@fristi Using LLM's *at all* is completely retarded - as trying to use a chatbot for any practical use is insanity.

While a LLM can be used to degenerately copy code badly, the end result is always unmaintainable slop.

Ah yes, the ideological issue of not wanting software to break for the dumbest of reasons (a massive idiot turned it into proprietary slop with a LLM).

Maybe LLMs can be used to find security vulnerabilities, but there seems to be more noise (fictional bugs) than signal (actual bugs).

What happens when the noise to signal ratio gets worse, due pretty much all-LLM-findable bugs being found (just one more prompt bro) and the company hosting the proprietary software and SaaSS comes to collect the actual costs of providing the disservice?