and of course, provided you have the expertise to check the output, LLMs can be very useful! i just dont know why I, the layperson, am being pushed the idea of writing and releasing code that i am wholly incapable of checking

Show Post

the use case of LLM code, i imagine, is similar to libraries that ppl use all the time, but in that case you would be using libraries that at least other people are capable of already knowing before you use them. LLM code spawns a new library that no one has the opportunity to check before release

Show Post

im sure itll get better but there isn't really any degree of improvement that'll make the risk proposition to me as a non-programmer more reasonable. what if i find out a month after release that its storing passwords in plain text or some other unforeseeable issue i cant predict cus i didnt code it

Show Post

I wouldnt know what permissions it has or what files it's accessing or how it's storing user data or anything because it's fundamentally from the word go a nondeterministic coding language

Show Post

i don't know how else to explain that im literally never going to release a piece of software to the public if i don't understand the code. even if your magic box makes exactly what i want im not going to tell ppl to run my code if i do not know what it is doing to their computer

Show Post