3 ms·
I dunno, I could see it working. I do something similar with reviewing code: I have one agent write the code and another reviews it, then they go back and fort
by Esophagus4 3mo ago
I dunno, I could see it working.
I do something similar with reviewing code: I have one agent write the code and another reviews it, then they go back and forth for a bit improving the code. Seems to yield better results than one agent alone.
Seems like a similar principle.
- scheme271 3mo agoThe difference is that in the code situation, you can run unit tests on the code, compile it, etc. Unless your LLMs are ordering diagnostics and reviewing the results, there is no further information that the LLMs have on the situation. Having a second LLM review the first is counterproductive, if the 2nd LLM is better, why not use it directly? If not, then what prevents it from sending the first on some incorrect tangent?
- RussianCow 3mo agoAlso, there are multiple "correct" ways to code something, so imperfect code that solves the problem is still useful. A medical diagnosis is either correct or incorrect.
- Esophagus4 3mo agoEn, I think you’re just trying to justify your pre-existing position that this can’t work. https://www.nature.com/articles/s41746-026-02619-0 https://www.nature.com/articles/s41746-026-02619-0 https://www.nature.com/articles/s44360-025-00007-8?fromPaywallRec=false https://www.nature.com/articles/s44360-025-00007-8?fromPaywa... Different prompt approaches and training doctors to use LLMs can improve accuracy of LLM-assisted diagnosis. It’s pretty reasonable to hypothesize that LLM “peer review” could improve that as well.
- RussianCow 3mo ago> En, I think you’re just trying to justify your pre-existing position that this can’t work. I never said it can't work. I just said that finding the correct medical digagnosis is different than finding a solution to a software problem.