AI and the Human Experience: An Irreplaceable Duo

Reality as a Competitive Advantage

David Beyer, known for his insightful analyses, highlights an essential tension in our technological era. He compares two medical companies to illustrate his point. The first, specializing in radiological image processing, benefits from an ideal application of AI. Images are static data, and with sufficient computing power, accuracy becomes accessible to all. This type of problem is solved by AI due to its ability to efficiently process fixed data.

In contrast, the second company, which manages medical insurance claims, faces a dynamic environment. Regulations are constantly changing, policies evolve, and billing codes are updated. This context requires operational knowledge gained only through direct experience, which Beyer refers to as "scar tissue." This knowledge cannot be simulated or learned outside of the real context, as it develops through interactions and continuous adjustments in response to system rejections and errors.

AI can certainly accelerate learning in environments with fixed rules, but it cannot replicate the unpredictability of the real world. It also cannot influence the speed at which regulators change their rules or anticipate competitors' moves. Thus, the speed at which a company learns and adapts is dictated by reality itself, not by the computational speed of machines. Reality, with its complexities and unpredictabilities, then becomes a competitive advantage that is difficult to imitate.

The Adoption Crisis: Recursive Technology ≠ Recursive Adoption

AI models progress recursively, with each generation of models improving upon the last. This advancement is undeniable and impressive. However, many extrapolate this recursivity to the economy, anticipating a rapid and massive replacement of the human workforce. Citadel Securities, in its analysis of the "Global Intelligence Crisis of 2026," deconstructs this idea by emphasizing that technological recursivity does not automatically translate into equally rapid adoption.

The adoption of AI in the real world is hindered by factors that do not progress at the speed of software. Among these factors are the physical capital needed to build infrastructure, the capacity of the electrical grid, regulatory approvals, and, most importantly, organizational change, which is often the slowest. To illustrate these constraints, one can observe manufacturing construction spending in the United States. Between 2021 and 2024, this spending surged from $75 billion to over $240 billion, a historic record. This leap reflects the need for massive physical support to accompany the promise of AI, a process that spans years.

Historically, productivity shocks related to AI have been positive supply shocks. They reduce marginal costs, increase output, and raise real income. John Maynard Keynes predicted that productivity gains would allow us to work only 15 hours a week in the 21st century. This prediction has proven false, as it underestimated the elasticity of human desire. As technology reduces costs, we do not stop working; we expand our consumption, demand higher quality, and create previously unimaginable new industries.

Recent data confirms this trend: since 2020, the creation of new businesses in the United States has reached unprecedented levels, remaining historically high. Far from reducing human activity, technological changes stimulate innovation and entrepreneurship. Moreover, contrary to fears of massive replacement, the demand for technical skills, such as software engineering, has stabilized at pre-pandemic levels, highlighting that technology complements our work rather than eliminates it.

Will AI Replace Us? The Wrong Question

The idea that AI could replace all humans is often debated, but it poses the wrong question. Instead of asking whether AI will replace us, it would be more relevant to consider how it is redefining our work and skills. AI, by automating certain tasks, frees up time for humans to focus on more creative and strategic activities. It does not replace humans but alters the nature of work, requiring continuous adaptation of human skills.

Ultimately, AI and humans are not in direct competition. AI is a powerful tool that, when used well, can amplify human capabilities and open new perspectives. The real question is how we can leverage this technology to enrich our work and our society.