Let’s kick off the Hot Takes series with something extra spicy.
Hot Takes #1: Robotics companies that use ROS don’t make money.
ROS stands for Robotics Operating System, and it is basically Arduino for robots. Ok for amateurs to play around with, making the “plug the mapping into the motion planning” hello-world blinky-LED of robotics, but not really suitable for real products.
ROS2 is the same. I can already hear the objections, and no, switching to DDS didn’t fix the build system, everything being async, the nondeterminisim, the untestability, the reliability, everything being a node, or, frankly, the sales pitch of where the hard part of the problem is. [I hear some people wanted to make more drastic changes but it was shot down to improve the backwards compatibility? Well, we are where we are.]
A fun comparison I like to make is starting your own car company. Would you start with a university lab or open source car project? What about a car project by a company that went out of business? Would you keep the chassis, engine, suspension and drivetrain which were mostly used by university students to put together a car that you wanted to sell commercially? Really? I mean for a first prototype sure, but bring it to production? There are so many things it isn’t optimized for that suddenly become far more important: reliability, supply chain, maintenance, resource consumption. None of those matter when you can throw infinite hardware at a single demo.
If I were an investor in robotics companies, “Do you use ROS?” is one of the highest-alpha questions I could ask on whether the company has the technical capabilities they need. There are a few aspects to this:
- You still need the skillsets that ROS lets you bypass in the early days. If you don’t have the chops to already solve the things ROS ‘does for you’, you will fail hard later at productization. Better to fail early, or at least find out who you need to hire early to not make those later failures.
- If you hire people who only have experience with ROS systems, you start off with a culture and knowledge gap even if not a technical gap.
- Yeah the technical gap too. ROS locks you into an inefficient way of doing things, and even if you then bring on the right people they will be stuck with all the cruft from the beginning. Digging your way out of that hole is probably about as hard as throwing everything away and starting from scratch. (Although of course you shouldn’t do that.)