I grew up in Kolkata, wanting to be a Physicist. For all its flaws, I have always wanted to solve really hard problems.^[1] I am currently a student at Rice University where I study Mathematics and Electrical Engineering. Some things about me:

• I’m currently building Optica Industries
• I love writing, creating at the intersection of intelligence & embodiment, and startups.
• I spent about a decade learning visual arts (Japanese Ceramics & Oil Painting) and about 2 years practicing astrophotography. From art I derive taste, and from taste, art.

I am working towards realising what I want the outcome of my life to be, there’s a lot I do not know, but here are some things I believe that guide my search:

• The physical world is under-indexed. The most important problems left are ones you can touch.
  • Intelligence without embodiment is incomplete.
  • Manufacturing is a leverage point.
  • Software optimised everything it could reach. What it couldn’t reach is where the real work is. The next decade belongs to people who can do both.
  • The gap between what we can simulate and what we can build in the real world is the most honest measure of how much we don’t understand.
• On How To Think
  • First principles thinking is only useful if you actually know enough principles.
  • Depth is a competitive advantage that compounds.
    • The interesting stuff is always one layer deeper.
    • Depth without urgency is just procrastination.
  • Contrarianism is not the same as being right. But consensus is almost never where the interesting questions live.
  • Uncertainty is information. Certainty too early is a sign you stopped asking questions.
  • The scientific method is underrated in industry: Form a hypothesis. Test it cheaply. Update hard.
• On Building
  • Taste is a technical skill. You cannot build something that matters without it.
    • Ambition without taste produces noise. Taste without ambition produces nothing.
    • Knowing what’s worth building is as hard as building it.
  • Startups are the highest leverage way to run an experiment on reality. Everything else is commentary.
  • Being early is indistinguishable from being wrong until it isn’t.
• On People
  • Curiosity is a skill, not a trait. It atrophies if you don’t use it.
  • Talent is more evenly distributed than opportunity.
  • The best collaborators make you think faster, not just work faster.
  • Compounding is the most powerful force in knowledge. The person who goes one level deeper every year for a decade is in a different universe from everyone else.

Most of this I learned by being wrong first. I’m probably still wrong about some of it. That’s the point. I am open to all problems, but here is a small subset that really interest me.

• A complete theory of intelligence. How does understanding emerge from experience?
  • Why certain arrangements of matter produce understanding at all. What is the physical basis of intelligence?
• The simulation-to-reality gap. How do you build a robot that learns the real world without ever having to live in it?
• Large-scale self-supervised learning. Can we recover the latent structure of the world directly from observations?