People see a drone and think ‘toy’. I see a drone and think ‘the future of physical work’. And after building and deploying a fleet of autonomous drones, I can tell you, the future is a lot weirder than you think.
I’ve been building companies in Silicon Valley for over a decade. I’ve had a couple of successful exits, RemoteTeam which was acquired by Gusto, and MovieLaLa which was acquired by Gfycat. I’ve also been lucky enough to be an angel investor in some of the most transformative companies of our time, like Anthropic, OpenAI, and Scale AI. But nothing, and I mean nothing, prepared me for the challenges of building a fleet of autonomous drones.
It’s one thing to move bits around. It’s another thing entirely to move atoms. When your code has to interact with the real world, things get messy. And in that mess, I found some of my most valuable lessons. Here are seven of the most counterintuitive ones.
1. The Dumbest Drone is the Smartest Drone
We started out trying to build the Starship Enterprise of drones. We wanted it to have a dozen sensors, a supercomputer for a brain, and the ability to make complex decisions on the fly. We spent months, and a small fortune, trying to make it work. And it was a disaster.
The more complex we made the drone, the more things could go wrong. A sensor would fail, a processor would overheat, a line of code would have a bug. And when you have a two-kilogram piece of metal flying at 50 miles per hour, ‘wrong’ can get very expensive, very fast.
We had a drone that was supposed to be inspecting a warehouse. It was a beautiful piece of engineering. It had lidar, thermal cameras, and a custom-built AI processor. On its first test flight, it got confused by a flock of pigeons, tried to avoid them, and ended up smashing into a pallet of brand new TVs. That was a bad day.
We went back to the drawing board. We stripped out everything that wasn’t absolutely essential. We made the drone as dumb as possible. All it did was follow a pre-programmed path and send a video feed back to a central server. The server did all the heavy lifting. And it worked. Perfectly.
The lesson? Don’t put the brains on the drone. Put the brains in the cloud. The drone is just a tool. The real intelligence is in the system.
2. Your Biggest Enemy is the Wind
You’d think that the biggest challenges in building autonomous drones would be things like battery life, or collision avoidance, or regulatory hurdles. And those are all hard problems. But the one that will keep you up at night is the wind.
We were doing a test flight for a client who wanted to use our drones to inspect their wind turbines. It was a beautiful, clear day. The sun was shining, the birds were singing. And the wind was blowing at a steady 20 miles per hour. No problem, we thought. Our drones are rated for 30.
What we didn’t account for was the turbulence. The wind hitting the turbine blades created a vortex of chaotic air that was impossible to predict. Our drone, which had been flying perfectly just moments before, was suddenly tossed around like a leaf in a hurricane. It managed to stay in the air, but the video feed was useless. It was like watching a found-footage horror movie.
We had to completely redesign our flight control software to account for turbulence. We ended up using a combination of predictive modeling and real-time sensor feedback to create a system that could anticipate and react to sudden changes in wind conditions. It was a huge undertaking, but it was the only way to make our drones reliable enough for commercial use.
3. The More Data You Have, the Dumber You Get
We were drowning in data. We had video feeds, thermal images, lidar scans, GPS coordinates, battery levels, motor temperatures, and a dozen other things. We had terabytes of data coming in every day. And we had no idea what to do with it.
We tried to build a dashboard that would show us everything at once. It was a beautiful dashboard. It had graphs, and charts, and gauges, and a real-time 3D map of the drone’s location. It looked like something out of a sci-fi movie. And it was completely useless.
There was so much information that it was impossible to see what was important. We were so focused on the details that we were missing the big picture. We were suffering from information overload.
We ended up throwing out the dashboard and replacing it with a simple alert system. The system would only notify us if something was wrong. A drone was off course, a battery was low, a motor was overheating. That was it. And it was a game-changer.
We went from being reactive to being proactive. We were able to fix problems before they became disasters. And we were able to do it with a fraction of the mental energy.
4. The Best Sensor is a Human
We were trying to build a system that was completely autonomous. We wanted to take the human out of the loop entirely. We thought that was the holy grail of robotics. We were wrong.
We had a drone that was inspecting a construction site. It was looking for cracks in the foundation. It was using a high-resolution camera and a sophisticated machine learning algorithm to identify potential problems. It found a crack, flagged it, and sent an alert to the site manager.
The site manager went to look at the crack. It was a piece of string.
The algorithm had been trained on millions of images of cracks. It was an expert at finding cracks. But it had no common sense. It couldn’t tell the difference between a crack and a piece of string.
We realized that we didn’t need a fully autonomous system. We needed a system that augmented human intelligence. We needed a system that could do the tedious work of scanning for potential problems, and then present those problems to a human for a final decision.
We changed our system to a human-in-the-loop model. The drone would still fly autonomously and flag potential problems. But instead of sending an alert, it would send a picture to a human operator. The operator would then look at the picture and decide if it was a real problem or just a piece of string.
It was a simple change, but it made all the difference. Our system went from being 95% accurate to 99.9% accurate. And our customers were a lot happier.
5. The Law is More Important Than the Code
We were so focused on the technology that we forgot about the law. We were building drones that could fly anywhere, see anything, and do anything. We were so excited about the possibilities that we didn’t stop to think about the consequences.
We got a call from the FAA one day. They were not happy. We had been flying our drones in restricted airspace. We had been flying them over people’s houses. We had been breaking a dozen different regulations that we didn’t even know existed.
We had to shut down our entire operation for a month while we figured out how to comply with the law. We had to hire a team of lawyers. We had to rewrite our software to include geofencing and other safety features. It was a nightmare.
We learned a hard lesson that day. The law is more important than the code. It doesn’t matter how good your technology is if you’re not allowed to use it.
6. The Most Valuable Real Estate is the Sky
People are fighting over land. They’re fighting over oil. They’re fighting over data. But the next great frontier is the sky.
Right now, the sky is a chaotic mess. You have airplanes, helicopters, and a handful of drones all competing for the same airspace. There’s no system for managing it all. It’s the wild west.
But that’s going to change. In the next decade, we’re going to see the rise of a new kind of air traffic control system. A system that can manage millions of autonomous drones, all flying at the same time. A system that can deliver packages, inspect infrastructure, and even transport people.
And the company that builds that system is going to be one of the most valuable companies in the world.
7. The Future is Not About Drones, It’s About Autonomy
I started out thinking that I was in the drone business. I was wrong. I’m in the autonomy business.
Drones are just the first step. The same technology that allows a drone to fly autonomously can be used to make a car drive autonomously, a ship sail autonomously, or a robot work autonomously.
We’re at the beginning of a new industrial revolution. A revolution that will be driven by artificial intelligence and robotics. A revolution that will change every aspect of our lives.
And I, for one, can’t wait to see what happens next.
Frequently Asked Questions
How do I know which items apply to my situation?
Start by honestly assessing where your biggest bottleneck is right now. The items that address that specific constraint will give you the highest return on your time and energy.
How were these items selected?
Each item on this list comes from direct experience, either from building my own companies or from patterns I've observed across the 200+ startups I've invested in. I prioritize practical, actionable items over theoretical concepts.
Are these recommendations still relevant in 2026?
Absolutely. While specific tools and tactics change, the underlying principles remain consistent. I update my thinking regularly based on what I'm seeing in the market and across my portfolio companies.
Which item on this list has the highest impact?
It depends on your stage and context, but in my experience, the items near the top of the list tend to have the broadest applicability. That said, sometimes the less obvious items create the biggest breakthroughs for specific situations.