Would whoever is downvoting care to defend the article's claims? Maybe this one: > the timescale for light to cross an atomic nucleus is a factor of 100,000 sh

Would whoever is downvoting care to defend the article's claims? Maybe this one: > the timescale for light to cross an atomic nucleus is a factor of 100,000 shorter than the time to cross an atom, and so if there are any physical effects that preferentially affect these longer timescales as compared to shorter ones, the difference between atomic and nuclear clocks could reveal them. Because that claim is nonsense. It is very, very, very difficult to resolve discrepancies between two measurements, one of which is 10^4 times more accurate (the article's number, repeated multiple times). It just disappears into the uncertainty of the less accurate measurement. So you need to improve the less accurate measurement, but you can't do that because the way you gained your accuracy was also the way you gained your signal. Yes, this can be worked with, but it's not something nuclear clocks will ever do alone. My main thesis, that nuclear clocks do not serve any market between the current cesium/rubidium standards and the national metrology labs' best, still stands.