Your work on Mobile First makes me particularly interested in your perspective on the transition now underway. What can the shift from desktop to mobile teach us about today’s move from graphical interfaces to AI, and are we approaching a moment when “Mobile First” gives way to something like “AI First”?

Every time computing shifts platforms, the definition of "an application" changes. Mainframe apps were text terminals to a computer that wasn't in the room, desktop apps were packaged software with graphical interfaces, web apps lived behind a URL, and mobile apps had to be lightweight, touch-first, and built around location and cameras. AI is the next row in that table.

So what does mobile teach us? First, the early phase always looks the same. When mobile came around, everybody shrunk their website. When the web came around, everybody copied their desktop app or put a brochure online. I think we're still in that stage with AI, where we haven't found the final form of an AI application yet. Bolting a chat box onto your existing product is the 2026 version of cramming a website onto a phone.

Second, the better question is always about capabilities. What can you do with mobile that you couldn't do with the web? What can you do with AI that you couldn't do with everything that came before it? With mobile, a big part of the answer was things like location: on desktop we knew with 99% certainty you were in the US, while mobile could get you down to about 50 meters. With AI, natural language becomes the primary interface and interfaces get rendered dynamically based on context.

Third, these rewrites are inevitable. Everybody rewrote their mainframe app for PC, their PC app for the web, their web app for mobile, and every single person is going to rewrite whatever app they have now for AI.

So yes, I think "AI first" is the right successor. But the two philosophies point in opposite directions. Mobile first was mostly about constraints. Going from a 1024 by 768 screen to a 320 by 480 phone meant losing 80% of your screen space, which forced you to do the really hard thing called prioritization. That limit made products better. AI flips this. If mobile first was "embrace the constraints to find focus," AI first is something like "embrace non-determinism to find value." AI can do a lot, so you start big and work backwards toward what's actually useful.

My practical advice is the same one I gave during the mobile transition: you don't see the true potential of a new platform unless you dig in, play with it, and try to build things. The people who figured out mobile weren't the ones shrinking websites. They were the ones asking what a phone could do that a desktop couldn't.