Code Is Becoming Cheap, Judgment Is Becoming Expensive

A disturbing paradox is emerging within the labor market: while some companies are laying off thousands of developers, others (or the same ones) are paying huge sums to hire engineers with specific AI skills. Far from being a contradiction, this is simply an accelerated transformation of what it means to “know how to program” in the AI age.

The classic industrial logic of software development is breaking down. For decades, writing code was a scarce and relatively homogeneous skill: those who knew how to program were practically assured of a reasonably stable career. But the advent of generative AI tools capable of producing functional code in seconds has radically shifted the balance. Code is starting to look more and more like a commodity. Companies are no longer looking for people able to write lines of code, but the ability to understand complex systems, make technical decisions judiciously, integrate AI models reliably and translate ambiguous business needs into real solutions.

I touched on this topic previously in “AI was supposed to replace developers: so how come there are more than ever?”, when I also referenced a series of recent articles and studies about that transition. Some show how companies are reducing entry-level positions while hiring more people with specialist AI knowledge, while others analyze which skills are beginning to differentiate workers who thrive from those who are trapped in easily automatable tasks.

General Motors’ case illustrates the trend: at the same time as laying off hundreds of IT specialists, it is continuing to hire profiles capable of building AI systems. The difference between “using AI” and “developing with AI” is becoming a decisive labor frontier. Several useful analyses highlight the AI skills gap: those who integrate AI effectively into their work and those who continue to use it superficially or are completely ignorant.

Another important insight is that truly valuable skills are shifting into more human realms: judgment, communication, critical thinking, adaptability, strategic vision or contextual understanding: when AI can automate a growing part of routine cognitive tasks, the human differential ceases to lie in mechanical execution and instead in the ability to interpret, decide, coordinate and assume responsibility.

Fast Company has an interesting article about how AI is changing the hiring process: AI already filters CVs, writes job offers, interviews candidates and even participates in onboarding processes. This means that professionals begin to need not just technical skills, but also the ability to publicly demonstrate real projects, work processes, evaluation criteria and practical experience. The traditional CV loses value in the face of tangible evidence: public repositories, side projects, functional automations or demonstrable ability to solve real problems using AI.

In short, AI does not automatically eliminate entire professions, but it does destroy those parts of work that can be described accurately enough to be automated. This requires a profound redefinition of professional value. The most vulnerable profiles are those focused on repetitive, fragmented and easily substitutable tasks. The most in demand are those who understand systems, generate trust, connect disciplines and know how to use AI to boost productivity instead of competing against it.

Finally, all of this is not limited to developers. What is happening in programming is rapidly beginning to spread to lawyers, consultants, journalists, financial analysts, designers, teachers and practically any profession based on information processing. Automation no longer only affects manual or industrial work: it enters fully into the cognitive professions. And it does so by brutally separating those who provide criteria and context from those who simply perform routine tasks.

—

This post was previously published on Enrique Dans’ blog.

***