Is life a form of account?

IN 1994, a strange and pixel machine appeared on the computer screen. I read a series of instructions, copying them, and building a cloning of himself-as expected by the American Hungarian Polima John von Neuman half a century ago. It was an amazing demonstration of a deep idea: that life, in essence, may be accountable.

Although this is rarely fully estimated, von Newman was one of the first to put a deep link between life and reckoning. Reproduction, such as the account, as shown, can be implemented by machines that follow encrypted instructions. In its model, based on the global Aolan Turing machine, self -repetition systems read and implement instructions as DNA do: “if The following instructions are cga codon, then Add arginine to the protein under construction. “It is not a borrowing to call the DNA” program ” – this is literally.

Of course, there are significant differences between biological computing and the type of digital computing by the PC or your smartphone. DNA is invisible and multi -layer, including phenomena such as genetics and proximity effects. Cellular DNA is not anywhere near the entire story. Our bodies contain (constantly swap), an countless number of bacteria and viruses, each of which runs its own symbol.

Biological Computing “”Parallel“Decentralization, and Sufficient. Your cells have a place in the 300 neighborhood Quintillion Ribosomes, all work at the same time. Wonderfully complex floating protein factories, in fact, is a small computer – albeit a random device, which means that it cannot be fully predicted.

Ferry components movements, capturing and release of smaller particles, and processing all chemical bonds are all randomly, reflected, and not vibrant, through continuous thermal buffet. Just a statistical lack of consistency prefers a direction over the other, with smart origami moves that tend to “lock” some steps so that the next step becomes likely to occur.

This differs greatly from the operation of “logical gates” in the computer, the basic ingredients that treat bilateral inputs in the outputs using fixed rules. It is irreversible and engineer to be 99.99 percent reliable and repetitive.

It is not a borrowing to call the DNA “program” – and this is literally.

Biological computing is computing, however. And its use of randomness is an advantage, not a mistake. In fact, many classic algorithms in computer science also require random (albeit for various reasons), which may explain the reason for the TURING statement that Ferranti Mark I, an early computer that helped design it in 1951, includes random number instructions. Thus, randomness is a small, but important conceptual extension of the original Torring machine, although any computer can simulate its numbers by calculating the inevitable numbers but with a random or “liar” appearance.

Parallel, also, is increasingly essential for computing today. Modern Amnesty International, for example, depends on both the huge parallel and Random – as in the parallel “SGD” algorithm, used to train most of the nerve nets today, the “temperature” settings used in Chatbots to enter a degree of randomness in its output, and parallel to the graphics processing units (GPU), which most AI’s strength in data centers.

The traditional digital computing, which relied on the central implementation of the instructions, was a product of technological restrictions. The first computers needed to implement long accounts using the lowest possible parts. Originally, these parts were cortical: they were expensive vacuum tubes, for example, they tended to combustion and repeated by hand. The natural design, then, the CPU (CPU) (CPU) was working on the bits that were transferred back and forth from an external memory. This has become known as “von Neumian brown”.

Both Torring and the Neumman realized that computing could be made by other means. Torring explores, near the end of his life, how biological patterns such as leopard stains can arise from simple chemical bases, in a field called formation. The Torring model of formation was a biological -inspired shape of the highly parallel account distributed. It was also his previous concept of an “unorganized machine”, which is a random nerve network similar to the infant’s brain.

These were visions of what might look like computing without a central processor – and what it is Do It looks like, in living systems.

Von Newman also began exploring parallel methods in a great way for the account until the 1940s. In discussions with the Polish mathematics scientist Stanisao Ayam in Los Alamos, the idea of ​​”cellular automatic” is depicted, pixel -like networks of simple mathematical units, all of which obey the same rule, and everything that changes its cases simultaneously by communicating only with direct neighbors. With the distinctive Bavorra, von Newman went to the extent of design, on paper, the main components of Self -reproduction Cellular autoaton, including a horizontal “tape” of cells that contain the instructions and masses of the cellular circles for reading and copying and implementing them.

Automaton cellular design is much more difficult than regular programming, because every cell or “pixel” changes its condition and environment simultaneously. Add randomly and impacts of micro -reactions, as in biology, and it becomes difficult to think about the “program” or “Debug”.

However, Turning and Von Neumann have absorbed something essential: the account does not require a central processor, logical gates, or bilateral or serial mathematical programs. There are countless methods for an account, decisively, are all equivalent. This insight is one of the greatest achievements of theoretical computer science.

The “basic system independence” or “multiple investigation” means that any computer can simulate any other device. If computers have different designs, the simulation may be slow. For this reason, the autonomous Automaton autonomous von Newman-although this will be fun!

That demonstration in 1994 –The first successful simulator to automate von Newman– It cannot happen early. The serial computer requires a dangerous treatment force for the throat through the Automaton 6,329 cells during 63 cells one billion The time steps required to complete the reproductive cycle. On the screen, I worked as it was announced: a two -dimensional two -dimensional machine, wandering in an educational tape with a length of 145315 cells to the right, pumping information outside the tape and communicating with the “writing arm” slowly to print a working cloning from the top to the right to the right.

It is likely that the serial computer mimics a parallel nervous network, and the heir of a “non -organized” top. Consequently, the operation of large neurological networks such as those in the convertible Chatbots has become practical only, thanks to the continuous progress in miniaturization, speed and parallel to digital computers.

In 2020, my colleague Alex Morfinsv Completely neurological networks, formation in Torring, and iPhone cellular automation in “Nervous cells” (nervous automatic “(NCA), replacing the simple pixel base for each pixel for classic cell automation with a nervous mesh. This network, capable of sensing and influencing some values ​​that represent local morphine concentrations, can be trained on “growth” any style or desired image, not only brutal donkey lines or tiger spots.

Real cells do not literally contain nervous networks within them, but they run very sophisticated, non -written and ugly “programs” to make a decision on the measures they will take in the world, while giving external motivation and internal condition. NCAS provides a general way to mix a set of potential behaviors of cells whose actions do not include movement, but only changes in the case (here, represented in color) and absorb or release chemicals.

The first NCA Alex showed me was from the emoji lizard, That can renew not only its tail, but also its ends and head! It was a strong show of the complexity of the multi -cell life “thinking locally”, but “behavior worldwide”, even when each cell (or pixel) operates the same program – completely as each of your cells runs the same DNA. Simulation like this shows how an account can produce vibrant behavior through standards. Based on the designs of Von Neumann and extends to modern cellular nervousa, it provides a glimpse of the mathematical foundations of living systems.

This story is reprinted with permission from Reader. It was adapted from “What is intelligence?

To read more about the mathematical basis for life and the ideas of Alan Torring in NutelosCheck these stories:

In the beginning, there was an account: Life is a symbol, and the symbol is life, in nature as it is in technology.

The man who tried to recover the world with logic: Walter Bates rose from the streets to the Massachusetts Institute of Technology, but he could not escape himself.

Torring patterns appear in a small crystal: The idea of ​​Alan Torring in 1952 extension of leopard stains to the atomic domain.

Lead: O-IHI / Shutterstock

• Blaise Agüra y arcas

  Spread on September 24, 2025

  Blaise Agüra Y Arcas is the Vice President/colleague in Google, where CTO was for technology and society, and the founder of intelligence models, an organization dedicated to basic artificial intelligence research. He is the author of the book ” What is intelligence?and Which is adapted to this article.