Astronomers do not need leap seconds, because even with this adjustment UTC cannot be used to determine anything in astronomy.
Astronomers need either true time, which is TAI, to be used in computing the positions of celestial bodies, and they need for observations the so-called Sidereal Time, which is not a time but the angle between a coordinate system attached to the Earth and an inertial system of coordinates attached to distant celestial objects that have negligible angular movement (in the past those were distant stars, now they are distant galaxies or quasars).
The Sidereal Time can be computed in a complex way from TAI, because it is determined by the periodic rotation and precession of the Earth and by various superposed periodic or random movements.
The UTC is not adjusted to match the current true rotation angle of the Earth, which you can measure by looking up to the stars, but it is adjusted to match within 1 second a fictitious angle that would be the rotation angle of the Earth-Sun direction corresponding to an Earth that would rotate uniformly both around itself and around the Sun, so that the duration of a day would have been constant.
In reality, the duration of a Solar day, i.e. the time between 2 consecutive noons, varies a lot during the year, by a large fraction of an hour (by about a half of hour peak-to-peak), so using UTC directly for estimating the position of the Sun gives a very big error, of many minutes of hour.
So what you need for astronomy is to know the current TAI and you need a Sidereal Time calculator, which you need for knowing in what direction to point your telescope, to find a given celestial object.
UTC cannot be used directly in astronomy, but only after passing either explicitly or implicitly through TAI. The fact that astronomical almanacs are published using UTC in their tables is obfuscating this, because the values in the tables have not been computed using UTC, but everything has been converted to UTC to match the time that is presumably shown by the watch or clock that the almanac user may have.
[0]: https://en.wikipedia.org/wiki/Leap_second
That said, no one wants to admit it, so contemporary science follows Falsification, where we find ways to not actually make claims about reality. (Which as an Instrumentalist/pragmatist, I love Karl Popper, its just not metaphysical truth. And that would break Popper's heart)
I think it matters. No the planets are not doing circles around the sun. Circles don't actually exist, they are doing elipses.
Also 'real' has quite a few meanings. If I ask the question 'Are you closer to a keyboard or the gym?' does that question exist?
This kind of stuff does end up mattering. It becomes much more noticeable in psychology (and biology). If you read Freud, Adler, or Jung, you will say 'Oh extrovert! I've seen that before!' But then you realize its vague and almost always true. Its like a horoscope.
So if we think there is a truth to reality, we look for perfect relations. If we think its impossible for humans to figure out, we look for best fits.
I'd argue the opposite is true for anyone who has studied statistics which is largely built on Instrumentalism (think George Box: 'All models are wrong, but some are useful') and Popperian falsification (Null Hypothesis testing). We are absolutely taught to treat models as predictive tools rather than metaphysical truths.
And taking fluid dynamics, we used renyolds number, which is a made up ratio that helps for decision making... Its not like when we answered questions, we could answer the grey area we are discussing.
If I had to guess, I think its due to western civilization being built of Platonism (and even Aristotle was infected). Our science and morality is later built by platonic realism. Only in the last 100-ish years are we starting to get over it.
I can see how someone could misunderstand or forget what they're taught though.
I have read numerous explanations, but haven't found a really authoritative discussion.
[0] https://en.wikipedia.org/wiki/Leap_second#Rationale
> In 2021, it was reported that Earth was spinning faster in 2020 and experienced the 28 shortest days since 1960, each of which lasted less than 86399.999 seconds.[24] This caused engineers worldwide to discuss a negative leap second and other possible timekeeping measures, some of which could eliminate leap seconds.[25] The shortest day ever recorded was 29 June 2022, at 1.59 milliseconds less than 24 hours.[26] In a 2024 paper published in Nature, Duncan Agnew of the Scripps Institution of Oceanography projects that the water from increasing ice cap melting will migrate to the equator and thus cause the rate of rotation to slow down again.[26]