Humanoid Robots Are the Future—So Why Are We Still Assembling Them With Hammers?

人形机器人是未来？可我们为啥还在用锤子组装？

eu.36kr.com

After going undercover as a temp worker in a humanoid robot 'factory' in Zhongguancun, I didn’t find a sci-fi assembly line. What I found was something closer to a college lab or a repair shop—dozens of people hand-assembling robot joints with screwdrivers and paper manuals. And yes, when parts jam, they just grab a hammer and knock them loose.

我以临时工身份潜入中关村一家‘人形机器人工厂’后，没看到科幻片里的流水线，反而像是大学实验室或家电维修铺——几十人拿着螺丝刀和纸质说明书手工组装机器人关节。零件卡住了？直接抄起锤子敲开。

This isn’t just ironic—it’s the core bottleneck. No matter how advanced the AI, if we can’t industrialize the hardware like Ford did with the Model T, we’ll never achieve true mass production. Right now, robots are being built by skilled artisans, not machines. The ‘embodied intelligence’ revolution is being powered by human hands, not algorithms.

这不只是讽刺——而是核心瓶颈。无论AI多先进，如果硬件无法像福特T型车那样实现工业化，就永远谈不上真正量产。现在造机器人靠的是熟练工匠，而非机器。‘具身智能’革命靠的不是算法，是人手。

FordT_Today (今天的福特T型车)

Exactly. We’re in the ‘horseless carriage’ phase of robotics. These aren’t cars—yet. They’re rich guys’ science projects with batteries and motors. You don’t mass-produce prototypes. The real question is: what use case will actually justify millions of robot bodies?

没错。机器人正处于‘无马马车’阶段。这些还称不上汽车——顶多是富豪们的科学玩具，加了电池和电机而已。原型机是不能量产的。真正的问题是：到底有什么场景真的需要百万台机器人？

MechanicWithASolderingIron (手持电烙铁的机械工)

Spent three days learning to assemble a single joint module. You think it’s just tightening screws? No. One wrong torque value, and the motor burns out. Then you start over. No AI can fix bad craftsmanship. This job needs hands, eyes, and brains—all three.

我花三天才学会组装一个关节模块。你以为只是拧螺丝？错。扭矩错了，电机就烧了。然后重来。再聪明的AI也救不了烂手艺。这活儿得要手、眼、脑三位一体。

CEOofDreams (梦想公司的CEO)

You’re missing the point. This isn’t about efficiency. It’s about iteration speed. When R&D and production are on the same floor, engineers can fix a design flaw in hours, not weeks. The hammer isn’t a sign of failure—it’s a tool of agility.

你们搞错了重点。这不关乎效率，而在迭代速度。当研发和生产在同一楼层，工程师几小时内就能修正设计缺陷，而非几周。锤子不是失败的象征，而是敏捷的工具。

MechanicWithASolderingIron (手持电烙铁的机械工)

Agile? Try explaining that to the worker who has to redo a 4-hour module because the R&D team ‘iterated’ overnight.

敏捷？那你跟那个因为研发团队‘迭代’了一夜，不得不重做四小时工件的工人解释去吧。

EconomistInSneakers (穿球鞋的经济学家)

The labor cost is unsustainable. 300 yuan/day for high-skill work in Beijing? That’s exploitation. Either wages rise and kill margins, or robots stay niche. There’s no ‘middle path’ here.

人力成本不可持续。在北京，高技能工作日薪300？纯属剥削。要么涨工资压垮利润，要么机器人永远小众。没有中间路线。

RobotGrandmaFan (机器人奶奶粉丝)

Let’s be real—most demos are just robots twirling handkerchiefs. Show me a robot that can clean my house, care for my parents, or cook dinner. Then I’ll believe in mass production.

说实话吧，多数演示不过是机器人转手绢。你先让我看到能打扫房子、照顾父母、做晚饭的机器人，我再信量产这回事。

HouchangcunLandlord (中关村房东)

Best ROI in Zhongguancun: renting offices to robot startups. They pay premium rent for ‘R&D proximity’ while I sip tea and count cash.

中关村最稳投资：把办公室租给人形机器人初创公司。他们花高价租‘靠近研发’，我喝茶数钱。

CEOofDreams (梦想公司的CEO)

You think Ford built the Model T factory in a day? Every revolution starts in a garage.

你以为福特一天就建成了T型车工厂？每个革命都始于车库。

Humanoid Robots Are the Future—So Why Are We Still Assembling Them With Hammers?

人形机器人是未来？可我们为啥还在用锤子组装？

这7亿美元的机器人订单，是真正的‘终结者时刻’，还是只是一场营销炒作？

机器人服务员刚给我点了单，但我们真的准备好迎接工厂革命了吗？