LLMs are surprisingly bad at finding medical papers
skeptichuman.substack.com
Over the past 5-6 years, it has become very clear to me that the mainstream guidelines in almost every subfield of medicine is suboptimal, partly due to corruption, partly due to the immense biases and inability of medical researchers to admit they were wrong despite mounting empirical evidence against their pet theories. Thus, instead of relying on the mainstream guidelines, I have often had to exhaustively find and carefully read all published literature about important health questions. Just reading the abstract is not enough: it is also very important to read the methods section to see if the experiments really support the claims in the abstract. Also, ideally, looking for replication of results by groups with no conflicts of interest is also important. Doing all the above for every medical decision is very tedious and for years, I have wanted to find a way to automate it.
LLMs are surprisingly bad at finding medical papers
LLMs are surprisingly bad at finding medical…
LLMs are surprisingly bad at finding medical papers
Over the past 5-6 years, it has become very clear to me that the mainstream guidelines in almost every subfield of medicine is suboptimal, partly due to corruption, partly due to the immense biases and inability of medical researchers to admit they were wrong despite mounting empirical evidence against their pet theories. Thus, instead of relying on the mainstream guidelines, I have often had to exhaustively find and carefully read all published literature about important health questions. Just reading the abstract is not enough: it is also very important to read the methods section to see if the experiments really support the claims in the abstract. Also, ideally, looking for replication of results by groups with no conflicts of interest is also important. Doing all the above for every medical decision is very tedious and for years, I have wanted to find a way to automate it.