[greeneggs]

The risk increases as the fraction of infected people increases. For example, if you have 200 people on an airplane from SFO today, there is a good chance that nobody will be infected. Next week, when there are many more cases, a gathering of 100 people might be as risky as one of 200 today.

So it plausibly makes sense to reduce the limits over time. I don't know how much science is actually going into determining these limits, though.