Dougan Milne

Building Typeahead from -1 to revenue changed my definition of an AI-native company. Today we’re releasing Typeahead 2.0 on Product Hunt. It is the first product from the new company I am building that people can buy. producthunt.com/products/typ… For months, I have been writing publicly about using AI to do better product and marketing work. The work has ranged from audience research to positioning, websites, launch systems, content, and demos. Typeahead is where we put those ideas under pressure. The product started with a belief. AI is going to move closer to the work. Cloud AI is already part of my daily workflow. Local AI is where I think a lot of personal software starts to feel different. That is the bet behind Typeahead. We wanted AI writing help at the cursor, inside the apps where writing already happens. Getting there took more than building the app. First, the thesis had to become positioning people understood. From there, the positioning had to become a website that could sell. Demos needed to make the behavior obvious. Launch materials had to make the bet legible. And the product had to become something people could actually buy. AI helped with all of that. It helped us research the audience, test angles, shape pages, create assets, draft copy, pressure-test the story, and move through weak versions faster. The surprise was where the work got harder. AI made more directions possible. That meant our taste had to get sharper. We had to decide what was true, what sounded like us, what customers actually cared about, and which details deserved more time. As the product improved, the launch got sharper. The loop got tighter. That is the lesson I keep coming back to. AI-native company building is about shortening the distance between learning and shipping. Research sharpens positioning. Positioning shapes the website. The website exposes what people understand. Demos make the product promise visible. Feedback changes what gets built next. Each pass leaves better evidence and a higher bar. Typeahead 2.0 is the clearest result of that loop so far. It is AI autocomplete for every app on your Mac. You type where you already work. A suggestion appears inline. You accept what helps and keep moving. It runs locally, works offline, and keeps your writing on your device. The first version proved people understood the behavior. This version is faster, more native, easier to control, and much closer to the product we wanted from the start. For me, this launch is proof of two beliefs. AI-native companies are going to learn and ship differently from day one. Local AI software is going to matter a lot more than people think. We are live on Product Hunt today. I would love for you to take a look and tell us what you think. producthunt.com/products/typ…