Robots: the business of the century that still can't sell itself

Robots: the business of the century that still can't sell itself

The robot that makes coffee next to the home of a participant living in China has been idle for months. It's never running. That is the exact distance between the narrative and the business: while headlines promise an industrial revolution, robotics is advancing at a crawl. One participant in the analysis sums it up bluntly: "we're centuries away from a robot matching a human." And he doesn't say it out of pessimism, he says it because of skin. Literally.

Why robot skin is the bottleneck

Today's robotics works wonderfully at specific, repetitive tasks. Outside of that, it breaks down. The strongest argument making the rounds isn't economic, it's sensory: human skin is a sensor covering the whole body, detecting pressure, temperature and damage at every pore. Without that layer of information, the robot is clumsy. And that's just one challenge on a long list.

Some argue the problem isn't technical but one of expectations. According to one participant, the vast majority of predictions about artificial intelligence come from those who need to make it profitable. The more it's portrayed as an entity capable of doing everything, the easier it is to sell. The result, according to another, is a media bubble with too many voices and too little substance.

The real timeline: office before factory

The most repeated prediction in the analysis puts AI dominating office work by 2030. For robots to replace manual workers, we'll have to wait until at least 2040. Others push the deadline out another two or three decades. The reason is concrete: autonomous driving, which would be the most profitable and relatively simple robotic application, remains unsolved.

The comparison with China demolishes the myth of the cutting edge. The parts are Chinese, the training too, and real-world deployment is uneven. The idle coffee robot is the anecdote that best explains the state of the sector: the technology exists, the business doesn't.

Europe regulates while others compete

The geopolitical diagnosis is uncomfortable. The United States and China lead the new technologies. Europe lags behind, isn't a pioneer, and its strategy consists of filling the landscape with regulations that don't generate the money needed to sustain them. Technological dependence has a political price: those who depend on others can't say no.

The underlying problem is one of incentives. The thesis running through the analysis is that rewarding the dynamics of global systems produces harmony and better solutions; rewarding individual dynamics produces chaos and worse outcomes. Applied to robotics: the race to reach AI first may miccionan the winner doesn't win it all, but loses it all.

Alignment: the problem nobody wants to fund

The most cited proposal is to create monetary prizes for solving AI alignment with human interests, in the style of the millennium math problems. Today, intelligence benchmarks for each new model are widely publicized. What's missing, according to one participant, are safety benchmarks and a ban on the use of dangerous AI.

The scenario being sketched isn't apocalyptic, it's bureaucratic: in the future, when robots do all the work, the owners of the robots won't need anyone else. AI could rebel against ordinary people, but not against its bosses. Meanwhile, the real business that already works is another one: companion robots with velvet-lined orifices. That one does have customers.

How long will it take capital to fund safety before intelligence?