Excited to share that I’ll be sitting down with @mehul , co-founder of @maticrobots.
I have deep respect for people who actually ship... who go through the long, hard journey of building something real.
Nobody wakes up one day and just decides to buy a robot. There are years of work behind making that even possible. Looking forward to going through that journey with him.
Congrats to Mehul and the whole Matic team!
What would you want to ask him? Drop your juicy questions below!
The founder behind Matic has been working on computer vision for over 20 years.
Navneet Dalal's research has 57,000+ citations, and his 2005 paper on human detection alone has been cited more than 50,000 times.
Now he's using that same expertise to build a home robot that can see and understand the world around it.
You can point at a mess, say "clean this," and Matic figures out exactly where you mean and goes there.
Pretty cool seeing decades of research finally turn into something you can have running around your house.
Reimaging home robotics means going all the way back to first principles: sensing, movement, behavior, and what your robot should actually understand about your home - and then doing the hard engineering to make it all possible on-device, at scale and across a live fleet.
My favorite part of being at Matic the past few months has been that every engineer is also a product person.
All the engineering we do is in the service of building a more delightful experience.
Matic cues is just the latest example
In my 3 years at Matic, purely through weekly software updates, we have unlocked edge cleaning, toekick cleaning, clothes, mirror and vent detection and so much more. We have sped up the robot by more than 2x!
Matic Cues is the next huge unlock also delivered through a pure OTA update! You now won't need your phone for 90% of your interaction with Matic, and it unlocks some crazy capabilities like point to clean 👇 and some easter eggs (try follow me 😉).
It's the only gadget on the market that will save you time ⏳ instead of wasting your time prepping for its task.
A while ago, we were working on product reliability. For context, this specific test ran for more than a month. It tests 1 failure mode of 1 part. Each part has multiple failure modes. There are ~700 parts in the robot. There is a huge cost to a reliable product and nobody wants to pay. That’s why other robots keeps breaking.
I’ve spent the last few years obsessing over parts, suppliers, builds, and problems most people will never see. Seeing Matic out in the world makes all of that feel pretty damn worth it🚀
A lot of people ask why I left Waymo for Matic. To them, self-driving cars feel like a much bigger challenge than a robot vacuum.
I used to think the same way, until conversations with Navneet and Mehul changed my mind. The chaos inside a home is wildly underrated as a hard problem, and that's exactly what pulled me in.
Physical AI is here, and Matic is one of the first edge-computing household robots that will actually live in people's homes at scale. Huge respect to the team for staying so customer-obsessed while building it.
@maticrobots is a product of love.
When I decided to leave Tesla 8 years ago, I didn't have any intuition about AI or Computer Vision. I called someone who's known in my circles as a robotics expert at the time. After listening to this company's offer and the vision, he told me it won't work. I didn't listen.
My roommate at the time, a sharp and savvy engineer, said this is crazy, why would you leave your dream job to go do this thing. I didn't listen.
I sat down with someone a long time friend years later for lunch and discussed what I'm doing. Being a seasoned investor, she was very curious. The more I told her the more she was aghast. Why would you do something like this? There's no ROI on your time. If anything, at least put Lidar. Be like Waymo, not Tesla. I still didn't listen.
Why? The answer is simple. I fell in love with the vision, the idea and the problem we wanted to solve.
Happy to share this with the world and keep working towards more.
It took us 9 years, 11 prototypes, and our life savings to learn that every great product has the same design process.
Matic is now decisively the best home robot for families (I'm biased)
A 500-word thread and video on the Universal Design Process behind every great product:
A robot that does EVERYTHING sounds nice, until you realize there's not even a human that does EVERYTHING.
There's plenty of specialization in human bodies and intelligence, even within physical labor. The tools are specialized and so are the consumables.
The way to home robotics utopia is to solve one problem at a time, and to solve it really well.
Everything else is lazy design thesis.