So either their business is a dead end, the inefficiency is at the management layer, or AI isn’t actually making workers more efficient.
Its why I'm not the biggest believer in "Jevon's Paradox" when it comes to software. Most software projects scale and as such they aren't really cost constrained. Or another way of stating this is "if the idea is good and can make lots of money the cost of dev's isn't a limiting factor" - hence why AI doesn't necessarily increase demand all that much. This is unlike say physical industries where cost can absolutely matter especially if the good is constrained already by affordability.
I think most SWE's know this - they see it with outsourcing as well where cheaper costs don't necessarily mean more software is done; but there is some "hope" that things will be OK.
I'd bet on all three being generally true.
Instead we're still talking about cost cutting. Seems the market is not focused on investing outside the AI moonshot gamble.