Dave xxxiv 🇨🇦📷🔭✨2d ago
Unnamed TNG skant beefcake

Me: Spends several years crafting a very careful systematic review and meta-analysis showing that most AI approaches applied in a specific medical context are overfitted to that context, and are therefore useless generally

Them: "AI models have played an irreplaceable role in this domain" [cites me et al]

Mar 20 at 12:37pmWeb
Show threadUnnamed TNG skant beefcake2d ago
I swear, the Web of Science "your work has been cited" email should be the "your work has been misinterpreted and cited as if it supports the exact opposite of what you found" email
Show threadUnnamed TNG skant beefcake2d ago
Concept: I'm allowed to write a lay-title for a paper that has swears and isn't neutral, so it's harder to mis-read, like "AI tools in stroke healthcare are absolute shit and you shouldn't use them"
Show threadRachel Rawlings2d ago
@researchfairy I realize research papers are usually narrowly-focused, but I think you could do away with the word "stroke."
Show threadGeoff Wozniak2d ago
@researchfairy Does anyone read even the abstracts anymore, or do they just assume citation means endorsement?
Show threadCaptain Superlative2d ago

@researchfairy

“But I read an AI summary of your paper and it said using AI in healthcare was super duper awesome sauce.”

Show threaddasgrueneblatt2d ago
@researchfairy 😱