Principles
I think clearly when I separate three things: what matters, what's true, and what I'm willing to do about it. I think most of my personal and professional friction comes from mixing up these questions and overestimating how clear the answers are.
First: Values set direction. Evidence sets method.
"Should we?" is an evaluative question. It can't be solved by more data.
"Can we?" is an epistemic question. It can be solved—slowly, imperfectly—by experiments, measurement, and iteration.
So I try to route decisions accordingly:
- Route "should" upward (to values, identity, responsibility).
- Route "can" downward (to practice, prototypes, tests, and feedback).
From there: Capabilities are discovered, not declared.
I don't trust self-images—mine or anyone else's.
"I can do X" is a scientific claim: it becomes real through repeated attempts under stated conditions, with honest accounting of failures.
Constraints matter (failure teaches), but I try to orient around capability creation: expanding what I can reliably do, not just shrinking what I avoid.
Then: Codify what works.
When something is repeatedly true, I compress it:
- into a principle when it requires judgment,
- into a system when it can be made deterministic.
This is how knowledge becomes leverage: fewer decisions require fresh willpower every time.
But always: Protect what matters from what's measurable.
Metrics are useful, but they're not neutral.
What's visible tends to become "real," and what's "real" tends to start running the show. I try to keep slow, non-measurable commitments (care, integrity, meaning) from getting colonized by fast proxies.
Therefore: Earn autonomy. Don't gamble for it.
More freedom should come from more validated capability and tighter feedback loops—not from pretending risk is virtue.
A Sikh triad for knowledge creation
A mapping from Sikh tradition:
Naam Japna — humility (epistemic honesty)
Humility is the beginning of science.
It's the stance that says: I might be wrong; reality gets a vote; my model is not the world.
Practically: keep a log, look for disconfirming evidence, update without drama.
Kirat Karna — hard work (experimentation)
Knowledge is built, not wished into existence.
Hard work is running the trials: prototypes, reps, iterations, measurements.
Practically: choose a method, do the work, repeat until the signal is real.
Vandh Chakna — sharing (replication + diffusion)
A result that can't be shared isn't stable knowledge—it's personal luck.
Sharing is how claims become communal: others can test them, extend them, and build on them.
Practically: publish the artifact, teach the method, share the credit.
humility → experiment → share → collective capability.
Science, at its best, is discovery in service of expanding what we can reliably do together.
Continue
These principles inform the work I do with organizations and the writing I publish.