Edward Fredkin and Computational Reversibility

THE PROBLEM OF REVERSIBILITY

If, as Edward Fredkin, we think that the universe is ultimately a gigantic computer, which processes information according to that particular software which are the laws of physics, we have a problem: the problem of reversibility.

I computer in fact they are not reversible: once the inputs have been processed and the final result has been reached, it is no longer possible to go back to the inputs.

In other words, if I add 5 and 4 from the 9 I get, it is not possible to know which numbers I have actually added, they could be 6 and 3, or 8 and 1 and so on for all possible combinations.

For this reason it is said that computers waste information and dissipate it as heat.

The heat that our computers give off is nothing more than information going to waste.

THE CANCELLATION OF BITS

And this deletion of bits, discovered by Rolf Landauer in 1961, it is formally and rigorously expressed in the formula called "Landauer limit”, which we do not report due to its obscurity, but which is accepted by the scientific community and which allows us to calculate exactly the heat dissipated as a function of the bits erased.

It might all seem okay, but it really isn't.

The laws of physics, in fact, are reversible and describe a substantially deterministic universe.

This means that if I could know the position and velocity of all particles, I could not only predict the future, but also go back to previous states of the universe. But for this to be possible, it is absolutely necessary that all information about the past be preserved in some form.

It follows that the idea that the universe is a giant computer contradicts two basic principles of physics: the reversibility of physical laws and the conservation of information (which are two sides of the same coin).

This is the dilemma that Edward Friedkin, deeply convinced of the computational nature of physical phenomena, is faced with.

FRIEDKIN'S LIFE

Fredkin is the son of Russian immigrants who met in Los Angeles, the city where Edward was born in 1934. He attends high school reluctantly, and manages to get admitted, with very poor grades, to the Caltech, who, however, to be on the safe side, immediately dropped out in the first year to join the Air Force.

This is where he comes into contact with a huge IBM computers: it's love at first sight and Edward immediately learns how to program it. Then he joins the MIT Lincoln Laboratory and resumes his physics studies.

He later left the Air Force and worked for a large IT consultancy firm, which he soon left to found nel1962 la International Information Incorporated, a high-sounding name for a company with no assets, no customers and only one employee, and not even a graduate.

Soon, however, Fredkin developed a system to translate optical signals into digital bits by modifying a simple projector. The device was a huge success among some business customers and the III it is listed on the market.

It is 1968 and At just 33 years old, Fredkin becomes a billionaire. He buys a ranch in Colorado and an island in the Caribbean.

Also in 1968 Edward was called, at the suggestion of Marvin Minsky, to collaborate with the MAC (Machine-Aided Cognition) MIT Project, of which he became director in 1971.

In 1974 he returned as a professor at Caltech, sponsored by Richard Feynman.

After another brief stint at MIT, he taught physics at Boston University and held various positions until his death in 2023.

COMPUTATIONAL NON-REVERSIBILITY

Fredkin made two major contributions to theoretical computer science, both of which formulate an answer to the dilemma of computational non-reversibility:

• The Billiard Ball Computer
  A computer designed as a billiards table, with balls that interact with each other, as particles do in physics. Instead of relying on electronic signals, it uses the motion of the balls as vectors of information. In fact, this type of computer is completely reversible and tends to reduce its entropy to zero, not dissipating information and therefore heat. The billiards ball computer was never practically built, but what mattered was its logical possibility.

• The eponymous logic gate
  Fredkin's other invention is the eponymous logic gate: by combining Fredkin's gates we obtain a universal computational architecture and also in this case absolutely reversible. Therefore, reversible computational machines exist: consequently we can imagine that the universe is one of these: an enormous reversible Turing machine; in this way the apparent aporia between the digital conception of the universe and the reversibility of the laws of physics is eliminated. For Fredkin, the universe is digital, it works like a reversible computer, it contains all the information that has been processed and is therefore completely deterministic: if we knew the state of all the particles we would be able not only to predict the future, but also to reconstruct the entire past.

In addition to the important implications for physics, Fredkin's theories on computational reversibility are also confirmed by the most recent developments and problems of theoretical computer science: quantum computers are in fact reversible; and the issue of energy consumption, and therefore of entropic dissipation of information, is on the agenda especially in connection with the most recent developments. of Generative Artificial Intelligence, which cause an enormous waste of energy.

We thank Paolo Riccardo Felicioli for his contribution