"We" refers to humans, human-generated tools, AIs, etc.
Control the weather refers to being able to cause every climate condition which is within the normal range for some region, under some reasonable timeframe.
For example, a necessary, but not sufficient, condition would be that, if it rains on one day per week in Paris, then we can decide to make it rain on Monday, and not on any day for the rest of the week, whether or not that would have been the case without our intervention.
See also:
/Bayesian/will-we-be-able-to-control-the-weat-f47455a10483
/Bayesian/will-we-be-able-to-control-the-weat
/Bayesian/will-we-be-able-to-control-the-weat-ae971df8b4cd (This one)
And for a much stronger requirement in terms of weather control:
/Bayesian/will-we-have-total-control-over-the
@Fion Hmm, I don't think it discredits the control if there's an occasional 1 in 100 fluke, but 5% seems too high. ie that's not strong enough control for the purpose of this question. good question