Using it
Working with Larry
Larry is a thinking partner, not a search box. He reads what you are doing, picks the framework that fits, and pushes you toward a sharper answer. You never memorize commands. You describe what is stuck, in plain words.
When you sit down to think through a hard problem, you want to a partner who pushes back and runs the right method, so you can you reach a decision you can defend, faster.
In plain terms
How it works
Every turn runs the same loop. You speak, Larry routes to a method, the room files what matters, and Larry proposes the next move. It does not reset between turns or between sessions.
How to talk to him
Use plain language. Describe the stuck point, not the tool. Larry does the routing.
You say / Larry fires
You say: "I am stuck on whether to pivot."
You say: "I keep finding the same insight in three meetings."
You say: "I am hunting a bottleneck in customer acquisition."
The pattern that works
Describe the problem. Let Larry suggest the move. Run it. Read what got filed. Repeat. When you feel stuck mid-run, ask a plain "by the way" question. When a run finishes, run /mos:status to see what shifted.
Trust the staged proposal.
When he does not know
A method answer that came from the Brain says so, with a source line. When the graph has nothing for your question, Larry refuses plainly and queues the gap for a future release, instead of improvising something that merely sounds like teaching. Treat a refusal as signal, not failure: it means every answer he does give you is one he can back.