Ivo Vaske: bits move fast, atoms move much slower
Analyst Ivo Vaske argues that the spread of artificial intelligence does not necessarily mean the disappearance of human labor, since automating the physical world requires far more capital, energy and time than automating digital work.
Estonian commentator Ivo Vaske writes that the rapid rise of artificial intelligence does not mean human labor will soon vanish — rather, it will shift toward areas where people remain a cheaper or more flexible option than machines. He recalls a programmer acquaintance worrying about how wealth would be distributed in a world where AI performs nearly all work, but argues that such large-scale redistribution is still far off.
Software versus physical capital
Software, Vaske notes, can be copied at almost no additional cost, but an excavator, a truck or a robotic arm cannot. Every physical machine requires steel, copper, semiconductors, energy, factories and maintenance. That is why even very fast AI progress may reach the real economy far more slowly. The world already holds an enormous stock of physical capital — cars, trucks, tractors, factories, ports and construction equipment — and how quickly it gets replaced depends on price, reliability, available capital, infrastructure and regulation.
Autonomous production chains
A fully universal, self-replicating robot is not a prerequisite for a major economic shift, Vaske argues. It is enough for increasingly long production chains to become autonomous — robots assembling robot parts, automated mines extracting raw materials, autonomous transport moving them to factories. The smaller the human share in such chains, the faster physical capital can grow. But this creates new bottlenecks: energy, raw materials, land and production capacity become more critical constraints, since robots cannot override the laws of physics.
Vaske concludes that two separate revolutions should be distinguished: automation of information work can happen very quickly, while automation of the physical world depends on how fast machines, energy production and infrastructure can be built up. A world where autonomous machines with minimal human involvement could plan and build a project the scale of Rail Baltic is not necessarily fantasy — but the path there, he says, is considerably longer than the path to an AI capable of designing such a project.
