Most AI Products Aren’t Very Good (Yet)
Let’s be honest. Most of them aren’t that good. They work, but they don’t hold up. They’re usable, but not especially useful. AI has massively scaled our ability to produce output, but it hasn’t scaled the taste that differentiates great products from average ones.
Drawing on experience building and designing AI systems, I argue that taste is not a soft skill but a critical one. It is a muscle built when curiosity meets discernment, and exercised through decision-making. As making becomes easier, the challenge shifts to deciding what is good, what matters, and what should exist at all.
Amanda Baughan
I’m a UX Researcher turned Product Manager at Maincode, where I translate cutting-edge AI research into products that fit seamlessly into real-world workflows. My work focuses on advancing human–AI interaction methods and exploring how AI can reshape the way we work.
Before joining Maincode, I earned my PhD in Computer Science from the University of Washington, where I studied human–computer interaction and social computing at the intersection of design, psychology, and technology. My research explored how technology foregrounds or prevents connection online, and has been featured by the Wall Street Journal, Scientific American, and NPR.
I’ve also worked as a UX Researcher at Google Maps and AI2’s Semantic Scholar, improving human–AI interactions in research discovery, and I periodically teach design ethics at the University of Washington Information School.