People compare themselves not to their earlier selves but to the best visible result.
At work this looks reasonable: there are people who do it better, so there is room to grow. The mistake is not the comparison. It is the sample.
What is visible is not a random sample, it is a shop window. Releases, but not the months when nothing shipped. The report, but not the draft. The talk, but not the three rejected proposals before it. You are comparing your whole process with someone else's result. That comparison is lost every time, at any level.
A leaderboard is also built so that you never come first. Someone is always above, there is one place at the top, and even when taken it holds for a week.
There is one reference point that works — your own state six months ago. It has two properties.
It is honest: you saw that process whole, without the shop window, and you know what it cost. And it is dull: growth on it looks less like an achievement and more like "it is slightly less frightening to open someone else's repository" or "I stopped being afraid to delete my own code".
There is nobody to share that with, and it gives no reason for pride. It does, however, exist. The leaderboard does not.