3 ms·
Something else less related that I find really useful about these chatbots is their ability to take some abstract explanation of some problem/attempts to tackle
by sva_ 3y ago
Something else less related that I find really useful about these chatbots is their ability to take some abstract explanation of some problem/attempts to tackle it, and it'll often spit out a name for it, and explain that this sort of problem is already well studied. That can be really difficult with other more traditional tools.
I feel like this could be a big deal in science in general. People of different fields often do unnecessary extra work (as in reinventing the wheel) because they're not aware that there's, say, 10 years of research on it.
To take an extreme example, just think of the poor fella doing medical research who rediscovered integration[0].
0. https://fliptomato.wordpress.com/2007/03/19/medical-researcher-discovers-integration-gets-75-citations/ https://fliptomato.wordpress.com/2007/03/19/medical-research...
- skydhash 2y ago> I feel like this could be a big deal in science in general. People of different fields often do unnecessary extra work (as in reinventing the wheel) because they're not aware that there's, say, 10 years of research on it. Maybe an introductory search in other fields is what required? Or heading to the library? Even in software engineering, we see people reinventing the wheel because they can't be bothered to read a computer science book.
- jltsiren 2y agoOn the one hand, this is something I also hope LLMs will be able to do. On the other hand, it seems to be something LLMs are particularly bad at. There are many similar but subtly different concepts. People use the same words for different concepts and different words for the same concepts. Every paper invents new terminology. And there is often very little written material available.
- Avamander 2y ago> People use the same words for different concepts and different words for the same concepts. Left unchecked, it will be amplified by LLMs. Detecting and ignoring the poisoned data requires way more "intelligence" than the current approaches can achieve. Not sure if it'll exactly be bad, some naming schemes are just bad anyways. Old man yells at a cloud using words incorrectly...