Post #3416390
2026-06-27 00:14 UTC
For those doing work with LLM agents herdr has become a non-negotiable for me during my work day. It's a good way to keep track of your agents in the terminal especially when you have several projects going at once.
I hate that I'm at this place with LLMs now but I feel like I'm keeping a decent grasp on responsible usage and minimizing errors.
My token usage is like 2% of my limit and I simply do not understand how people can max it out without their work being dogshit.
#llm #ai #tmux
Replies (1)
-
@ahau@tribe.net 2026-06-27 00:44
@brett@bitflipp.social LLM is not a superhuman. Your responsibility is not to "herd" an LLM, but to build a system that can harness an LLM's capability and potential. You, as a human, is responsible for making no mistakes. Blaming an LLM is not an option. You are not a gatekeep. You are the one who is supposed to build a deterministic system that includes an LLM that can increase your own productivity and other people's productivity. You don't know this tool and you pretend to.