What the World Needs: Pivoting away from Robotics

There’s an Japanese concept about one’s purpose in life called Ikigai.1 It’s built on four criteria:

  • What you’re good at
  • What you like doing
  • What people will pay you to do
  • What the world needs

I like applying this concept to finding jobs; it’s a good mental model about the tradeoffs in different opportunities, and it has illuminated my thoughts and past choices around my own career. Up until sophomore year of undergrad, most choices I made about school and my job revolved around what I liked doing: programming.2 I’m lucky enough this choice also happened to satisfy two other criteria: I’m a good enough software engineer, and software engineering pays well.

The final criteria has been tricky though: what the world needs. It was most certainly been a blind spot in my choice of career for a long time. I got some internships at Google in undergrad, and started researching and pursuing a Master’s Degree in the robotics, not because I thought the world really needed it, but because I liked the work. Robotics has a real geometry and permanence to it that other software focuses like web design or databases lack, and still has some incredibly interesting and hard problems.

However once I graduated with my Master’s, moved to Boston, and started working for a robotics startup, I had a bit more time on my hands. Slowly, I started feeling the itch of “what the world needs.”

Robotics and What the World Needs

I always found robotics to be a morally strange field. There are red-herring moral arguments about “Terminator/Robot uprising” situations, but we’re still pretty far away from the general AI needed for a Terminator situation. “The robots are going to take our jobs” is another common refrain. It might be true, but there’s a chance that it’s overhyped and the amount of people displaced by robotic workers will be much lower than expected, at least in the near (10-20 year) future.3 I’m not trying to determine if it’s true however, so let’s assume for now that it’s not. There are also more reasonable arguments, such as “Search and Rescue” projects just being “Search and Destroy,” a whole lot of surveillance issues, and, the biggest one to me, who do robots actually help?

At this point it might be useful to take a step back, and note that this is all my subjective experience and opinion, and really only focused on my experience, not anyone else’s.

What the world needs is a pretty philosophical, and usually political, question. It depends on what you mean by “the world,” on what you mean by “need,” and on the context to those two answers.4 In my opinion, “the world” shouldn’t mean the entire world, as untargeted reforms don’t seem to help many people, and not understanding enough about the context and the community that you’re trying to help means you won’t really help them at all.5

There’s enough info about productivity from scientific research slowing down, likely because we just have really hard problems now. I don’t think I’m going to become the next Ada Lovelace or Alexander Fleming and invent something as world changing as the generalized programming or penicillin. So, it’s a much better bet to find a community, listen to their problems and try to understand those problems from their perspective. I also hold that most people will improve things for themselves if given the chance and ability. The Earned Income Tax Credit is a good example of that. In short, I think that helping people help themselves is a good philosophy.

For most robotic projects in industry, the answer to “who do these robots actually help” is the rich. I tried to subvert that in a few ways. Over one summer, I designed a curriculum for a robotics class at a day camp for late elementary to early middle schoolers. While the program was reasonably successful (the instructor we hired to teach the class stayed on with the org running the camp to do more robotics classes), it didn’t satisfy me as much as I had hoped, partially because I wasn’t able to actually teach it. I imagine I’d feel more fulfilled if I had the opportunity to see the curriculum come to life in the classroom.

I have also tried diving heavily into open source software. I was a maintainer with MoveIt for a while, and mentored a Google Summer of Code student for the project. I had a general vision of software for most robots being free, or close to free. But even in a perfect world, where the software we wrote was easy to use and understand, was bug free, and had all of the features general people wanted, not just robotic specialists, who would actually have access to useful robots? Even most robotic arms are tens of thousands of dollars, and while there’s some progress being made with Berkeley’s Open Arms, I don’t think there are good incentives for useful robots to be made at consumer level costs. This makes it difficult to justify to myself that robots can be used and controlled by the masses, and not just by the subset of people with the funds to own robots as capital. That might be fine with some people, but like I outlined above, it doesn’t align with what I would consider what the world needs.

Like I mentioned earlier, I wrote this post only for myself; what sort of positions my skills and situation would lead me too. To be clear, I think robotics can be a morally good field, and I would highly encourage anyone reading this to consider a career in robotics if you are interested. There are applications in space travel like at NASA (or other orgs doing space travel that’s not rooted in a desire to escape earth), disaster technology, like the robots deployed to help with Fukishima cleanup, and wearables and exoskeletons (check out the BiRob2020 proceedings if you’re interested in stuff like this). There are other robotic applications that focus on sustainability and food generation, but the line between allowing businesses to exploit workers and helping the world gets blurry, and, more importantly, there are plenty of other capable people in robotics working on those goals, so it’s not necessarily neglected. If you are interested in any of those above fields, I’d also encourage you to be aware of the effects, both positive and negative, that your work actually has. If you’re in academia, see who’s citing you, and who’s using your open-source projects. If you’re a practitioner, think about the effects your company and industry has on the rest of the world, and the communities around you. Know how you’re perceived, and don’t brush it off as “they don’t understand.” Know that to fix societal issues, you need a deep understanding of those issues.

What’s Next

Where does that leave us? Staying in robotics really doesn’t seem to mesh philosophically with what I think is good for society, despite how much I like it. It’s a hard situation, and I have a lot of trouble trying to explain to friends and colleagues why exactly I wanted to change careers. It’s the main reason that I wrote this post. Robotics is hot right now, and will be for a long time. But “what the world needs” still dogged at me.

By February of 2020, I had made up my mind about moving out of robotics into something else, eventually. I had some thoughts about “getting better at solving human problems” jotted down in my work notebook. But the key word was eventually. I had no idea what steps to actually take, or even what problem I wanted to try to solve outside of robotics. But I knew that my next step was to find a small problem that I could engage with and hopefully, make an improvement in. That… is a topic for another post. ;)

  1. I heard about Ikigai from an episode of NPR’s Lifekit. I really don’t know any of the cultural background of the term, or whether it’s actually a common way of thinking in Japan, but I like it, and feel like it applies to me. 

  2. I may have been socially pressured to choose engineering; I group up in a very STEM-centric world, with a general “Engineering is the way” mentality. But I do like to think that I exercised a small amount of free will by choosing programming specifically. 

  3. This gets complicated with the general exponential rate of change in technology. Before the pandemic, I would have said that most robots aren’t sophisticated enough to take over most people’s jobs, but now with millions unemployed because of the pandemic, automation has been deployed at an increasing rate. I suspect that it won’t take long before I’m proven wrong here. 

  4. While I don’t agree with them all the time, 80k hours has a lot of good writing about philosophy in your career choice, and even if you don’t agree with the results of their thinking, just the idea of approaching your career as something to do good with is a great idea. 

  5. A great example of this is the failed collaboration between Boston Public Schools and MIT to redesign the bus schedule