I don’t worry that an LLM “writing” code will introduce costly subtle mistakes. It’s that people believe it writes. What it actually does is regurgitate existing code it has indiscriminately spidered. I worry that it does not INTRODUCE costly subtle mistakes. And that it perfectly performs the task to which it was set.
=> More informations about this toot | View the thread | More toots from integerpoet@sfba.social
text/geminiThis content has been proxied by September (ba2dc).