Beyond ChatGPT: A Real Example of AI Solving a Hard Problem AI is starting to break “iron laws” in manufacturing. Paint has one of those rules everyone accepts: If it dries fast, it looks worse. PPG decided to test that rule with AI. They built a database of product properties, then constrained it with chemistry. A system that can search a design space a human brain can’t hold. Then they asked for a new product: A fast-drying clear coat for auto body shops. Within minutes, the model suggested a chemical combo their scientists hadn’t tried. Lab tests said it worked. Real-world result: Heated drying time dropped from ~30 minutes to ~5 minutes at 140°F Air-dry was under an hour versus roughly two hours for competing products The product launched in March and shops say it lets them squeeze in extra jobs That last part is the point. Innovation that changes throughput is cash. And this isn’t just paint. P&G is using AI to generate new scent blends for body wash and laundry products. Mars redesigned a thinner-walled gum bottle, cut development time ~40%, and saved 246 tons of plastic. 3M used AI to design a sanding disc that improves dust collection and grinding performance. Products have too many variables, too many interactions, too many “if you change this, you break that” tradeoffs. Humans search locally because we have to. AI can search globally because it doesn’t get tired, forget, or fall in love with yesterday’s answer. One 3M exec put it well: AI becomes the fourth expert in the room. This is a far more interesting use of AI than speeding up emails and decks. Product development has always been trial and error within human limits. AI removes the limits. When machines can search what people never could, new products stop being rare accidents. This still feels early. https://lnkd.in/ecJcy4s6