“You’re making software for someone who spends 95% of their time in the field, isn’t trained in accounting, and needs to understand their balance sheet from day one. It’s gnarly — that’s what makes it rewarding.”

“Working for the real economy felt more tangible than the abstract problems AI is solving today. Being on the ground in these messy environments — literally, physically — shows you what technology actually needs to be built.”

“I’m building a self-healing loop that improves our codebase, so agents run faster and engineers ship quicker. No one’s mastered this, and no one’s published it — so I’m building it in house.”

“The technically challenging part is sheer scope — one product, so many use cases. There’s a lot of fluff in how companies talk about AI tooling. We’re empirical about what actually moves people through product faster.”