Why you should move that button 3px to the left

That point is central to the post and yet not really followed IMO. Costly changes to the UX should be justified in objective terms, e.g. consistency, color or style matching, differenciation, focus improvement, branding, readability... Not a "it feels better", or "it's delightful" or "I'm a creative person, I know what's right" (<- if it's his/her personnal project this one is OK)

Designers are not in a good position to appreciate the trade offs. It's very easy for a designer to think 'this infelicity decreases delight, and fixing it involves just moving this thing to there'. This is probably an excellent spot of a real design issue, but that isn't the only thing that determines whether the work should get done or not.

I have no solutions, but problems like this are generally because of an insufficient appreciation of the other professions challenges, and if you're working in an environment with skilled practitioners of both disciplines, more communication is likely to fix it.

Here are some suggestions that might help designers talk to developers: Never use the word 'just' when talking to developers, frequently apparently simple things have huge ramifications and just as a designers specialty is understanding how the little things create a full experience, a developers specialty is understanding how little changes affect other things in the full system. Talk to them about the effect you're aiming for, rather than just the change itself - if the change is trivial, the worry is that you're not respecting their time, but if you convey that the trivial change will have a large specific effect then they will understand and may be able to do other things proactively to help you achieve your effect. Show that you care about more abstract values such as 'maintainability'; the ability to add features quickly in the future and having a low bug count are both massively beneficial to the user experience. Where you can, show that your fixes are not merely matters of opinion but are based on empirical evidence, good developers respect evidence. Get involved early in the process and talk regularly to the people implementing the design - show that you care about their opinion, but educate them in seeing things from another viewpoint too.

As a developer, you are probably used to having a discussion with the person who reported a bug for technical issues. Do the same for design issues - even if it's just 'move this button three pixels', talk to them about it, find out why it's important, find out if they really mean that it needs to align with some other element. If you have a legitimate reason to not want to do it, explain the problem - you might be able to work out a better solution together.

I’m a professional software and web developer, and I currently build web-based user interfaces for a living.

but zooming in on many web pages can easily destroy the design

Sorry, but that’s simply not true if you have any idea at all what you’re doing. The tools to support designs that follow the original intent but are flexible enough not to break just because someone has different default fonts or zoom levels have been around for many years, they still work as well as they ever did, and they are entirely compatible with the more adaptive/responsive designs we often use today.

(e.g. using px's for your base font size instead of em's or %)

These days, it’s more likely to be the other way around IME.

Too many people still rely on received wisdom from the days when browser zooming didn’t adjust the whole page, as all major browsers now do. The original arguments for avoiding px-based font sizes were about allowing users to configure their preferred size in their browser preferences and have web sites respect it instead of overriding it. Today the default font specified by most sites is larger than it used to be (a good thing, up to a point) and if that doesn’t fit the user’s needs then every modern browser will scale it up when the page is zoomed.

Too many people are also making trendy design decisions under the banner of “mobile first” that result in a poor user experience on desktop/laptop systems (or even, ironically, on tablets). Consequently, we get silly things like specifying 30px thin fonts and main page widths/margins as percentages with no other limits, which probably look awesome on the designer’s chosen development device(s) but unfortunately look terrible and can’t be fixed using the usual browser adjustments on much larger and/or smaller screens.