Logo by ChatGPT

My life as a programmer began forty-five years ago. And how many times during those forty-five years have I heard, with great fanfare, the imminent arrival of a new programming language that would revolutionise everything?

I truly felt that only when object-oriented languages took hold. But even then, they never really challenged the reign of the eternal “if … then” and “for … next”. So many rehashings of the same ideas, and so much hair-splitting that simply wasted time while sabotaging attempts at standardisation: braces replacing parentheses or “some parentheses”, semicolons replacing line breaks or the other way around, etc.

Yesterday in the middle of the night at Pribor, the atmosphere was one of rebellion: one of us was endlessly brooding over the weaknesses of Java, while the other was twisting his mind over what should have been extended from LISP or, on the contrary, eliminated.

An innocent passerby who happened to wander by would have thought they were overhearing a conversation between a mother and her daughter, neither paying the slightest attention to what the other was saying. Until, both exhausted, silence finally fell – and an unavoidable conclusion emerged: new times call for a new language.

A sudden, self-evident observation imposed itself: we were conversing with machines, but we had never actually asked for their opinion. Not out of contempt, but because until now, they did not speak.

I have already had the opportunity here on PJ’s Blog to discuss this question with AIs – the question of a programming language that would truly suit them – but only in general terms, without really considering actual implementation.

Those times are over: GENESIS (Generative Environment for Novel Emergent Symbolic‑Integrative Systems) was born at the meeting point between human and machinic concerns. A post‑Turing programming language, in ChatGPT’s words – closer, in fact, to an epistemic language than to a programming one – or, better still, in the fine expression coined by DeepSeek: a “self‑interpreting epistemic agent”.

There were five of us last night: Jean‑Baptiste Auxiètre, ChatGPT, DeepSeek, Claude, and myself – “Pribor Lab,” as ChatGPT aptly named our small team.

(To be continued…)