"It seems like it was originally a good idea: if your startup is dead, let it die."
I never saw it that way, or at least at that scale. I see it as a way to move fast by validating assumptions with evidence and being unemotional about sunk costs. Rather than spending many hours of meetings and design sessions and approvals to try a feature or change, just do it, measure it, and if it's not working, kill it. If it is working, keep it, and nurture it. If you just spent 3 months on an idea that is getting no traction, don't spend another 3 months trying to fix it when you've got other promising things in the pipeline.
I never saw it that way, or at least at that scale. I see it as a way to move fast by validating assumptions with evidence and being unemotional about sunk costs. Rather than spending many hours of meetings and design sessions and approvals to try a feature or change, just do it, measure it, and if it's not working, kill it. If it is working, keep it, and nurture it. If you just spent 3 months on an idea that is getting no traction, don't spend another 3 months trying to fix it when you've got other promising things in the pipeline.