Folha: ‘Random numbers and quantum mechanics’
Reproduction of Marcelo Viana’s column in Folha.
In an article published in the scientific journal Nature in March 2025, a group of 32 researchers announced that they had managed to produce genuinely random numbers using a computer. The problem dates back to the advent of electronic calculation in the middle of the last century, but until recently had no satisfactory solution.
Many uses of information technology (in cryptography, for example) depend on being able to produce random numbers quickly and in abundance. It’s also important that the product can be certified: when a customer (a bank, say) buys random numbers from a supplier, they need guarantees that they are good (really random) and fresh (not previously used).
But traditional computers are incapable of generating random numbers. The best they can do is create sequences of pseudo-random numbers, calculated deterministically from “seeds” extracted from physical phenomena: the randomness is all in the seeds. A curious example: the cybersecurity company Cloudfare extracts cryptographic seeds from the evolution of lava lamps (you know those decorative lamps with colored wax that goes up when it gets hot and goes down when it gets cold?)
With the advent of quantum computers, this is changing. The reason is that, unlike classical physics, quantum mechanics is intrinsically random. For example, the spin of an electron is determined randomly at the time of measurement, as if the universe were drawing heads or tails between the two possible values, -1/2 and +1/2.
Although the engineering of the quantum computer still faces many difficulties, there are already several working prototypes, some of which are accessible on the internet. The team behind the study published in Nature used one of these machines, a Quantinuum H2-1 with 56 qubits, to generate a sequence of 70,000 random bits (zeros and ones). This is equivalent to a text of some 7,500 characters, more than twice the length of this column.
How can we verify that these bits are really random, that they are not the result of a traditional, deterministic algorithm, but with information hidden to make it appear random? That’s where things get even cooler!
In 1935, physicists A. Einstein, B. Podolsky and N. Rosen pointed out that, according to quantum mechanics, particles that have met at some point remain related (“entangled”), in such a way that a measurement made on one of them influences the other instantaneously, even if they have moved very far apart in the meantime. To avoid this “ghostly action at a distance”, considered aberrant, it has been proposed that quantum mechanics is incomplete: the meeting of particles would create links between them that the theory does not take into account (“hidden variables”), which would satisfactorily explain the fact that they are related.