It represents a groundbreaking advancement in the field of LLMs However, privacy and security on Deepseek remain a hidden concern.

I tried to use it in place of a different LLM we were using, but using the same prompt. And it seems to give a different output. Not sure if it needs to prompted differently?

If you run it locally there should be no security concern, some of the distilled models can run on regular laptops.