Investing Partner, @a16z speedrun. Did some important-ish stuff at Reforge, Slack, Instacart, Runkeeper, Zynga, and Pixar.
Berkeley, CA
Joined March 2007
- Tweets19.1K
- Following1.9K
- Followers13.5K
- Likes39.5K
Pinned Tweet
Agent-native products are coming.
Every product on the internet was built for a human with eyes, a cursor, and a credit card. Agents have none of those things.
Most companies are teaching agents to pretend to be humans. That's a hack. The real opportunity is products designed for agents from scratch.
Everything inverts:
• Discovery → protocol registries, not ads and billboards
• Trust → machine-readable reputation, not brand
• Onboarding → full capabilities upfront, not a narrow slice
• Payments → spend authorization, not checkout flows
• Retention → zero. Agents switch between API calls.
30 years of human product design. Day one of agent product design.
It’s that time of year - speedrun apps are open!
SR008 APPS ARE OPEN. Up to $1M + $10M+ in credits. 12 weeks in SF, Feb through April. Built to help you land capital, customers and talent. Apply by Nov 1: speedrun.a16z.com/apply
Yesterday was @speedrun Demo Day, and in the past few weeks we've had a wave of product launches and announcements.
A roundup of recent SR007 launches from teams I've worked with most closely. Satellites, fintech, hardware, robotics, consumer, enterprise AI, we got it all. 🧵
Elu
Self-healing software. Detects issues hitting real users, generates the fix, validates it, auto-merges to production. 8x ARR in two months.
lnkd.in/p/gc7MVe-F
This is less than 20% of the cohort.
If you're building something ambitious, SR008 applications are open early at speedrun.a16z.com
I've been lucky to see this project as it's unfolded and there's so much love and thought put into this space in Berkeley - a must-do event.
Hey, so this Saturday you can come to the Grand Opening party for Tabletop Library in Berkeley all day and night. It'll be worth the drive over, you know, if you want.
@BerkeleyTTL is a side quest I've been on with @andrewmason, @tabletopgrl & david c to make a whimsical place that gets us off our phones, playing with each other, and holds board gaming up for the awesomeness that it has become recently. It's been a labor of love where we've also built a dewey decimal system for games, a board game recommender service, used a card catalog from UC Berkeley as the front desk, and a dozen other little details I'd love to show you.
tabletoplibrary.com/news/gra…
And after the visit you can go less than a block to the comic shop (escapist) and sci-fi bookstore (dark carnival) in what must now be the nerdiest single block anywhere.
Someone needs to do this same deep dive for Berkeley public schools, because whatever is happening in PA, I assure you it's more extreme in Berkeley.
Over the weekend, I happened to have a long conversation with Avery Wang, who is running for Palo Alto school board. You may know him as the parent who successfully sued the district after his son was denied access to more advanced math. The district ultimately moved his son into the class and settled the case.
PAUSD is obviously still a very high-achieving school district. But I find its approach to advanced math concerning.
The district has repeatedly put restrictions in the way of kids who want to accelerate in math. One argument is equity. Another is that offering increasingly advanced math creates too much pressure and isn’t good for student wellness.
But why is math different?
We let kids take extremely difficult classes in other subjects. We encourage them to compete intensely in sports. Kids spend enormous amounts of time getting better at music, debate, robotics, etc. If a kid happens to be really good at math, enjoys it and wants to go faster, why should we make that harder?
Is limiting opportunities for kids who happen to be good at and enjoy math actually going to help kids who are struggling with it?
I think we’re conflating two completely different problems. We should absolutely help kids who are struggling. But I don’t see why doing that requires putting a ceiling on kids who want to go further.
Fareed Mosavat retweeted
Today, we’re officially launching Galileo Space where we’re building superintelligent satellites for the future.
Galileo Space gives government and commercial customers live satellite intelligence on changes anywhere on Earth. We’re doing this by creating a real-time intelligence layer with computation done in space.
We've gotten a lot of traction with Space Force and if you'd like to learn more let's chat or feel free to check out our website: galileo-space.com.
"But the more I sit with the question, the more I come back to an almost disappointingly simple answer: I spend my days in a room full of spiky people, each wrestling with a hard problem of their own."
We hit $1M ARR at Speedrun this summer. But what I’ll miss most is the room: founders a few desks away, each wrestling with a hard problem of their own, and willing to let me learn from it. Bye bye, Speedrun 🫡
Fareed Mosavat retweeted
a trend we’re seeing right now: more companies want to own and control their models, rather than rent their intelligence from the big labs
Speedrun team Zorbe has gone from 0 to $1.6M cARR in just 6 weeks - building custom distilled models that give companies frontier-like performance while letting them own the models themselves
founders @yash_iyer21 @SadaniPranay are cooking 🔥
