Two beliefs guide my work:
Building is thinking. You don't truly understand something until you've built a version of it. I've carried this from homemade breadboard CPUs to billion-user products. As AI drops the cost of building to near zero, thinking and building become the same act. That's why I still build things in my spare time.
The phone and apps will be deconstructed. The phone's monolithic screen and grid of apps exist because humans need direct manipulation, but AI and AR are changing that. Most people assume this means voice interfaces, but consider how terrible a McDonald's drive-thru would be without the physical menu. Voice and screens dynamically working together is the better path. But it doesn't stop at the interface. Instead of tapping through five different apps to get something done, you'll say what you want and an agent will do it, checking with you only when it matters. At Meta, I've spent the last several years building that vision, first on AI glasses and now on agent-first ways of working.
Before
Mid-gutting