The matching machine
In 1950, André Maurois ponders, in the Nouveaux discourses du Dr O’Grady, upon his time’s evolutions: atomic energy, sexual liberation, existentialism, birth of computer science, psychiatry and the progress of medicine, imperialism and space research…
The intellectual enthusiasm of Socratic dialogues and the amazement of Alain’s words live in his discourses. In “The future belongs to rats”, he writes:
- This is our time! he said. How can you expect this world to work properly when a cultured Frenchman ignores the name and the work of one of the greatest intellectual organisers of this time? My dear Aurelle, Dr. Vannevar Bush is the man who made possible, by his method, the use of atomic energy. Now he has written a revolutionary article on the superman’s working processes. Tomorrow he will have at his disposal machines for computing and even for thinking, so complex, so perfect, that they will free him from the mechanical side of mathematics and logic. They will be the “housekeepers” of the scientist; they will solve, in a few minutes, equations with such a number of unknown variables that a human team would spend years working on them in vain. The superman will possess libraries on microfilm so small that all the books published since there have been men who write will fit, Aurelle, in your room. The great Larousse will be thinner than a matchbox. Each microfilmed book will have its own code number. By forming this number, you will make the title page appear on a screen placed in front of your desk and, if you wish to find, in the volume, a passage or an information, you will have in front of you changes of speed which will allow you to pass one, ten, or hundred pages per minute.
- What a shame!” said Aurelle.
- That’s not all,” said the doctor. All these books will be linked together by ideas’ matching machine which will make available to you, in a matter of seconds, if you want to research, for example, the treatment of shingles by Tibetan doctors, or the role of pederasty in the founding of empires, everything that has ever been written on these vital subjects.
- Are you serious, doctor? And is such a machine possible?
- But of course,” said the doctor. Have you not seen in the American armies those trucks containing the cards, carefully cut and filed, of every man in a division, and which, at the touch of a button, could at once produce a complete list of all the soldiers who were blacksmiths, Baptists, and Finnish speakers? The principle will be the same. But there will be more. The superman, when he wanders around his laboratory, will have a walkie-talkie in front of his mouth, a mobile microphone to which he will entrust his observations; these will be immediately typed up from a distance, photoelectric cells transforming the sounds into signs; he will have on his forehead a photographic camera, not bigger than an olive, which will record on tiny films what the observer sees. Thus everything that is said and done in the world will be fixed and classified in microfilms’ libraries.
- So that, doctor, nothing will be lost any more; that the beneficial filtering of oblivion will no longer take place; and that the archives of humanity will swell at an accelerated rate, to the point where no one will be able to consult them usefully… And you call this the time of the superman? I fear that this is a new stage in the decadence. Plato and Descartes had no walkie-talkies, no microfilms, no card vans…
- And you see,” said the doctor, sarcastically, “where Plato and Descartes have led us. I think we should now give Dr Vannevar Bush a chance.
- What can he do for us?
- Everything, my boy. He can provide our brains with the necessary relays to master our new knowledge; he can give us machines to predict, machines to govern… But yes… Many problems of government are in fact problems of calculation; but the data are so abundant that no one dares to undertake the analysis of situations. Instead of counting, governments and peoples choose to rant, shout, or fight in the streets. The machine would make everyone agree.
- Example, doctor?
- There are thousands of examples. Your farmers are asking for a “fair” price for wheat or beetroot. Give the machine all relevant data: harvest, world price, industrial prices, legitimate profit, etc. It will give you an irrefutable figure. It is easy to imagine a banking machine. It would tell you the reasonable discount rate in such and such circumstances, the volume of credit justified by production, the respective values of currencies.
- O Doctor Right! You reason as if the real data were not human passions. A crisis, even an economic one, is not essentially the result of figures, but of hopes, fears and errors.
- And where did you learn, my boy, that passions are not subject to statistical laws? The governing machine will take thoughts into account. Not of the individual, but of the masses. Think about it, Aurelle. In all the D.C.A. stations there are machines solving this problem: an aeroplane of a certain type flies at such and such a speed, in such and such a weather; its position at the moment was known; what will be its probable position at the moment t’? This position t’ depends not only on the aircraft, but also on the aviator who knows he is being shot at and who is manoeuvring. Yet the machine is efficient; it takes into account the reactions of the individual. The probability will become even greater when the reactions are those of crowds.
- In short,” said Aurelle, “you’ve come to the conclusion that the planet will be easier to govern than the Principality of Monaco?
- That is obvious,” said the doctor.
It is useful, for a generation born with computers, to take a step back and reconsider what last century’s thinkers imagined. They saw in the ever-growing archive of ideas, or data as we would say today, a formidable source of knowledge to harness. What they lacked was the “matching machine”. Such a machine would necessarily also be a forecasting machine, and therefore also a governing machine, if we believe, with Dr O’Grady, that the masses’ ideas can be predicted.
In this discourse, the seeds of statistical rather than procedural reasoning appear, the invention of ‘machines for associating ideas’ and predicting probabilities, rather than revealing a single truth hidden in the data.
To my mind, this is the most ambitious and the richest definition of artificial intelligence: a matching machine, based on statistics, able to predict and govern.
For a definition of A.I.: read again Maurois’ discourse, replacing “ideas” with “data”, “microfilms” with “big data” and “machine” with “software”. Everything’s there: the weak A.I. as a “scientist’s housekeeper”, automatizing repetitive tasks, and the strong A.I., identifying patterns in series of data and able to make predictions that would take human reactions into account.
