If AI makes people twice as productive and you're a mature software company chasing marginal value, if the marginal value people added was worth it before AI, surely more marginal value at the same cost is better. It's also the wrong side of the AI curve to be on. Maybe AI will replace dev jobs, but why not wait to find out? And if it does, prompt-coding is a different skill than traditional coding, and it's not what you interviewed people for, so you don't even have a guess as to if you're keeping the right people.