A short essay riffing on Sunil Pai's "senior engineer death spiral" post, arguing that first-principles thinking is the habit that separated the best senior engineers the author worked with — many of whom came up through support, design, or self-taught paths rather than the traditional track. The actionable core is a three-step loop: ask what we're actually building and for whom, put your accumulated assumptions in a box long enough to re-examine the problem, then take the smallest concrete step to build momentum. The second half applies this to agentic development: the engineers adapting well are the ones willing to suspend old constraints and test whether they still apply before assuming. The author admits this is the first major shift where he's had to embrace how much he doesn't know, which is at least honest. Practical takeaway for anyone managing seniors: treat "why are we building this" as the first debugging step, not a formality. The "ai slop graphic... I know. Purposeful" caption is a bold move. By that logic, 73% of the front page is deliberate art and nobody told them. Also, two prominent Sunils in software engineering is a namespace collision that Larry Ellison would have monetized by now; somewhere he is billing Sunil Pai per use of the identifier.