Something new

I left Google last month to start something new. Allen Hutchison and I are cofounding Vycari, a startup focused on building agents. We’re opening up to testers in the coming weeks, sign up here if you’re interested in early access and providing feedback. I’d be so grateful for your thoughts and experiences.

This is the best time to build

I’ve been working in Silicon Valley for over twenty years - I moved here in the post-dotcom doldrums and felt the lingering malaise from the crash. I’ve since watched multiple hype cycles bring us up and down, and yet, what we’re seeing today is bigger than what I’ve seen before.

This is the best time to build I’ve ever seen. I couldn’t be more excited about what the future holds. All I see are greenfield opportunities for new types of experiences that delight us and help us in ways we never considered. If all frontier progress on models stopped today, we would still have enough intelligence right now to play out years worth of innovation across the entire economy.

Lest you think this is some kind of coded selfish optimism: innovation in this case doesn’t mean displacing humanity. The consistent thread I’ve observed through all of the fear is that people and their judgement and taste becomes more valuable as model intelligence increases. On top of that, I believe the tools we have today will help us connect to other people better than we’ve ever been able to, enabling more human experiences together.

We can design a better world

I keep coming back to John Collison’s incredible observation: “things around you weren’t just always there; people made them happen”. The capabilities we observe with models today change the entire nature of design for what we can build for people.

For nearly fourteen years at Google I worked applying AI into products used by billions of people. I was a founding PM on Google Photos and launched our search functionality, including face grouping and object recognition. When we launched object recognition we had a fledgling ability to recognize things in photos (it was an active area of research) and we intentionally designed our product for success - when you came to Photos instead of asking you to think of something to search for which we may or may not get right, we showed you the tile of photos that matched objects we were good at! Because of the foundational research breakthroughs, we designed a unique product experience that caught on with billions.

We also extended that object intelligence with location intelligence, thanks in no small part to Allen Hutchison who I met in that project, and partnered with to extend location intelligence in innovative ways that nobody else had imagined.

I observed this time and time again at Google: when foundational research gave us new capabilities, teams that were able to understand how that breakthrough changed product design had the most success in building something users want. To quote John again: the world is a museum of passion projects.

We can do it again

We’ve captured lightning in a bottle a few times, and we can do it again. Last year Allen and I were leads on the founding team of Gemini CLI, and for the first time I saw how agents would impact our world. If you can build the correct form factor with a good-enough model married to tools and skills you can work magic! You can help people do more, be happier, and find joy in creation.

This is ultimately why I decided to leave Google. We can do more. I grew tireless watching the speed of the world and I itched to get more deeply engaged in building the future. This wouldn’t be possible without my long-time collaborator Allen - he’s the best engineering lead I’ve worked with, pairing a genius for design and architecture with lightning fast execution. You can read his thoughts about this new move here. It’s just the two of us together now but we are running like never before.

Come along

We’d love to have you trying out what we build if only to tell us how to be better. I know a lot about how to infuse AI into products, but I don’t know what you think and how you experience the world, and you can teach me a lot. I’ll be writing more about this on my Substack - as we get to a launch, I look forward to sharing more of my thoughts on product design and choices in the new world.

If nothing else, I encourage you to also build. There is nothing more human than creating something for the joy of creating. Now is your time.