3 ms·
I mean it's almost inherent to LLM's right? Like they only know about version before it's knowledge cutoff. I guess it's a big argument for not putting exact ve
by im3w1l 1y ago
I mean it's almost inherent to LLM's right? Like they only know about version before it's knowledge cutoff. I guess it's a big argument for not putting exact versions in files generated by LLM, only major (+minor?)
- jaimefjorge 1y agoYes. My point is that because of training cutoffs it should be mandatory to run SCA scans when dealing with AI code generation. Not putting exact versions would be a good idea. But that’s not what’s happening today.