In 2010 I left MIT to found my first startup, a Platform-As-A-Service called Nodejitsu. Since then i've had some interesting successes and failures. But I've been working on infrastructure and developer tools for over a decade. But i've been involved in P2P communities for over 15 years.
I avoid that sort of thing. I try to listen to people objectively and decide if what they're doing and saying makes sense. But at the end of the day, founders are just contributors. And a founder is nothing without a team. So if I can say what teams I look up to, I think Tailscale, Oxide Computer. These are remarkable teams. I definitely get inspired by their work. They're solving hard, meaningful problems that drive the field of computing forward.
3. What book has had the greatest impact on you as a founder? (this helps us get to know what makes you tick!)
Probably the Improvised Munitions Handbook or the Anarchist's Cookbook. I found the munitions handbook at a garage sale, it was vintage and from the 50s. Years later I traded it to a street punk for the Anarchist's Cookbook. These things are cute compared to what's online these days. But as a teenager these books fascinated me because they made me realize how anyone can learn, teach, and build anything.
Our ah-hah moment came when we started seeing these "web3" products. We said, wow, these are bad. They aren't even decentralized. They're just SaaS products and investor subsidized EC2 nodes on AWS, and that's a problem, it forces most of these products into negative pricing... We thought ok, people wan't to create a decentralized web, they also want to be liberated from the cost of the cloud. Let's make it happen.
We told some Web3 investors about what we thought the problem was, and they said ok, here is 3.5M. Get to work. We went heads down and coded. Here we are today with a whole new open source runtime that allows WEB DEVELOPERS to write apps that CONNECT DIRECTLY to each other.
The way it works is, developers use HTML, CSS, and JavaScript, to write cross-platform, native apps on desktop or mobile. We expose easy to use P2P APIs built on UDP and Bluetooth. So after about a year and a half of coding, now you can really build decentralized apps.
5. What market category are you in? Are you creating a new category or transforming an existing one?
I think this category is called Ubiquitous Computing. It's the next, more interesting chapter in computing. Peer-to-peer is a whole new class of software and it will require a whole new class of developer tools. We're the company making them.
In a market down-turn when everyone's stock is in the toilet, sage c-suite execs start asking where the burn is, and cloud is a huge burn. Supplementing Cloud infrastructure with P2P can bring Cloud costs to zero. But this is just a foot in the door. There's just the start.
When the cloud was created, there was a fraction of the hardware in the world, and AWS where there to provide high availability. Today they're doing more of the same, they have roughly 38 Data centers in 87 Availability Zones in 27 Regions. That's about 50-80K servers in each data center, so maybe 3M servers.
But all that absolutely pales in comparison to the compute and storage in the wild. There's ~15B mobile devices, ~3.5B laptops, countless IOT devices, there's sensors everywhere.
The more hardware there is, the more it will be connected and the less it will be making round trips to the data center. Instead, all this hardware will form large, ad-hoc peer-to-peer networks. And the cloud will be used mostly for super computing tasks.
It's important that to clarify, because there's going to be a lot of sweaty nerds in the room saying you can't compare a smartphone or laptop to an AWS server. And, of course you can't. None of the hardware in the wild is a 1:1 replacement for a server. If you try to monopolize a user's resources and act like a server, thy'll just delete your app. But it's unnecessary because even at a small scale, short bursts of connectivity, and a tiny bit of compute and storage adds up.
Imagine if everyone scrolling Instagram was forming a peer-to-peer cdn, you could cut bandwidth, storage, and caching costs significantly. Imagine people scrolling Twitter were do something similar. Now imagine if these networks could piggyback off each other. Now imagine your average web developer can do all this. That's what we have created and this is the problem you can solve today.
Public sector, startups, enterprises, etc. Anyone who builds software or web apps is interested in what we're doing. But recently we've been talking to folks in the German government about their expensive and difficult to audit cloud solutions. In the US we know this as FedRamp. Most countries have something like this. Is's expensive and takes time. It's added cost and complexity that can be prohibitive. Now imagine that you don't need the Cloud at all. Imagine your IT org controls everything. Suddenly you have autonomy and there's a smaller surface area and less potential for incidents.
MetaMask is actively exploring re-writing their apps with the Socket Runtime, We've had discussions with Bloomberg about how bloated Electron is and how we can replace it. Folks from Signal's desktop team have been actively evaluating it. We're talking to a lot of smaller companies too. Just recently we quietly open sourced the runtime, we're just starting to talk about it.
9. Startup funding is booming and there’s a lot of noise out there. How are you rising above all this noise?
I don't really notice the noise. I don't really care what people are talking about, only what they're doing. A little over a year ago we raised 3.5M in pre-seed from about 4 investors, it happened fast. We also have a low burn, about 30 months of runway so we don't think there's an urgency to raise, we still take calls from people and we've slowly soft circled about 10M in follow-on for a seed.
10. Going to market with an innovative product is never easy. What’s been your greatest challenge? How’d you overcome it?
Bottom line matters. As you scale, Cloud gets less reliable, more expensive, and more complex. The inverse is true about P2P. As you scale, it gets more reliable, less expensive, and it's complexity is a fixed-constant. Cloud is unpredictable, and anyone with any experience knows that, so people are looking at this and thinking ok, it's low risk, let's give it a try.
It's really exciting to follow that little white rabbit. Every day I get to see the very edge of the computing universe, and it's expanding really fast. That's a lot of fun.
In a surprising plot twist, our tools and open source will have significantly displaced cloud services and the things built on them. Where you once paid rent on a cloud service, you'll be spending that money developing features instead. You'll be free of vendor lock-in, and you'll have regained the autonomy you lost.