First Principles Thinking

A recent post from Sunil Pai is making the rounds because it hits a nerve: senior engineers can get very good at what they do, then end up trapped by that same experience when the work changes. The post argues for momentum over outcomes. When things feel stuck, the move is to shrink the problem until there’s something concrete you can finish, then use that small win to find the next step.

That idea leads straight into first principles thinking. The strongest senior engineers, Pai writes, don’t just lean on pattern memory. They keep asking why something is being built, what it’s for, and how the code connects to the people using it. Some of the best examples he’s seen came from support, services, design, or even entrepreneurship, but the common thread was the same: they could strip away assumptions and get back to the simplest useful move.

He says that habit matters even more as development shifts toward agents. The people adapting best to that shift are the ones willing to set aside what they already think they know, at least long enough to test what’s actually possible. In other words, don’t let an old technical limit answer the question before the question is clear.

That’s the hard part. Experience is useful, but it can also become a shortcut to the wrong answer. Pai’s point is that a familiar limitation or a past project shouldn’t get to decide the shape of the current problem. Step back, ask what’s being built and why, then try the smallest thing that could move it forward.

And for him, that’s where the agentic era gets interesting. If working with agents speeds up the loop between trying, learning, and trying again, then first principles thinking isn’t just a philosophy. It becomes the way to keep up without pretending you already know the ending.