It’s January. Everyone’s thinking about the “1% better every day = 37x improvement in a year” math.
It’s inspiring. It’s also not how improvement actually works.
The math assumes every gain stacks on the previous one. But most improvements don’t compound. They reset. You get better, then you lose the thread. You make progress, then you can’t tell what’s working. You ship something, learn nothing, and start over.
James Clear said it well this week: “New goals don’t deliver new results. New lifestyles do.”
But here’s what gets missed in that framing: a lifestyle requires infrastructure. You can’t sustain what you can’t see.
I learned this the hard way building Revarta. I was iterating fast on the AI interview experience, shipping improvements, feeling productive. Then I hit a wall: outputs weren’t landing and I had no idea why. I hadn’t set up LLM tracing. Every conversation was a black box. I was pushing hard without the foundation to learn from what I was pushing.
I had to stop, instrument properly, and essentially restart my learning curve.
Compare that to a decision I got right: investing in observability before I had much traffic. It felt like a detour from building features. But within weeks I was seeing rage clicks, bugs in unhappy paths, friction I never would have found through user feedback alone. Every fix made the next problem visible. That’s compounding.
Paul Graham put it differently: “Organizations that can’t measure performance end up measuring performativeness instead.” The same applies to personal improvement. Without visibility into what’s actually happening, you’re not iterating. You’re just moving.
The question isn’t “how do I get 1% better?” It’s “do I have the foundation to capture the gain?”
Build the foundation, not the feature.
This year, before you set the goal, ask: what would make the improvement visible?