Garrett McCurrach (Pipedream)
The CEO of Pipedream Labs reveals strategies behind building hardtech companies fast and publicly, the engineering process, and the future of hyperlogistics.
Welcome to the 6th issue of Master Plan, a series of conversations with hard tech founders solving really difficult problems.
Garrett McCurrach is the CEO of Pipedream Labs, a hyperlogistics company starting with an underground package delivery network. Previously, he was VP of business development at Martin Bionics. Garrett was kind enough to share some insights on hyperlogistics, building fast, and some of the engineering process behind Pipedream.
Garrett introduces Pipedream and Hyperlogistics
My name’s Garrett, I’m the CEO of Pipedream. Pipedream is a Hyperloop for packages—a really fast underground delivery network.
We don’t know where the first Pipedream will be built yet, but there are a lot of interesting ideas on where the best application for one is right now. And really what we’re looking for with the first implementation is somewhere we can build quickly:
All private land
A high volume of usage
Packages aren’t high-risk
For Phase 1, we’re looking for a campus where everything being moved around is low-risk so if it doesn’t end up through our system in 30 seconds, there’s another way to get that item.
We see delivery evolving towards an end state that we call hyperlogistics. Your experience with Pipedream changes depending on how close we are to that end state.
Hyperlogistics is just like the internet, but for physical objects instead of pure information. It shares features with the internet, including:
Routing. You’ll have a physical router and nodes on the system where you’re putting in deliveries and sending them out
Bidirectionality. If there’s anything that the user is sending back, you’ll accept it through that same router. You have bi-directional interaction with the consumer at a node.
We’re not quite at that stage yet. If you move closer to where we're at now, our first big implementation will be creating portals between districts in a city.
Portals for intra-city delivery
Let's say there’s this grilled cheese restaurant—let's call it Cheesy Joe's—and they have a lot of grilled cheese sandwiches they need to get all over the city.
The way it works right now, DoorDash is dispatching a driver to go and grab this grilled cheese from Cheesy Joe's. That driver is driving five districts North to a neighborhood and dropping off that order at a person's house. They’re sitting in a two-ton vehicle and probably for 45 minutes, we have paid a human being to go and pick up a single sandwich and drop it off at someone's house.
And it's a big joke—we all know those economics are tough because I pay delivery fees that are double the price of that sandwich.
But it works for food because it's already a high-margin product and I’ll pay a premium for it since I don't want to drive all the way to the restaurant.
The way we would use Pipedream in that situation is we have these portals in each district. So if the restaurant has 10 grilled cheese sandwiches that need to go all over the city, a GrubHub driver would come to pick up all ten grilled cheese sandwiches, take them to the nearest portal station.
Pipedream moves those orders at high speed to different portals near the end user’s house where they're batched with other deliveries from around the city and a driver comes, picks up that whole batch, and then delivers it to the individual people's houses.
The portals are just to make intra-city delivery more efficient, eliminate cart minimums, and reduce the amount of time it takes to make those deliveries. We’re taking existing systems, making them more efficient, and supercharging them.
This also opens up the ability to transition cities to completely autonomous delivery.
We have technologies that could make a city fully autonomous right now—sidewalk robots, drones, ZipLine. They're just not long-range or quick enough. But if they're only interacting with those individual portals in the district, then they only need to do that last half-mile to 400 meters. And that's something that they could handle today.
So we think that's currently the most interesting use case, eventually building up to a future where everyone's connected to this network. But the five-year goal is just getting those districts connected through high-speed tunnels.
The delivery cupboard
10-20 years from now, you'll have a door in your house that connects directly to Pipedream. If everyone on our team dies and Pipedream doesn't exist, you'll still have that door.
It is a certainty of logistics that you will have this magical cupboard, where ordering something online and getting it delivered to that cupboard is just as easy as going to any other cupboard in your house and grabbing what you need. We just think we can be the first to make it a reality.
On the Lean Startup
I like Lean Startup. There’s nothing wrong with it. I think where we're at with SaaS right now, there are so many tools to build software that if you want to make $100M, you can just do the lean startup model and build a software company. That’s awesome.
But that ultimately, it's just too iterative. A lot of the innovation we need requires a paradigm shift. Hyperlogistics, by definition, is a complete leapfrog of technologies from where we are in logistics right now to where we’re going.
So the lean startup model doesn't work—we don't want to iterate on logistics. There needs to be a step-change in logistics, and you don’t get there by making existing technology 10% better.
Working backward to an MVP
We took a different path with Pipedream. Usually, we'll come up with an idea and build the easiest, quickest MVP. We can just get something and build it and get it out there. But we took a much longer, more thoughtful approach here.
We really started with the idea and the end goal in mind:
With teleportation as the end state, what's something that we could do that's a little more plausible today than teleportation, but gets you almost all the benefits of teleporting objects?
We call this new industry hyperlogistics. What is a version of that we could get done this decade?
We took a first-principles approach to building a hyperlogistics company. We started with the end goal and looked at:
What technologies are missing that would make this possible?
What can existing technologies can we stitch together?
We think that low-cost tunnels are needed to make hyperlogistics possible. And we’re building to prove our assumptions wrong right now and just making sure that everything works like we think it’s going to need to.
But if it ends up not being that, it’s okay. We can pivot. Everything is focused on just building towards that end goal.
I tried digging test tunnels in my backyard and just do like a little circle, but ultimately, my backyard is not very big, so that circle got pretty small. I actually dug tunnels out in the middle of nowhere, and just put a long 100m test track there with a trencher.
A trencher is this giant chainsaw on a little car and then you drop it down and it just chainsaws the dirt up. And you just draw it down in a line and then you put your pipe in that trench that you just dug and then you put the dirt back over it. When I explain it, it sounds super easy for one guy digging the trench. But it took me forever. It was super hard.
Everything is Storytelling
We’re definitely focusing on storytelling when talking to customers—although you still need to back up your storytelling from the engineering side. Pipedream is such a different use case than what governments are used to talking about, so it’s really about convincing them of the benefits of a system like this.
What we’re really doing is eradicating personal ownership of things and moving all tools and consumables to the rental economy, and that’s not super obvious. A lot of what we do is also detailing the ramifications of what we’re doing.
Sometimes I can tell when people are asking the right questions and there’s an incubation period of the idea where you don’t really understand it the first time you talk about it.
But if they've asked the right questions and then they think about it for a few days, they always call back in like, “Oh my goodness. I thought of like a million things that are going to change once we realize this future.”
The engineering process applied to prosthetics
I love the engineering process and that's what attracted me to work at Martin Bionics before Pipedream. A lot of the technology development process around prosthetics was centered around questions like, how do you make a knee better? How do you give it more intelligence? How do you make it aware of the ground? How do you make it communicate better with the person, make it more usable, and make the battery last longer?
And Martin Bionics took a really counter-intuitive approach. In talking to people, they discovered the biggest problem people had with prosthetics was how they fit onto their limbs.
There needs to be an interface between the machine and the body, and everyone thought a hard rigid socket was the best design. And Martin Bionics said, can we make a socket from pliable materials?
They took this really big, hard problem and they said, let's just keep iterating and designing until we find the most pliable socket we can create. And that is ultimately what attracted me to them is they were doing something really hard with a huge impact that has never been tried before.
I learned a lot about the engineering process, iteration, pivoting the business model, etc. from working there, and I took those learnings and applied them to Pipedream.
Paradigm shifts come from outsiders
If you look at a lot of the paradigm shifts we’ve had in industries like space or computing, it comes from outsiders who take a unique perspective on something experts say isn’t possible, and say “Okay, let’s try to find the most possible variant of this idea.”
Not by theorizing, but actually building it and seeing why it’s not possible.
I always love this story about the physicists who said Elon Musk landing rockets was a fool's errand and he had no experience in the space industry. I love the story about the New York Times article that said we were millions of years away from human flight, then a few months later the Wright brothers built their first aircraft.
The same thing happened with prosthetics. Jay Martin, the founder of Martin Bionics, got a grant to pursue the pliable socket idea and hired a team of experts. Three months in, they were all like, “we’re not going to be able to do this.”
So he fired them, hired a bunch of college interns, and they built it. They didn’t have any sticking points of perspective. They had a dream, a lot of hope, and the engineering process to iterate until they built something that worked.
I don’t have any experience in tunneling. And I think that’s good because we’re not focused on tunneling. We’re actually focused on making hyperlogistics possible. I think if I was an expert in tunneling and we found a reason that tunneling wouldn’t work, I would be way less likely to pivot away from that.
The only thing we care about is making hyperlogistics possible and sticking to a really fast iteration cycle in our engineering process.
Advice on advice
We were talking to a ton of experts while also building things in parallel. Experts have a lot of experience and industry knowledge. You just need to be really careful about the advice you choose to take from them.
A stumbling block for young engineers and young entrepreneurs is they reach out to experts and they take all their advice. You definitely have to filter the advice that you get from those experts and understand the mental biases they have when giving you that advice.
But there’s still a lot to learn from experts. Usually, when I’m talking to a logistics expert, they’re like, “Oh yeah, this would be great if it were possible.” Okay, that’s fair. So maybe it’s not possible, but the next question to ask is—how could it be possible?
What specific problems are we going to run into when implementing this? Help us get to those problems you’ve seen in the past. We don’t want to repeat any mistakes. We want to make new mistakes.
On finding co-founders
My co-founder and I were college roommates. I think it's a classic story. We were college roommates freshman year and stayed roommates all four years of college.
I think you can over-optimize on finding the perfect team. The metaphorical yin to your yang in a co-founder. But the most important thing about having a good cofounder is the same as having a spouse—you need good, efficient communication and a lot of trust in each other.
You're making a lot of decisions really fast. You need to be able to communicate and understand when the other person needs help or when they’re falling into an old habit and you should be able to help them with that.
I’m also honored to have him as a co-founder since he’s the smartest engineer I’ve ever met. Absolutely brilliant. We have a lot of history and trust built up. The great thing about my co-founder, Drew, is he knows so much. Not to oversimplify it, but we have a Jobs-Wozniak type relationship, where he is really an engineer and I’m a little more visionary.
But he’s also really creative. If I have an idea that’s kind of crazy or out of the box, he’s really good at saying “That’s probably not going to work, but here’s how we could break down this big idea and make it possible. Let's build it and find the reason why it doesn't work.” That’s super-valuable to have in a co-founder. He's my favorite person to work with.
Garrett Builds Fast
I was meeting a lot of entrepreneurs and people with ideas where they wanted to build this thing, whether it was an app, a website, or a physical product. And they were always waiting to raise money so they could hire a technical co-founder. Or maybe they had a dev shop that they wanted to bring on or an agency that was going to build it for them.
They had this mental barrier of “I need someone else to do what I want to do” and sometimes that’s true. I love the current state of the internet where if you want to do something, you just need to Google it long enough and the internet will bequeath it to you. The information is out there, you just need to look for it.
So I set up a bunch of calendar slots on garrettbuildsfast.com because I love brainstorming with people and saying—okay, you have this thing:
Can we walk that back from a feature standpoint to something where you can build it yourself?
Can we find a tool that will help you build this thing for cheaper than you think you can?
For hard tech, there’s a bunch of free tools out there. Onshape—an in-browser CAD tool—will let you have a hobbyist account for free. It’s really intuitive and you can build whatever you want on it. Then you can just take that design and upload it to Protolabs where they’ll 3D print it for like 20 bucks and ship it to you. And now you have a prototype.
The same thing applies to code. You have tools like Webflow, Memberstack, Airtable, and Bubble that let you build out an entire MVP. It might take a little bit to learn, but you’ll be able to get a product out within a weekend.
I love trying to figure out the hack from “I’m not technical” to “I’m going to will this thing into existence.” And I think people just need help on figuring out the first step, and it’s so much easier to figure it out when you’re brainstorming it with people.
I also have a tendency to want to build everything and those brainstorming sessions alleviate that a bit for me and help me stay focused. Talking with other people about their things helps, and I just love it if I could find out what every single person in the world is working on.
I love to hear about what people are working on, what they’re passionate about building. It’s my favorite thing to do.
There are a lot of really interesting projects from those sessions that involve interesting intersections of different creative things. As soon as you hear the idea, you’re like, “Of course, that’s going to exist someday.”
Garrett Gives Grants
I’m working on giving out grants to people working on launching things. Hopefully, it’ll be done in January.
I just don’t like how in doing some things, people get really caught up on a $40 domain name, or a certain API they need to pay for to launch what they’re doing. I want to remove that very small barrier for people to test their ideas.
So I said, “all right, we’ll pay for one month of your domain name as long as you build lodge and tasks. And if you keep hitting your KPIs, we’ll keep paying for it until you’re able to bring in enough revenue to pay for it yourself.”
Removing that $50 monthly barrier for someone launching something they’ve built is an easy way to really make a difference in a project for some people. I’ve definitely been there. I’ve had a couple of things that I’ve built where I’m like “I don’t have $30.”
Building in Public
The day I tweeted that was the day I decided to seriously use Twitter. Growing up, I was taught that it was best to “let others do the talking for you.”
So I grew up with the mentality of “I’m just going to build and put my things out there, and I’m not going to talk about it.” However, I followed KP on Twitter, and his building in public mission and taught me that this mentality was flawed. I needed to be a lighthouse for my mission and attract other people who also want to build towards that.
I wanted to pin a tweet that was exactly my mission, very succinctly, that would attract other people to it. And ultimately, that was:
I will bring 30 second delivery times to the world this decade, no matter what.
That is my decades-long mission.
Productivity
There’s so much power in visual KPIs. I have a dashboard where I set really small daily goals that are easy to hit. And then over a long period of time, if you do a little bit every day, you’ll have a huge outcome at the end of the year. The dashboard is color-coordinated so you can see at a quick glance if you’re trending in the right direction or not.
For setting goals within Pipedream, we’re moving really fast right now. We have big project goals, but we don’t have any KPIs to track. We’ve just been setting ambitious deadlines and then working backward from those. Ambitious engineers are also perfectionists, and we like to set really big goals. But then when we feel like we’re not accomplishing them, we get stuck in trying to figure out how to make them perfect.
So to avoid that, we try to set really aggressive deadlines. And if we’re not able to hit them, we walk what we were trying to get done back until we have something we can get done by that deadline.
How to get more people into hard tech
It’s been so interesting this year, just the amount of VC money and communities of support that have popped up. I think the more people that are publicly doing ambitious, hard things and telling their story, the faster those crazy ideas will go from fiction to reality.
Elon kind of blazed that trail for everyone. Not to be an Elon simp, but he was among the first to:
Become obsessed with a really big, ambitious hard tech goal
Talk about it nonstop
And, for the most part—actually pull it off
Ambitious founders like Elon have given our generation so much permission to dream big. Founders are saying “hey, I’m building factories in space” and instead of dismissing it as crazy, people’s first reaction becomes “Right on, let me know how I can help.”
The more people that feel comfortable about having those ambitious goals. It may be towards the end of the decade, but I think we’re about to see some sort of Moore’s Law type growth in manufacturing, where it’s going to be easier to build big ambitious hard tech stuff.
Side note: this is a sentiment echoed by Justin Glibert in our 3rd interview
I think if we can bring Moore’s law to the physical world, we’re going to see the same set of extremely cool companies and life-changing products we have seen in the information world. And the key to doing that is making manufacturing so much better that we won't even call it manufacturing anymore.
Things that seemed fanciful in the early 2000s are starting to seem a lot more reasonable.
It’s time to build
“It’s time to build” is something I live by. My Twitter header is “just build it.”
If you’re sitting on what you think is a good idea and don’t overthink it. Just build it, get it out there, and test it. When you get lost in the what-ifs you lose so much time making sure this idea is perfect. It’s important to just have a bias for action. Just build it is a huge life-changing ideology.
I’m in a couple of Discord channels where almost every morning, someone alerts everyone and just goes “it’s time to build.” And it gives me so much hype for the day. It’s totally time to build.
On no-code
There's so much room for no code. No-code for everything.
I wish I could rename it. I think low-code is definitely the move where you can change the frontend and data handling. We're not talking about something like Webflow, which is focused on design, but more about data handling where you have the same logical flow as code, but made visual. That’s going to have huge ramifications. The ability for someone to have an idea and build it really quickly is crazy.
10 years ago, there was a very small amount of people who could build things on the internet. And the builders had all the power because they were the roadblock between an idea and its existence. In Web 1.0 and 2.0, the builders were the only ones who were able to build their ideas, so products reflected an engineer’s ideologies. With buildings, an architect needs to find builders to build their building. But with no-code, the architect is basically 3D-printing that building and not needing to hire builders.
You will have a big shift from builders to architects 5 years from now, where it’s going to be way more valuable on the internet to be a creative than to be a builder.
When you’re able to build things a lot quicker, more people will have a say in what gets built. I think we're already seeing that this year.
Back in my old job, I built all our internal automation systems on Bubble. We had an internal set of software for all our processes. So when we acquired a new company, we just plugged them into our software suite and they were ready to work with all our systems out of the box. Not only that, but I also could teach everyone in every department how to add new features to our software.
That way, there was no bottleneck in making the software better because, for example, the people in customer support who are using the software for customer support could also go and build things onto the software suite.
There's so much room there where the person with the problem has the power to solve it for themselves.
Pipedream’s Master Plan
Our original plan was to build tunnelers, sell them to governments, and use that money to build the first Pipedream.
We discovered that we can make it happen with technology that already exists. No need to build our own tunnelers.
So the first version of Pipedream will be built with existing technology and utilize the district portal system. And then in the future, there will be one connected to every home. We call that the Loop.
I love this idea of hyperlogistics and we think it's the inevitable place that logistics is going to go.
I'm going to keep building until this technology exists. If someone beats us to it, we'll open one of the first businesses on top of it. That's going to create a whole new opportunity for businesses. I'm in this industry for the rest of my life.
Pipedream is hiring! If you're inspired by building impossible dreams, why not join the Pipedream team below? 👇Thanks for reading! You can follow Garrett on Twitter @thegarrettscott if you haven’t already. And if you’ve learned something from this newsletter, try sharing it with your friends.
If you want to learn more about how hard tech founders build ambitious companies, sign up here. You’ll get 1-2 interviews per week and learn about how founders go from a crazy idea, to hiring a team, to solid execution.
For any questions, feedback, or requests, send them to me via Twitter.