The strategy James da Costa outlines is exactly why the next wave of category defining companies won’t come from adding random AI features to your legacy stacks. They’ll come from serving new companies at their formation, embedding helpful intelligence into those company's

Dropped out of Google to go back to Stanford

The team that nails seamless scene transitions will carve out their own market in AI filmmaking. AI can make stunning clips, but it still struggles to make them flow together. The great AI filmmakers are skilled at this. I couldn’t get a natural transition from shot 1 to shot 2

Great read. For AI-native startups especially, cinematic launches and waitlists aren’t marketing gimmicks. They’re how you test narrative-market fit (before pmf). You’re learning whether people care enough to sign up, share, or associate themselves with your story. It's probably

Students are learning to build with AI by default. They're getting a head start, building the defining skill of the next generation of great founders + employees + builders.

The first generation of AI apps has showed us what’s possible. But most are toys. Fun & impressive demos that got attention, not retention. Few have been sticky. The next wave must evolve from entertainment to utility to be successful. True winners won’t throw AI onto old

Takeaway for me from this piece from James da Costa and Alex Rampell : The PC made Microsoft. The internet made Google. Mobile made Apple. Each product cycle redefined what dominance meant not by destroying incumbents, but by changing the terrain they stood on. AI will do the same.

I posted this meta glasses video of me working the friers while volunteering at a local festival, and one of my classmates asked me what my Veo prompt was 🫢

Last week at Stanford, we met with the co-founder of PixVerse , Jaden X , to talk about building consumer apps in the AI space. They recently were ranked #25 on a16z’s GenAI consumer mobile products list. But, what surprised me most from his talk wasn't about their product:

It’s wild how easy it’s becoming to build real, playable, (actually fun) games in-browser...no code, no engine, no asset sourcing, no rendering. I used Gambo AI and built a fully playable side-scrolling shooter in minutes using nothing but a prompt. As a teenager, I'd spend

Ratemyprofessor but for cofounders

Clarity of your vision is the hardest part of your story to tell, but the most important.

Really glad Justine Moore brought up the 'Japan phenomenon' here and wish there had been a bit more convo about it. When we launched Pixel Studio in Japan (mobile app built on Imagen), we saw tons of new use-cases & user needs there that we wanted to address (eg stickers >

Being mean to AI makes it more accurate. The AI revolution will not be polite...