We need to augment our command line tools and design APIs so they can be better used by LLM Agents. The designs are inadequate for LLMs as they are now – especially if you're constrained by the tiny context windows available with local models.
No. Just no. You don’t get to change the problem to fit your inadequate solution.
If you want your silly little LLM to use a cli, you have to make it able to do so. If it gets lost in a fucking directory structure, your LLM is failing and you have to build a better one.
I think watching the agents use our existing command line utilities get confused and lost is a strong indicator that the information architecture of our command line utilities is inadequate
No. It’s a strong indicator that LLM are bullshit technology that shouldn’t be near any cli whatsoever.
No. Just no. You don’t get to change the problem to fit your inadequate solution.
If you want your silly little LLM to use a cli, you have to make it able to do so. If it gets lost in a fucking directory structure, your LLM is failing and you have to build a better one.
No. It’s a strong indicator that LLM are bullshit technology that shouldn’t be near any cli whatsoever.