Matthew Pulsipher

I've had 10 founders contact me over the last 3 weeks about their 'AI smartphone' idea (basically a phone OS built around AI tools and customization) so I figured I might as well post here rather than explain the same thing to each one. I'm intrigued by the idea but smartphone is so core to my daily activities that switching would require a massive benefit or low investment. I believe ppl mostly use smartphones for 3 things: - entertainment (primary use, I'd include chatting with friends here) - work - personal communication A product is the sum of its features/benefits/coolness, then you subtract the friction points. A great new product feels like the pros are 10-100x better than the cons. If your product is right on the edge (barely better than the cons), it's not a standout first-gen product. Think about all the first gen products you used and how much better they felt than the status quo, even given their friction points! Okay, this is all generic consumer product startup-y advice. On to the AI OS specific stuff. Being able to get more things done and/or customize my phone to my exact preferences (with AI tools) could be very awesome. I vibecoded an OG instagram app for my wife (Liztagram) that strictly shows close friends stories and posts. She loves it and I'd love to share it with her friends but even though it's RN, it's still hard to share things like that with iOS folks. My dream is that an 'AI OS' could make it easier for folks to build and share 'apps' that interpret their OWN data in a better, more personalized way (instead of mass-market consumer apps). Beeper's a good example of this, but it took 20+ people working for 3-4 years to build. Most people are smartly starting by building a ROM that gets installed on a Pixel rather than building their own hardware. This is a good test - if you or your friends use this and have a great experience that's so much better than their existing Android or iPhones that they switch completely to using your Pixel+ROM device. Then you're on a good track. The truth is that it's harder than it looks to switch to a non-Google Android rom. The biggest issue is they don't support Google Wallet/tap to pay or some banking apps. There's tons of other little papercuts as well (remember - think about how friction works against you). Since you're probably build building for AI-pilled people in US - most of them seem to have an iPhone. If you don't have a solid iMessage/iCloud story, that's another blocker (openbubbles.app works well but is $15/m). I deeply looked into building a smartphone for smallandroidphone.com. For that project, there were two problems: - No one made a display in the 5.4-inch form factor (the best we could find knockoff iPhone 13 mini replacement screens) - It's practically impossible for a small vendor to make a good camera experience comparable to Pixel/iPhone. It's not the sensor. That's easy to source. It's the camera app. AOSP-derived OSes don't include a good camera app! Each Android OEM pours millions of dollars and time into building their own. Google doesn't license PixelCam. It's possible to make a new phone company, but extraordinarily hard. Just look at nothing - Carl was a cofounder of oneplus and even with the support of oppo he is still on a super difficult path. as much as I would love to have a small android phone or a new AI OS phone, for me at least right now, the cons outweigh the pros.