For a variety of reasons the Earth isn't always perfect about making it around it's access in the expected amount of time. The implemented answer to this is Leap seconds. Periodically we add a second to our time keeping systems. Since 1970 we've added 26, and we're slated to add the next at the end of the year. I think that this system is stupid for a variety of reasons. (also let's just state off the bat that all numbers in this post will be approximate)
The first thing is the question of why adjust. Why do we need astronomical noon to be perfect noon in part of London? Not all of London mind you, just part of it. At a latitude of 51 degrees 30 minutes, a second's worth of the earth's rotation is only .179 miles. London is over 600 square miles, which is equivalent to circle with diameter 27.6 miles. which at that latitude would take 154 seconds to rotate through. At the current rate of leap seconds it would take over 200 years for the perfect moment of noon to make it all the way from one side of London to another! And it's not even perfect noon since the sun is always slightly to the south. And with s system that takes so long to pay off London might be growing faster than the need for leap seconds to keep noon in London. Will it matter when we've got colonies all over the solar system where in the sky the sun is at noon in London?
Another big problem with this is the effort it takes to actually implement this for all our time keeping systems. Human's don't notice that big a change but computers sure do. So in all our computer systems we keep having to code in and fix weird edge cases like adding a second. Adding this leap second is hard and it's buggy. A stupid amount of human effort goes into pulling this trick off. And probably more effort goes into fixing the things that break when it fails. And this time Google's solution is real time to not implement the extra second but to slowly stretch the seconds on other side. Hopefully nothing high precision will be using googles clocks that day. Plus there is the question about what they do going forward. Normally things keep track of all the time unit seconds since the beginning of 1970, and have a table to look up when there have been extra seconds, but if google computers don't ever live that extra second then how are they eventually going to take care of it. Are the all the computers just eventually going to register an error of a second and skip ahead? The effort that is involved just doesn't seem worth it.
This also has the gross effect of eliminating all other units of time. If the span between 11:59:30 and 12:00:30 is a minute, then how long is a minute? It depends on which one. If it's one of the ones that has had a second added then it's 61 seconds. And so on for larger units. This already impacted years with leap years, and moths weren't regular anyway, but now everything above a second is not just a calendar indicator instead of an actual unit. And because of weird implementations like google's, a second isn't even really that any more.
One potential solution for this is to do it in time zones. Right now to tell you your local time. your computer has to look up the time zone and modify the time based on that and then look up the leap second chart and modify the time based on that. Why modify the time twice? Why not just make leap seconds or any other minor adjustments for local pride part of the time zone? That way locale's that want noon to be special can finally have it that way, and then they can keep it that way for as long as they desire.
But what we're really going to eventually drive at is global / universal time. The inefficiency of having so much space/stuff dedicated to one person is eventually going to falter and we'll end of with round the clock people trading off space/stuff throughout the day. Not to mention globalization which increasingly creates communities in many time zones that all are on one schedule. All the effort that has gone into local time, and keeps going into maintaining local time is eventually going to be set aside. And then we'll look at the legacy of the leap second look up table and ditch it. Hopefully by then it won't be so late that making that change in and of itself won't be a major headache of an undertaking.