OK, I ran the numbers and my gut feeling was very wrong. Beverage grade CO2 is 99.9% pure, so I looked at how long it would take for the oxygen permeating through the O-ring to equal 0.1% of a 1-liter headspace. At that point the oxygen entering through the O-ring would equal or exceed the oxygen already in the CO2. That seemed like a reasonable value to use as a threshold. The time for that to occur with a Buna-N O-ring was only about 9 days, so silicone would be 1-2 days. Both are way shorter than a month, so even Buna-N O-rings will not keep beer oxygen-free for months. Considering that both are so short I'm not sure it matters a whole lot that one is a lot shorter than the other. If Buna-N were 10 months and silicone 1 month then I think it would really matter, but not 10 days to 1 day.