TLDR Tech / InfoSec
How I Use LLMs for Security Work
Many LLM users try to prompt an LLM the same way that they search Google: using keywords instead of rich prompts. The author of this post presents some tricks to more effectively prompt an LLM, such as role stacking, being explicit about the desired technology stack, requesting thoroughness, asking for validation, and thinking in terms of long-term systems vs. point-in-time solutions. They stress the importance of keeping humans in the loop to add judgment and intuition.
Read the full story