Definitely agree about the loss of precision. It's also a problem for machine readability and SEO. The solution I like is using the HMTL5 <time> element for semantic markup of timestamps. If you do something like this:

    <time datetime="2019-10-11T13:53:36.861" title="Fri Oct 11 2019 13:53:36 GMT-0700">2 minutes ago</time>

Then you get what in my opinion is the best of both worlds: machine readable timestamp in the datetime field, human-friendly detailed timestamp as a tooltip (the title field shows when you hover over the element), and relative time as the default display. Bonus points if your React component or ERB template or whatever switches from relative time format to absolute for the inner text when the timestamp is older than a day or so.

"when you hover over the element"

On the website I'm looking at Google Analytics for right now, fewer than 35% of the visitors are from devices where "hover" means anything...

Well, it would require JS, but you could turn it into a button to show/hide an HTML tooltip or additional displays of the time. That iOS supports a mouse (in accessibility settings) or an Xbox controller but doesn’t properly support an “abbr” tag... it was acceptable in 2007 when we were happy to see the phone load nytimes.com with all its columns. It’s feeling very out of date in 2019, when 12 years later we’re adding downloads and “desktop safari” but we never finished the “desktop-equivalency” parts of the HTML5 spec. It would be amazing if Apple ported some of their iOS native components to web components or to enhance HTML5 elements. The farthest we’ve gotten so far are keyboard search button hints, if I recall correctly. Which is really quite disappointing, that an app made for jQuery UI would work as well as an app from today if you flatten and modernize the design, and add a bit of offline caching...

I prefer the HN method of 28 min, yesterday, 2 hours ago, etc... for HN. The older a comment or post, the less relevant it is to a particular discussion. This site is for real-time discussions of articles, so it is optimized for that use-case. The older a comment, the less relevant it is towards the immediate discussion. Thus, older comments have less precision, because those details don't matter for the discussion.

I'd expect a site/application that is designed for students to manage homework would have different date/timestamp requirements.

HN may be optimized for current discusions, but is also a repository of knowledge; I frequently land on years old threads from Google searches. For older times, an explicit ISO timestamp has at least two benefits: 1) I can compare it easily to other dates (e.g. is that comment older or more recent that event X, or software version Y), and 2) time of the day is a half-decent proxy for telling whether the US or Europe was awake then.

And while you can justify the inprecise stamp on HN, there's no excuse for GMail and a score of other messaging software doing the same. When I'm looking at the date of an old e-mail, I almost always need the exact date and time. "2 weeks ago" is useless. "2019-09-27 08:43" tells me it happened before work, on a day after an important meeting, etc.

ISO8601 date/time can easily be transformed into the fuzzy format HN uses, but not vice versa since the fuzzy format is lossy.

If HN sent dates to the client in ISO8601, the client could then transform that format into the format of their choosing with client-side javascript (either a userscript, or delivered to the user by HN.)

It's hard to be sure this isn't satire.

You mentioned 'for HN' twice in the same sentence.

Your 2nd, 4th, and 5th sentence all try to assert the same thing - older comments are less relevant than newer comments - despite there being no evidence, anywhere, that this is the case. Your 3rd sentence is partly true, though there's some historical record aspect to HN comment threads that you're discounting.

Your second paragraph is a non sequitur. Sites that correlate points in time with timestamps are ubiquitous, and there's rarely a compelling case for using wildly different approaches to displaying same.

They could easily do it the Reddit way and still provide a timestamp but keep the simplicity and relative time format. Just have it pop up the timestamp on mouse hover. That's still very simple and can be effectively done using just CSS.

You don't even need any CSS for that — the title attribute alone would be sufficient.

Personally I like HN the way it is. Very little UI clutter. If we start doing ISO dates and maybe options to switch between high precision and default low precision, we lose the good feel that HN has produced.

I'd prefer both.

I have one word to tell you: timezone

What about them? ISO8601 supports time zones.