Sukumar Rajagopal shared a LinkedIn post about our meeting on 11 June at Chennai, saying:

Had a fanboi moment recently with Anand S who I have been following for a long time. He is prolific. It’s very difficult, no impossible, to keep up with him.

He generously shares his work with full source code; approach; working examples; the whole nine yards. All this I knew before I met him in person. The in-person meeting took me by surprise.

I was hoping to ask him a number of questions about his methods etc. But instead he peppered me with a lot of questions. Curiosity and humility is another hallmark of thought leaders. Anand showed me the way on how to be curious even in a topic like Gen AI where he possesses encyclopedic knowledge.

(etc.)

After cringing for a bit, I jotted down a few things.

Why I still learn: I get a kick out of it. Some get a kick out of wealth. Or power. Or fame (me too). Or learning (me too). It’s just a dopamine hit.

Why I learns topic I already know: I’d rather learn something I’m interested in. Who cares how much I know? It’s how much I like a topic. (Like re-reading books.)

How I learn from people: My current process / philosophy is:

  1. Everyone has something worth discovering, but not every conversation is worth my time right now.
  2. So, meet new people and ask good questions.
  3. Continue if there’s emotional / intellectual stimulation (surprising, interesting, moving, connecting, energizing, challenging), else exit warmly with respect.

Why I share generously: I’m trying to be liked and become famous. (Really.)

Why I am humble: It helps me learn and makes me more likable.

Anyway, I opened my notes from that meeting, where I had jotted down what I learned from Sukumar.

Let me share what I learned from that conversation. You’ll see who’s really generous with their knowledge.

Habits:

  • Start small. “I started with two pushups. Now I do sixty at one go.”
  • Anchor to something you already do. “After I brush teeth, I have to do it.”
  • Hard work is hard to habitualize. Start easy. “If there’s a lot of cognitive or physical effort, it can’t form as a habit.”
  • Streaks have a hidden cost. “Streaks create pressure. Once the streak stops, it’s hard.”
  • Prefer cycles. “I do it for 6 days. 7th day, rest. You hit a sixer.”

Leverage:

  • “I don’t have time” is not a useful perspective. “You have no leverage. You can’t create time.”
  • Find alternative perspectives. “Look at the same thing from as many viewpoints as you need to.”
  • Find the Trim Tab: the one step that transforms the system. Uber changed only “the ride-hailing step… everything else remains the same.”
  • Set extreme goals, not incremental ones. “I want to move the rudder with zero energy and zero time.”

Reflection:

  • Worse than failing is not noticing. “You’re not achieving higher order outcomes - and you don’t even know.”
  • Small successes blind us to bigger opportunities. “Zumba, yoga… exciting for three months. Then it dies.”

Changing people:

  • Look for hunger, not pedigree. “People from disadvantaged backgrounds have the hunger to succeed.”
  • People ignite at different speeds. “Camphor lights instantly… banana stems don’t.” (Doesn’t translate well from Tamil.)
  • Don’t tell someone they lack agency. “Reactive buttons get pushed.”

Mentors:

  • A great mentor compresses years into months. “What takes 10-15 years can be done very rapidly.”
  • Quiet people make good mentors. “It’s the art of asking the right question.”

Trust and selling:

  • Anyone can sell, in their own style. “It’s just that their style may be different.”
  • The real skill is earning trust. “There are many methods. Flashy presentations are just one way.”

Working with AI:

  • Use models as sparring partners, not judges. “Generate it in ChatGPT… go to Claude and have it reviewed… go to Gemini.”
  • A model always finds something to say. “These LLMs can never say, ‘I don’t have any review comments.’”
  • Run a pre-mortem. “You travel in a time machine, you fail, you look back and see what would have tripped you.”
  • Use the simplest tech that works. “No database, no web sockets, nothing… Just polling.”

And he bought me lunch!