(Disclaimer: numbers are entirely made up).
Ask 90% of developers on this planet when "feature X" will be ready, they'll be honest and tell you don't know. They might be able to have a rough idea when (and we've been given this rough idea; originally it was "a few days", now it's "a week or two"). And a lot of the time this not knowing is not just because of difficulty in estimating the time implementing a feature takes in itself, but also because there'll often be other things to tackle at the same time and priorities can shift rapidly.
Another 5% might have a good idea because they're almost done, they have no big remaining blockers, and their current and near-future workload appears to be predictable.
Ask the remaining 5% of developers, they'll give you an answer but they're lying to you, they don't actually know, and most of the time they'll be way off the mark, they were just trying to please you.
Estimating software development is incredibly difficult.