Richard Diehl Martinez interviews co-host Pierce Freeman about his experience building Saywhat with Will McTighe. Freeman contrasts earlier browser-automation and quality-assurance tooling with a LinkedIn content workflow whose users already felt the problem and paid for alternatives.
The lesson is less about a universally superior technology than finding a willing buyer. Freeman describes listening to users, benefiting from a co-founder's market knowledge and selling an outcome customers wanted. His account is a founder's retrospective, not independent evidence of market dominance.
Reusable infrastructure shaped the engineering approach. Freeman explains Mountaineer's Python-to-React data contracts, automatic state updates and checks for AI-written code. Type checking and linting can catch some mistakes, but they do not establish that every deployment or generated program is safe.
The product combines content inspiration, writing, publishing, analytics and engagement. A personalized writing model seeks to reproduce a user's style, with an option to turn personalization off when imitation produces unwanted results. Freeman favors practical research that clears a useful baseline over an unaffordable standalone research lab.
Unsolicited customer enthusiasm, returning users and improving economics inform Freeman's view of product-market fit. He discusses lowering hosting costs and keeping a small bootstrapped team, while treating future competition, defensibility and his next project as open questions.
Watch on YouTube
