02 Jun 2026
keywords: motivation, goal setting
back to blog
A good friend of mine gave me the advice that the hardest part of writing a blog is not the first post, but the second. Unfortunately, he was right… so after more than a year of sitting on my thoughts, I once again try.
Not that this is the first attempt; I have been sitting on this for an unreasonably long time (knowers know), which contrasts with the motivations I described in my last post. I think my first “one hour” drafts for this were a bit too raw, to the point of oversharing. I also think my functional perspective on writing has shifted towards “what unique insight can you add to the melting pot?”, and drawing on more of my recent experiences to nuance my conclusions fits this goal well.
What’s changed since then? More than 2 years later, I’ve finished my Master’s program at Stanford, where I learned from really amazing minds at Stanford REAL. I also tasted the robot learning shift in industry while spending time with 1X and now Tesla Optimus, where I’ve been focusing on robot learning from human data. It’s been intense so far, but I have had a blast continuing to work towards robots deployed in homes (I am obligated to say that we are always hiring).
I would say that general purpose robots are the north star that I’ve chased for the past ~4 years. As a student, my mentors taught me to search for north stars: identify the dream demo at the end of the horizon, and then plan bi-directionally to slowly fill in the missing pieces. The path towards this north star of everyday robots in homes is winding and treacherous: before concerns about robots continually learning from experience, we first must tackle how to deploy robots to collect data. Even before that, we must consider which modalities we need to integrate, and the robot design to unlock these capabilities. The premise of personally growing to conquer the daunting full stack nature of robotics is one attractive reason that I continue to trudge up this trail.
Beyond career changes, my mindset has decidedly shifted. If you had asked me around the time of writing my last post, I would have said that I am someone who works on an idea not because of external motivators such as climbing ladders or even making social impact, but rather because of an intrinsic motivation to work on the problems I deem interesting. In my mind, there were certain problems that I had natural taste for. My personal north star was to identify and work on ideas that I resonated with; robots happened to resonate the deepest. Towards this, I was very deliberate with my time and my decisions; there had to be a solid reason to dedicate my entire identity and ego to an effort.
While the north star mindset readily applies to worldly or physical ideas, it does not cleanly translate to personal ones; it is not easy to bi-directionally plan your way towards essentially self-actualization. It wasn’t even clear what exactly my north star was; the best I could do was to venture a guess far into the future, and plan myopically towards it. My sense of motivation relied so heavily on working towards what I deemed was important, on resonating with some grand purpose, to gain a sense of fulfillment. Putting in 100% effort and feeling in hindsight that I was not working towards my north star, or not taking the fastest path to get there, was inevitable given the inherent uncertainty. This led to the intense burnout which I previously wrote about.
Rather than cast my north star far ahead into the future, I’ve come to almost an opposite conclusion that there is no true north; there is no grand purpose that I should be aiming straight towards. While this phrasing sounds pessimistic, it is also a very freeing realization. There is nothing inherent to myself which motivated what I have worked on. Rather than needing to find a reason to do something, the act of deliberately choosing to work on it is reason enough. “Intrinsic motivation” is not something intrinsic but is instead generated; it is a function of the environment you choose to surround yourself in and the goals you set.
In the spirit of identifying the missing pieces of a north star, it is both more sustainable and more achievable to set goals which best reflect the current moment, rather than setting platonic goals for grander, longer term, and more abstract problems. Though I have enjoyed my work over the past year, I can’t deny missing working on more fundamental robot learning research. I still want to make robot learning more human-like using ideas such as online learning, and there likely are places better fit to immediately work on such a grand and unsolved problem. This previously would have been a sticking point for me. But while I am at Tesla, there are certain ways that I want to grow: thinking about data collection beyond controlled environments, moving fast and breaking things, and learning about the lower parts of the stack towards becoming a full-stack roboticist. Decoupling ego from what I work on has also really been productive in being more present, making decisions faster, counteracting burnout, and making my relationship healthier with work. While this may sound obvious in hindsight, mentally subscribing to a new ideology and reaping the benefits takes time, especially for someone whose mindset largely hinges on putting in 100% effort to achieve what I 100% believe in.
I still believe in north stars, just not the true north star that I once believed I was predisposed towards. How do I find my north star then? My perspective is that life is a random walk, consisting of experiences in which I may change in unexpected ways, in places I could not have foreseen. No matter where I may find myself though, I believe very strongly in two things. Firstly, growth compounds. Secondly, nothing is strictly orthogonal from what came before. My PI uses the analogy that science is a spiral, where progress is made by revisiting old problems with new tools, and developing newer tools from previously unknown understandings. Similarly, people tend to return to the same problems, but presented in different forms. In the same vein that you can only connect the dots by looking backwards, I believe that north stars are something that you can only reason about by first looking southwards to identify the faint through lines in the breadth of your experiences, then extrapolating that same vector northwards. Affinities are not natural, but rather are carved by the trajectory of our experiences. With each experience, I gain a better approximation of the north star I am drawn to, which I can approach more directly and with clearer perspective.
My understanding of my north star has evolved much. Rather than being directed towards problems as I once thought it was, it is almost entirely personal: to become my own unique person by deeply understanding myself, to become self-sufficient to work on whatever I desire, and to ultimately find purpose through my work. There are truly no shortcuts to get there, and there aren’t even guarantees where the destination is; the path is winding, and there is much I will experience along the way. If anything though, it is a timely reminder for myself to enjoy the detours.
on the idea stack:
robin wall kimmerer’s braiding sweetgrass;
hermann hesse’s siddhartha;
JEPA;
quasi-direct drive actuators;
spiral dynamics.