Learn everything

Initially, that might sound impossible — of course, you can’t actually learn everything. But that’s not the point. Rather, it is to approach the world with boundless curiosity, seeing connections across domains and wondering which ones you’ll find next. In our ever-evolving world, the most valuable skills you can have are learning quickly and finding patterns.

Conventional wisdom has always celebrated specialization. Often in our youth, we’re guided to pick a field, go to school for it, become good at it and subsequently pursue a career. This isn’t bad advice. The traditional path is easy to measure and reward. Specialization is useful, and many engineers gain significant advantages in solving specific problems from it.

Domain specialists often gather immense experience in their field and get quite good at execution. But they are often primarily good at what they have always done. It’s the classic “when you have a hammer, everything looks like a nail” scenario. Let me be clear — specialists are incredibly valuable experts, but those with broader knowledge can do things that specialists cannot. For instance, while leading software development on my high school robotics team, I deeply integrated design and programming to optimize our mechanics using simulation.

Learning everything can help us pursue more global solutions, and that’s what makes great innovators. It doesn’t prevent you from gaining deep knowledge. It means knowing when you need that knowledge and having the ability to get it.

I found that the “CS kids” have a tendency to sit behind a screen with a code editor open and consider little else. Drawing from my own experience of working on many areas from starting the robotics team, I took an initially unpopular approach of having my subteam members try disciplines like fabrication, assembly, or design. The effect was clear — students realized non-obvious hardware limits that would stop their theoretical code from working in practice. While this is a small-scale version of “learn everything” from robotics engineering, the concept scales.

If you ever find yourself feeling like you know everything, you’re not looking hard enough. The best part is that, with the internet, the challenge isn’t finding the information anymore. It’s knowing to look.

The goal isn’t to accumulate broader objective knowledge — it’s more about training your mind to be of a certain type (thanks, Paul Graham). One that’s hunting for patterns and finding unusual applications and connections that others don’t see.

Having a collection of dots isn’t very useful without having a network to constantly connect them together. Information isn’t unique, but your insights and understanding are — and they accumulate over time. And, you gain practice making connections. Interesting insights come from finding anomalies — cases that break the expected pattern. These exceptions either reveal opportunities for improvement in the original domain or suggest novel principles that could enhance our understanding elsewhere.

This process can feel daunting when venturing into new territory. Unfamiliarity with a field is scary — especially when you’re used to the comfort of being an expert on your own. Exploration and expertise aren’t opposites but companions. We constantly explore outer space whilst only having explored a tiny fraction of our own oceans’ depths. True learning happens outwards into what we don’t know and deeper into what we do.

Immense knowledge of the different fields isn’t necessary to gain a benefit; a working knowledge is often enough. Yet a backward narrative is usually derived from the traditional track — people try to learn everything before trying to effectively apply it. Don’t wait. Momentum and necessity drive us to acquire deeper knowledge. We can leverage this by taking on projects that force us to learn new skills and fill knowledge gaps.

Little extra effort is required to become this sort of active learner. We all hear or see many things daily, but we tend to pay less attention to parts that aren’t directly relevant to us. It’s easy to take advantage of passive observation. Whether you’re overhearing a conversation in a restaurant or scrolling past technology on the news, there’s something interesting to be discovered everywhere.

Being an outsider is an incredible position. It allows you to ask “naive” questions that hit the fundamentals and ignite profound impact. As specialists gain more experience in a field, they tend to focus on narrower and deeper parts that might not have as much effect. After talking to many students and mentors, I’ve found that fields often become echo chambers. They chase local minimums together. But it only takes one person to ask, “Why not X instead?”

DeepMind revolutionized protein folding with deep learning techniques because it wasn’t limited by traditional biochemistry methods. Tesla is well-positioned to build intelligent machines because it approached cars as electric robots rather than as vehicles.

We don’t see the problems that exist outside of our specialization. Problems across domains often share the same principles.

I heard about an interesting example recently. It’s about how tumors move during respiration. This makes it hard for doctors to target radiation accurately. My aerospace engineering professor, the inspiration for this part, realized that control theory can predict the motion of these growths. This is generally used to model and control highly dynamic aerospace systems.

The true power of learning is in developing cognitive flexibility. It means being open to new ways of thinking about things. Looking back at the mindset for learning everything — comfort with discomfort, pattern recognition, asking questions, and making connections to new ideas — you might notice something. Adapting to change isn’t just about picking up new skills but also about evolving your understanding. By learning to evolve your knowledge you also learn to evolve with the world.

---

Thanks to Divij Motwani for reviewing drafts and suggesting ideas and Professor Puneet Singla for inspiring parts of this.