I think the "does it make sense to furlough" question is a little misplaced or at least inadequately contextualized.

Let's say demand slowly drops 10% and you have a slight excess of pilots. You expect it to pick back up in a year. Profit margin is down from 4% to 2%. In that case it makes sense to pay $100 million now to save $200 million next year.

This is different. Demand halved overnight essentially. Imagine CFO says cash burn leads to bankruptcy in 90 days. Furloughing 25% of the pilots will stretch that to 180 days. For 90 days you don't retrain anyone and just use what you have. Storm passes and you still have cash. Doesn't really matter what it costs to retrain; if your choice is bankruptcy or furlough, you furlough.