You’re absolutely right, that was indeed a friendly American f-15 plane we shot down. Would you like me to make a note not to shoot friendly planes in the future? ⚡ 💪
Heh, amateurs. You do not put such important orders in the prompt or other context that gets constantly compacted (and thus fuzzy). Fix your LLM client.
(Or… Maybe, I don’t know, put some guard rails in cold hard traditional non-vobe-code to begin with?)
Let me guess, Claude said to shoot them down.
Heh, amateurs. You do not put such important orders in the prompt or other context that gets constantly compacted (and thus fuzzy). Fix your LLM client.
(Or… Maybe, I don’t know, put some guard rails in cold hard traditional non-vobe-code to begin with?)
Had to be Grok or ChatGPT. Claude got fired from the DoD because autonomous target acquisition was one of Anthropic’s redlines…