I use large language models every day. Not as a shortcut but as a thinking tool.
There is a version of this disclosure that goes: "I used AI to help write things, but the ideas are mine." That is true, but it misses the more interesting point. The reason I use LLMs is not to produce more output faster. It is to think better.
Here is what that actually looks like in practice.
Brainstorming without an audience
Every idea I have starts somewhere private. Most of them are half-formed. A good LLM session lets me put an unfinished thought into the world without the social cost of being wrong or incomplete.
I can say:
"Here is my rough take on why most founders misread their first market signal. Push back."
And it will. Hard. From angles I did not anticipate. That is not something you can do at scale with colleagues without burning credibility or time.
The result is a stronger position before I go public with it.
Multiple perspectives, on demand
The most valuable thing an advisor does is not give you the right answer. It is help you see a problem from a frame you were not using.
LLMs do this well. I can run the same argument through the lens of an operator three years in, a CFO doing diligence, a first-time founder who has not failed yet. Each produces different objections. The objections reveal the gaps. That is the work.
A single trusted advisor has a fixed perspective shaped by their experience. Useful, but bounded. An LLM gives me many perspectives without the agenda, the relationship stakes, or the scheduling friction.
Here is the actual skill I built and you can try it out too!
Stress-testing before publishing
I do not publish anything that has not been challenged first. An LLM pressure-tests faster than any other method I have found.
I use Claude because it is more willing to argue the other side with low sycophancy. It will find the weak seam in a case I was sure was airtight. It will tell me when the structure of an argument is sound but the evidence is thin.
That kind of pressure makes the work better. Every time.
What LLMs do not do
They do not replace operating experience. They cannot tell you what it feels like to run out of runway with payroll due Thursday, or what your gut knows after twenty years of watching founders make the same three mistakes. They surface the shape of an argument. They do not supply the judgment inside it.
The operating experience in my writing is mine. The positions I hold are mine. The conclusions I am willing to stand behind are mine. LLMs helped me get there faster and think harder along the way.
That is the right use of the tool.