Decision Made: Leaving Meta
After four years building VPS, I've decided to leave Meta. The reasoning, the emotions, and what comes next.
I've decided to leave Meta in September. Four years, a platform built, a team that no longer needs me. Time for the next chapter. Eighteen months ago I wrote "I am not leaving yet." The yet was carrying more weight than I let on.
Why Now?
The work is in a good place, which is exactly the problem. VPS is running, scaling, and valued, and my job is no longer to keep it running - others can do that. Meanwhile the learning curve has flattened: year four taught me less than year one did, and most of what I do now is optimizing rather than pioneering. Add the AI moment, which I think is genuinely transformational, and staying put starts to feel like watching from the stands. I also still have energy for a new challenge, and I'd rather leave with energy than stay until it's gone. In December I said I was thinking seriously and had decided nothing. Six months of thinking produced a date.
What I'm Leaving
A working system that serves millions of queries daily, reliably; the core-infrastructure claim from the Quest 3 launch has become a boring fact. A strong team - 75 people who don't need me to succeed. Career capital, since Director-level at Meta is a strong position to walk away from. And certainty: I know how to do this job well. Written out like that, it reads like a list of reasons to stay.
What I'm Seeking
What I want next is AI-native building: products designed around AI from the start, as opposed to AI added to existing products. I want smaller scale - more direct impact, less coordination overhead - and problems that are actually unsolved rather than optimization passes on solved ones. Mostly I want ownership. My decisions, my outcomes.
The Options
- Big company AI team: Google, Apple, OpenAI - similar scale, different problems
- Startup: join or start something new
- Independent: consulting, investing, building
I'm leaning toward independent plus building. I want the flexibility to explore before committing to anything.
Emotional Processing
There's grief in leaving something I built from scratch, and excitement about what's ahead, and a fear that keeps surfacing: what if I'm wrong? Under all of it, gratitude for four transformative years. I'm trying to sit with the feelings rather than rush through them.
The Handoff
I have three months to transition. That means documenting everything that lives only in my head, transferring relationships, closing out in-flight projects, and saying proper goodbyes. Leaving well matters - for the team, and for my own integrity.
What's Constant
Whatever comes next, the spatial computing expertise travels, the AI/ML skills are universal, the leadership experience applies anywhere, and the network persists across roles. The foundation is solid. Now to build on it.
Last quarter at Meta begins.