Good thoughts here, but I'm missing on the whole "steps" thing. How can you trust an agent to do the right thing if you can't audit it? and if you can audit it, can't you reverse engineer the steps? and if you can reverse engineer the steps, you don't need an agent. and if you know the steps, you don't need an agent. I'm not comfortable just handing off stuff to LLM's
I don't audit the agent, I audit what it hands back. Same as Josh: I didn't watch him code, I looked at the page. Known steps get a script. The agent only gets the work where the input changes every time.
Good thoughts here, but I'm missing on the whole "steps" thing. How can you trust an agent to do the right thing if you can't audit it? and if you can audit it, can't you reverse engineer the steps? and if you can reverse engineer the steps, you don't need an agent. and if you know the steps, you don't need an agent. I'm not comfortable just handing off stuff to LLM's
I don't audit the agent, I audit what it hands back. Same as Josh: I didn't watch him code, I looked at the page. Known steps get a script. The agent only gets the work where the input changes every time.
And if it hands back garbage, what? Tune your prompt? Sounds like…programming