Estimated reading time: 5–6 minutes
“And what good did it do us?” an imaginary reporter asked me. “That one I’m throwing out,” I thought furiously.
On 25 March 1961, the MECIPT-1 computer (the Electronic Computer of the Polytechnic Institute of Timișoara) was brought into service.
I learnt later that the Faculty of Computer Science and Economic Cybernetics at the Academy of Economic Studies (ASE) in Bucharest had been one of the first institutions in the world to introduce cybernetics and computing officially as fields of study. (1964)
To put things in perspective: the Massachusetts Institute of Technology (MIT) established its Department of Computer Science in 1962, Stanford University established its Department of Computer Science in 1965, and in Europe things moved even more slowly — the University of Cambridge did not establish its computer laboratory until 1970.
In 1964, within the Faculty of General Economics at ASE, the section for “Mechanisation and Automation of Economic and Statistical Computation” was founded, almost simultaneously with the American pioneers. I entered this section in 1966, and the students in the years above me would tell me about the Cybernetics course.
Without realising it, we were pioneers in a field that changed the world, much like the researchers who work in AI today.
“And what good did it do us?” an imaginary reporter asked me.
“That one I’m throwing out,” I thought furiously.
Economic cybernetics, computers in the sense of digital processing as we know it today, and their programming were revolutionary fields at the time, combining systems theory, applied mathematics and information technology to optimise economic processes. The students of this faculty, myself included, were among the first in the world to study these new technologies in the applied context of economics and management.
This early start to computer science education in Romania made ASE one of the world leaders in economic cybernetics and information technology. The communists were not entirely stupid in this regard.
Looking back, it is very likely that they acted according to a plan. All my lecturers who specialised in information technology were young and had been educated in the United States. In order to be able to teach specialist subjects from 1966 onwards, they had probably been sent to train in the early 1960s, shortly after the advent of FORTRAN, COBOL and ALGOL.
If there was a plan, it was a good one.
And I, a proud student of the Faculty of Economic Computation and Economic Cybernetics, was destined, without knowing it, to become part of the first generation of Romanian programmers who would turn bits and algorithms into their profession.
Later, I came to understand what had actually been happening. The communists had been playing both ends:
— The Romanian Encyclopaedic Dictionary (the old edition, begun in 1958 and published between 1962 and 1966) initially defined cybernetics as: “a bourgeois theory of an idealist, reactionary nature, which attempts to mystify the real processes of nature and society through mechanistic analogies between automatic machines and living organisms or societies.” — On the other hand, on 25 March 1961, the MECIPT-1 computer (the Electronic Computing Machine of the Polytechnic Institute of Timișoara) was brought into service. A Romanian paradox.
But let us not get ahead of ourselves; the story is a long one. That is to say, it is long for me — in truth, it is surprisingly short, if we are speaking of the evolution of knowledge.
Now, at seventy-seven, whilst creating AI models, I feel a new beginning. Even now, as I write these lines, the programme that has been tormenting me for a week is running in parallel, fine-tuning an AI model. It is rather like taking a child to school, except that you must first write the textbooks yourself. I have an Alienware laptop — 64 GB of RAM and an NVIDIA GPU with 16 GB of video memory. The entire MECIPT computer, which filled an entire room, possessed an infinitesimal fraction of the computing power of this laptop that fits into a briefcase. Or the Felix C-256, whose disc units would shudder in place whenever you ran an indexed search — you would swear they were trying to leave without permission.
From punched tape to neural networks. From magnetic drums to cloud computing.
Before, you had to think like a machine. Now you teach a machine to think.