Post by Ralf Gehrig

"I could have drawn the Nike swoosh on a napkin in two minutes." That line came up when I sat down with Paul Wright on the Link Through Social podcast — and it's the exact trap most teams are falling into with AI right now. Here's the pattern: someone opens Claude or Lovable, mocks up a polished-looking design in ten minutes, and concludes the big team, the process, the time — none of it was necessary. It looks finished. So it must be good. Brand designers have fought this exact misunderstanding for decades, long before AI existed. People see the Nike swoosh and think "I could have done that on a napkin." They're not wrong about the shape. They're wrong about where the value actually sits. AI hasn't changed that equation — it's just made the "looks finished" trap much faster to fall into. Polish now arrives in seconds. Judgment still doesn't. At WongDoody, we've been building AI directly into our design process instead of bolting it on. We coded sidekick agents that sit on our Miro boards as partners, not shortcuts. One is a "contrarian" that pushes back the moment a team locks onto a single direction too early. Another checks for bias baked into the research and thinking — because every team carries one, usually invisibly. That's the real difference: AI replacing judgment versus AI sharpening it. If vibe-coding was really enough, we'd have seen a wave of 16-year-olds building unicorns on a Saturday afternoon by now. We haven't. The problems worth solving were never the easy part. Full conversation below — including a genuinely counterintuitive economic idea (Jevons' Paradox) that changed how I think about AI and jobs. Does a polished AI-generated mockup make you trust the output more, or less? https://lnkd.in/eM3uj7bk

Video Content