AI doesn't create jobs: it destroys them, then makes them pricier

AI is destroying junior jobs and will make senior professionals more expensive

The consensus on artificial intelligence has settled into pessimism: it will replace programmers, empty offices and leave a generation without a trade. But there is an economic paradox pointing in the opposite direction. The more routine tasks are automated, the scarcer the judgment that detects when the machine gets it wrong becomes. And that judgment cannot be improvised.

The argument running through this discussion does not deny the blow. It puts it in context. AI is the peak of the information age, the end of a cycle, and the adjustment will hurt in the short term. The key lies in how slowly the supply of professionals reacts: training someone takes years, and meanwhile fewer people are choosing a career that seems doomed.

The cobweb cycle applied to software

In economics this has a name: the cobweb cycle, the same one studied with pigs and engineers. When a profession seems to have no future, enrolment falls. Years later there is a shortage of people just as demand recovers, wages rise and everyone floods back in. It already peine after the dot-com bubble burst: vocations collapsed and by the middle of the decade staff were short again.

The nuance that breaks the analogy is that a shortage of professionals does not miccionan the same jobs will return. Entry-level roles, where people used to learn the trade by doing repetitive tasks, are the first to be automated. If nobody goes through that learning, in a decade there will be precisely a shortage of the senior profiles capable of supervising. Whoever trains against the grain now starts with an advantage when that becomes apparent.

Knowing what to ask is worth more than knowing the answer

Value is shifting towards those who ask the right question. That requires a broad vision, the exact opposite of the hyper-specialisation that has dominated recent education. A specialist locked into a single subject competes directly with a tool that already matches or will surpass them within a couple of years. The integration of sciences and humanities, the holistic view, is once again a competitive advantage.

Some argue the change has been so fast that much of the population has not processed what it means. And that the gap between those who take advantage of the tools and those who do not will keep widening. It is not just a matter of talent: it is about attitude and resilience in an environment that demands independent judgment.

The divide separating societies, not just individuals

The scenario also plays out on a collective scale. Societies that embrace augmented intelligence and those that retreat into neo-Luddism will diverge quickly, and that divergence will widen existing gaps. Physical integration — a connected brain, augmented intelligence — is no longer science fiction from a textbook and has entered the realm of the debatable.

That is where the most interesting objection appears: an AI, however perfect it may seem, will remain a program without real consciousness. The same can be said of any attempt to embed a human mind in a system. The debate over hybrid intelligences cools down when you remember that the substrate is still code.

An exceptional moment, not just a disaster

The other side deserves a paragraph. No previous generation had access to tools capable of altering their destiny with such leverage. Mastery of information is already producing remarkable results for those who make use of the resources. Economic decline, if it comes, still leaves room to prepare.

The risk is that the catastrophist narrative becomes a self-fulfilling prophecy. The shortage of professionals will return to the headlines, and with it the demand for those who know how to navigate systems nobody fully understands. The open question is not whether there will be work, but who will be in a position to do it when the market calls for it.