Rodney Brooks who founded Rethink Robotics has definately an interesting story. Especially knowing that Rethink Robotics had initially raised money from Jeff Bezos, Goldman Sachs, something like 150M I believe. But despite all those high profile backers, it had to shut down in late 2018 i believe and was acquired by the Hahn Group, a German automation company.
This is a great article. I always wondered why they didn't succeed. I remember wanting to apply for a job there, as they were apparently developing a visual programming environment to program the robots. But that ended up being right before they closed their doors.
The lab i works had had a rethink robot for research purposes. It seem to be part of a wave of consumer-facing robots that have all since folded. With the exception of Waymo, DJI, and skydio. You have to remember at the time eveyone was searching for the next big computing platform, and it wasn’t that clear the iPhone would be it. After the iPhone succeeded, hardware investors became even more convinced, especially when money was cheap.
I worked at a robotics startup for a while developing a robotic arm. The big problem we kept coming up against was that the robot was just part of the story. A robot arm doesn't solve a customers problem - a robotic solution does. Programming robotics is effectively software development - it requires logical thinking, branching logic, variables and often ends up being quite complex. It's a similar problem to no-code and visual programming - the complexity of software development isn't in the medium of code, but the constructing of logical systems.
The robot arm then becomes a piece of a much larger engineering puzzle - and so the people using it need to be highly technical. All the "user friendly" gimmicks are just obstacles in their way.
I think what your parent comment means is summarized by something said by, i think Colin Angle (iRobot CEO), about iRobot: "to become successful, we had to become a vacuum cleaning company, instead of a robotics company".
Nobody cares about what's inside the black box you are selling, as long as it solves your problem. Doesn't matter if that is vacuum cleaning with either a robot or a human, getting stuff from a to b with either a compliant arm with fancy computer vision or a human or custom, inflexible automation.
I'm currently in a robotics company, we make various mobile manipulators with some internal storage. Got acquired by, of course, a logistics company a couple years ago to be part of their overall solutions.
I think the other comment pretty much captured it. If you don't make a complete plug and play package, your robotics are just going to be part of a bigger solution developed by another business. The nature of putting together a complete process automation solution is that it's a highly technical job which takes a lot of time. Your customers are not the people working around the robot - they are the automation designer / integrator who is building the solution - and they are far less bothered about swish UIs and safety features than they are about good APIs, physical repeatability, good hardware availability - all the things OEMs care about when selecting a component in their system.
We found a lot of the time people "wanted" a robot arm, but then when they got it they sat it on a shelf - because the work required to turn a robot arm into a useful automated process is massive. Having a cheaper or easier to use robot arm doesn't stop you needing fixtures, jigs, conveyor belts etc.
8 comments
[ 4.4 ms ] story [ 24.1 ms ] threadIt seemed like a nice cobot, but with no real practical profitable application
The robot arm then becomes a piece of a much larger engineering puzzle - and so the people using it need to be highly technical. All the "user friendly" gimmicks are just obstacles in their way.
And what were some of your other conclusions having worked on an arm solution?
Nobody cares about what's inside the black box you are selling, as long as it solves your problem. Doesn't matter if that is vacuum cleaning with either a robot or a human, getting stuff from a to b with either a compliant arm with fancy computer vision or a human or custom, inflexible automation.
I'm currently in a robotics company, we make various mobile manipulators with some internal storage. Got acquired by, of course, a logistics company a couple years ago to be part of their overall solutions.
We found a lot of the time people "wanted" a robot arm, but then when they got it they sat it on a shelf - because the work required to turn a robot arm into a useful automated process is massive. Having a cheaper or easier to use robot arm doesn't stop you needing fixtures, jigs, conveyor belts etc.