3 thousandths faster --- how does that affect the leap year thing?
If we lose 3 microseconds a day (and from the article it sounds like this may just be a "one time" event) in 28,800,000 days (78,904+ years) we will have lost a day. I'm not going to change my schedule over it.
As days get longer, we will need fewer leap years (intercalary days) to keep the calendar aligned with the seasons. At 0.002 sec per day per century, it would take about 5,000 years to loose one complete day, which about as long accuracy of the Gregorian Calendar, so it should not effect leap years in our lifetime.