| ▲ | kragen a day ago | |||||||
It sounds like you have tried both of these, but most of us haven't. What results did you observe? | ||||||||
| ▲ | kingforaday a day ago | parent [-] | |||||||
The intent is different and as the model interprets what your intent is it must make inference decisions on how to proceed toward your intended action. Security and reliability have always had a fine line. Whichever you choose, just be mindful on the context and other associations from past behavior. Guardrails will mature but they will likely always remain an incredibly hard middleware policy management system. | ||||||||
| ||||||||