Restraint

I'll go against the grain and say the most valuable human skill to develop right now might be restraint.

It's easy to get agents to write thousands of lines of code.

It's hard to:

• Keep your architecture simple
• Build automations that generate more signal than noise
• Build the right thing instead of 5 wrong things

LLMs are not going to do that for you. In fact, they'll egg you on.

And to be honest, I'm not seeing much restraint from humans at the moment. They're tokenmaxxing... throwing stuff at the wall. In many cases, corporate is demanding it.

But to make the most of LLMs, we have to stay disciplined. Agents work best when they operate on simple systems with clear patterns.

It's only through self-discipline that we can give them that foundation.