But taking a week to do it costs money. People won't work productively if they might get laid off any second. By laying them off atomically without any advance warning, they can get back to work quicker.

Back-of-the-envelope calculation: It takes a week to lay them off. During this week, productivity will be down 50%. You have 654 employees (ex-post). Say they each cost $200k/year (taxes, office space, etc), that gives (1/52) * 0.5 * 200k = $1.92k/employee in lost productivity, or almost $1.3 million in total.

You think the remaining workers are going to “get back to work faster” just because their colleagues were canned fast?

That’s not how people work.

Even if I remained, I would start looking for a new job very fast.

Amen, brother. A demoralized workforce is not a more productive workforce.

If they'd be unsure for a week as to whether they would be fired or not, yes. If they are never uncertain, there is less time in flight.

That's a good calculation. Now compare that with the PR fall-out from their chosen approach.

What are the future losses from not being able to hire top candidates (and maybe even loss of business if customers start boycotting it), since obviously Bird is a shitshow of an employer?

Customers never care about that. Case in point, my computer presumably contains all sorts of rare minerals mined by slave labour in the third world. Is it regrettable? Yes. Does it mean I stop using computers? No, it's a price I'm willing to pay.

Customers don't stop using Facebook over the clear and present erosion of their freedom it presents, so thinking they would be concerned with how a company treats its employees is absurd.

As for prospective employees, it's going to be a buyer's market from here on out. Universities have been pumping out CS grads en masse, and the startups have been hiring them. With the startups going bankrupt, there will be a glut of unemployed programmers.