If we can measure variance inside of 33 days, and that's extrapolated over a few millions of years, that could skew the numbers a bit. [..] that has a hug

If we can measure variance inside of 33 days, and that's extrapolated over a few millions of years, that could skew the numbers a bit. [..] that has a huge impact on carbon dating. No, the effect is very extremely small and this has no impact at all. Lets suppose that: * we measure differences of 1/1000th in decay rates (this is reasonable given the differences the article is based on) * the average variation is 1/1000th, otherwise it would have been detected earlier. * the halflife of C14 is 5000 years (to make calculations easy) * we are determining the age of something 20000 years old (carbon dating is only used for ages in that order of magnitude). * the decay formula is y = x(1/2)^(t/tau), where y is the amount of C14 you will measure after t years, tau is the halflife in years and x is the initial amount of C14 we obtained via other means. We measure an amount of C14 that's 2^4 times smaller than current amounts and conclude that the amount of C14 was halved 4 times. If the decay rate was actually 1/1000th larger on average, then we should have conclude that 2^(-4) = 1/2^(t/5005) or t = 20020 years. In other words: the error is as large as the relative seasonal variation, which is quite small already and probably falls well within the regular error bars.