
Sr PM Dir, GenAI @Google Workspace. Previously: @OpenAI, @Google, @Twitter, 3x startup founder with varied success. Better half: @ericakochi
San Francisco, CA
Joined March 2007
- Tweets13.3K
- Following2.5K
- Followers23.2K
- Likes15.2K
Bay Area dads in 2016: Asking if you watch Silicon Valley
Bay Area dads in 2021: Asking if you’re moving to Austin
Bay Area dads in 2026: Asking if your coding agents are running right now
Build loops not point improvements
I'm now building an "improvements" feature into all of my personal vibecoded applications: a freeform text input that's hooked up to an agent that yolo implements whatever is submitted.
It's an interesting experience -- the activation energy barrier to sanding any rough edge or implementing any new feature becomes extremely low. Someone else who uses one of the applications I created has already implemented 150 improvements over the course of a few days. This pattern feels like a new kind of tool -- liquid software? -- and every application that doesn't have this feature now has a slightly rigid and anachronistic flavor. Why can't I just tell it how I want it to be different?
Now, it's not obvious to me how this pattern can or should scale. Does it become unmaintainable after a while? (Maybe model intelligence improvement outstrips my feature appetite such that it's fine?) How does collaborative software work in this model? Can software designed this way be secure and reliable?
Whatever the larger questions about broader deployment, it definitely works great in a personal context. It reminds me a little of the visions of Lisp/Lisp machines, Smalltalk, and HyperCard -- systems that the user can take control of and modify. Those visions never quite worked, though: the vast majority of software today is not end-user modifiable in any useful way. So, while this "liquid" affordance is a very small touch, the way it changes one's relation to the software feels a bit transgressive and empowering in a Lickliderian and Engelbartian way.
PSA for my VC friends: All the returns might be below the top 10 items on this list.
I redid the startup founder signals that lead to top tier VCs investing in the seed / A, after incorporating some feedback.
This added ex-frontier lab and cyber security near the top, IIT as a positive, and educated in China as a negative.
The main changes were;
- extended founding year to 2019-2026, ~2x'd the number of startups
- unified and added some signals
Here are the new signals:
Sunday AI run club with @richardprice100 - Today we talked about the end of being “in the zone” because we are all just managing agents.
RIP Patrick Stewart
Patrick Stewart, who will be remembered most for his roles as Captain Jean-Luc Picard on “Star Trek: The Next Generation” and as Professor Charles Xavier in the “X-Men” films, has died at 86.
variety.com/2026/tv/news/pat…
Catching up with @jans who is building Roma in the current YC batch.
We talked about how the bottleneck is no longer coding.
It’s having enough time to actually try the things the agents have coded.
My wife @ericakochi makes Japanese-made men’s shirts in US sizes. Her new collection went live 6 weeks ago.
Some things she has learned so far 🧵
I clearly need to post here more because these are the orders that have come in. All just one shirt each










