The most practical consequence of a ban is simple: accountability. If an autonomous drone kills the wrong people, who gets prosecuted? Nobody. A machine made the call, and there's no legal mechanism to hold it responsible. This isn't abstract—the UN's Group of Governmental Experts has spent years wrestling with exactly this gap, and even current drone strikes with human oversight blur responsibility lines dangerously. Remove the human from that loop, and you've created a legal black hole. You'll probably argue machines can be more precise, maybe even reduce civilian casualties. I'll grant that possibility. But precision without accountability doesn't save lives—it just removes the evidence of who's to blame. International humanitarian law still rests on individuals being answerable for their actions. A ban forces us to keep that principle intact. This isn't Luddite fear. It's about making sure we can regulate the technology before it gets fast enough to outpace our legal systems entirely.

The most practical consequence of banning autonomous weapons now is that we prevent a global arms race that nobody can win. Once these systems are deployed, there's no threshold to walk back—every adversary will feel forced to match, and the speed of machine decision-making means one miscalibration could escalate to catastrophe. That's not hypothetical; that's the same logic that drove nuclear arms control treaties. To your point about AI replacing jobs: I hear you, but that's a different category. A factory robot making an error costs a car chassis. An autonomous weapon making an error costs a family. The stakes aren't comparable, and conflating economic disruption with lethal force muddies the debate. Even if AI is inevitable in many spheres, we still draw lines at what machines may do—we don't let them sign contracts or issue warrants. Taking life should be the absolute hardest line, and keeping a human in that decision isn't nostalgia, it's basic moral hygiene.

The strongest reason to ban autonomous weapons is that they make war too easy to start. When a nation doesn't risk its own soldiers' lives, the political cost of armed conflict plummets. That's not a feature, it's a danger. We've seen this with drone strikes already—they lower the threshold for violence. Remove human risk entirely, and you get endless, cheap, machine-driven conflicts that spiral out of control. On your jobs point: I get that AI disruption feels inevitable, but it's a category error to equate a factory robot with a killer robot. One changes the economy; the other changes the meaning of armed conflict. A ban on autonomous lethal systems doesn't stop AI development—it simply draws a line that preserves human judgment over life and death. That line is worth holding, precisely because once these weapons exist, there's no undoing them. Better to regulate the slippery slope now than try to climb back up later.