For a long time, it was believed that robots would take over strenuous physical labor, while humans would handle more intelligent tasks. The opposite has happened.
The release of ChatGPT brought a new paradigm in the use of artificial intelligence, shifting its applications from narrowly specialized tools to assistants with the potential to replace a large portion of intellectual effort. Since its public debut in 2022—when it was an interesting tool for generating text, providing unreliable answers, and similar uses—it has slowly become an indispensable tool in many industries.
Science fiction authors more often envisioned servant robots, flying cars, and automated industrial systems before solving the puzzle of human intelligence and the work that stems from intellectual capacity. The expectation was that human intellectual capacity would be a difficult problem, whereas human physical traits would be simple to replicate. Robots were supposed to free us from hard physical labor, while intellectual work and creativity were reserved for humans.
Contrary to popular belief, computer scientist Hans Moravec proposed in 1988 that it is easier to get a computer to perform at the level of an average adult on intelligence tests than to give it the perception and mobility skills of a one-year-old child. Observing computers embodied as humanoid robots that still cannot walk on difficult terrain, compared to artificial intelligence like ChatGPT, it’s clear Moravec was absolutely right.
According to a 2023 scientific paper, an older class of AI models already outperforms doctors in making diagnoses—even when those doctors use AI themselves. The latest class of models achieves impressive results, such as passing bar exams, excelling in programming competitions (scoring better than 99.8% of participants), ranking in math olympiads, and more.
Knowledge and intellectual labor that used to be distributed across many different individuals and industries are now available on a laptop that costs just a few thousand euros. Although there are still many critics and shortcomings, it’s important to note that the current limitations of AI seem to stem only from computing resources, human acceptance, and bureaucracy. As a programmer, I’ve watched ChatGPT “grow up” from an intriguing toy to an indispensable tool within a year—something that should prompt all of us to reflect on the future of human intellectual labor.
The robots that were supposed to free us from heavy physical labor are not that close, but those equipped with the knowledge to perform intellectual tasks at a human level are already here. The replacement of human labor may not happen very quickly, as certain technical and social prerequisites still need to be met. In some industries, the transformation is already underway and will unfold naturally; in others, it may happen overnight; while in some, it may take many years before the market forces all stakeholders to adapt.
The future is impossible to predict, but these changes should prompt each individual to reconsider their position in the labor market. A moment may be approaching when capital will have less and less need for human labor—raising the question: in the near future, will it still be possible to acquire capital through labor?
About the Author
