard of for a while. I don't feel like this should need to be said, but don't trust an LLM for medical advice. Seriously.
@mike "OpenAI's new 'dedicated experience' that taps into your medical records..." What could go wrong š
On the other hand, as a womanā¢ I'm actually kind of excited for the advanced pattern recognition that AI could bring. It might help with better identifying women's health issues.
-
@mike "OpenAI's new 'dedicated experience' that taps into your medical records..." What could go wrong š
On the other hand, as a womanā¢ I'm actually kind of excited for the advanced pattern recognition that AI could bring. It might help with better identifying women's health issues.
@Gina Theoretically, AI may be able to better identify women's health issues than some rando doctor, but as a Americanā¢ the whole idea of giving all my medical information to a company that's more than likely to turn around and sell it to the highest bidder creeps me the hell out. Especially since those companies could very well be health insurance companies that could decide to deny me coverage based on conditions I may or may not even have. And that's assuming the "AI" is not some random LLM.
-
@Gina Theoretically, AI may be able to better identify women's health issues than some rando doctor, but as a Americanā¢ the whole idea of giving all my medical information to a company that's more than likely to turn around and sell it to the highest bidder creeps me the hell out. Especially since those companies could very well be health insurance companies that could decide to deny me coverage based on conditions I may or may not even have. And that's assuming the "AI" is not some random LLM.
@Gina unless the AI is localised with a database for health records at the very least, it shouldnāt be trusted. @mike concerns are real and totally valid. I urge people to be careful with giving out data because once itās out there it can also be used in unintended ways (like against you).
I would run the AI locally and query symptoms to webMD or an internal database of health conditions so that you could keep data as private and secure as possible.
-
Ooof. This sounds like the worst idea I've heard of for a while. I don't feel like this should need to be said, but don't trust an LLM for medical advice. Seriously.
@mike avoiding to even mention any sort of compliance with established requirements (HIPPA?).
If this feature is enabled in the EU they should get fined heftily, as it is against our current AI established regulation (health & law enforcement purposes fall under high risk AI criteria).
"Daddy" Trump will probably swing in with harsher tone to get us to loosen our legislation, if that's the case. -
@mike "OpenAI's new 'dedicated experience' that taps into your medical records..." What could go wrong š
On the other hand, as a womanā¢ I'm actually kind of excited for the advanced pattern recognition that AI could bring. It might help with better identifying women's health issues.
@Gina @mike Unfortunately, AI isn't really good at identifying women's health issues. So, "what could go wrong" could actually be fatal.
https://www.newscientist.com/article/2510065-ai-chatbots-miss-urgent-issues-in-queries-about-womens-health/
https://femtechnology.org/2025/07/30/built-to-fail-her-ais-invisible-bias-in-womens-health-and-how-we-fix-it-for-everyone
https://www.business-humanrights.org/en/latest-news/uk-ai-tools-downplay-womens-health-issues-study-finds/ -
Ooof. This sounds like the worst idea I've heard of for a while. I don't feel like this should need to be said, but don't trust an LLM for medical advice. Seriously.
@mike ... or a tech corp to keep your data safe, never mind not use it against you
-
@Gina @mike Unfortunately, AI isn't really good at identifying women's health issues. So, "what could go wrong" could actually be fatal.
https://www.newscientist.com/article/2510065-ai-chatbots-miss-urgent-issues-in-queries-about-womens-health/
https://femtechnology.org/2025/07/30/built-to-fail-her-ais-invisible-bias-in-womens-health-and-how-we-fix-it-for-everyone
https://www.business-humanrights.org/en/latest-news/uk-ai-tools-downplay-womens-health-issues-study-finds/@kdkorte @mike yeah chatbots are pretty useless, but applications like improved breast cancer screening sound interesting.
https://www.sciencedirect.com/science/article/pii/S0720048X24001736
-
Ooof. This sounds like the worst idea I've heard of for a while. I don't feel like this should need to be said, but don't trust an LLM for medical advice. Seriously.
@mike Hey AI, should I drink Drano for my sniffles? AI: "You're absolutely right."
/End scene
-
@mike "OpenAI's new 'dedicated experience' that taps into your medical records..." What could go wrong š
On the other hand, as a womanā¢ I'm actually kind of excited for the advanced pattern recognition that AI could bring. It might help with better identifying women's health issues.
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@mikkaels
In ogni modo ĆØ una pianta che ha un valore inestimabile per la sua singolaritĆ . L'altro fratello superstite, che sta in Piemonte, mi ha raccomandato di non farla fuori, dato che alcuni mesi fa avevo espiantato un mandarino, un clementine e un arancio per mettere delle piante di olivo. Di mandarini mi sono rimaste altre due piante, sto mangiando mandarini da oltre un mese e ne ho ancora per alcune settimane! Piante vecchissime, hanno piĆ¹ di 40 anni di etĆ , ma molto generose
-
@NullTheFool i don't bother with the NC or ND variants, because they mostly just make trouble for regular well intended community members, CC-BY-SA is sufficient to discourage big corporations, and scammers aren't going to abide by the license anyway
-
In the country
-
Minneapolis Mayor Demands ICE Leave City After Fatal Shooting | TIME
-
@mikkaels
Il giardino, molto trascurato e incasinato, ĆØ in comune con la mia casa e quella in cui abitava mia madre e un mio fratello, ma io non me ne sono mai occupato per vari motivi. Ma nell'ultimo anno e mezzo sono deceduti in successione mia moglie, mia madre e il mio fratello, quindi ho deciso di ridare vita e rigoglio a questo giardino. Ci sto lavorando dall'autunno del 2024 ma c'ĆØ parecchio lavoro da fare. Domani faccio una foto.
-
it's a map to a magical land
-
@NullTheFool i use Apache 2 for code because it is a permissive license that makes it clear that any contributions to the project are made available under the same license by default, thus avoiding the need to make contributors put it in writing.
for media, usually i pick whichever of CC0, CC-BY, CC-BY-SA best fits my intent for a given project.
-
@NullTheFool For free assets Iād just use one of the Creative Commons licenses. They offer a lot of options.
For code, I guess the question is permissive or copyleft? For a less common permissive license, thereās Apache. Itās like MIT/BSD but requires a more clarity on modifications and patents. Or for kind of an oddball, thereās the Artistic License 2.0, which has broad relicensing compatibility with GPL or similar copyleft licenses but different requirements on distribution.
