Mastodon Feed
dysfun@treehouse.systems ("gaytabase") wrote:
LLMs generate far fewer bugs in absolute terms, but they also generate far more code, so it probably balances out if everything is equal.
uhhh. well i agree on the latter part, but fucking hell do you know how to spot a bug? 😂 cause like i do and i have done a fair bit of evaluation and i still believe they cannot write code for toffee. in fact that's really too mild, they just shit all over your codebase.
Where LLMs win though is the severity. An LLM will write buggy code in the sense that it doesn't do what you wanted it, but not so much in the sense of security or data loss bugs. They get that right more than humans because they're trained to care about that stuff.
they might get this right more than humans who are bad at programming, but i can assure they do not get it right more than i or many of my friends do 😂