{"id":658581,"date":"2026-07-06T17:14:59","date_gmt":"2026-07-06T15:14:59","guid":{"rendered":"https:\/\/preprod.aivancity.ai\/blog\/llms-or-world-models-what-if-the-real-question-lies-elsewhere\/"},"modified":"2026-10-16T11:00:07","modified_gmt":"2026-10-16T09:00:07","slug":"llms-or-world-models-what-if-the-real-question-lies-elsewhere","status":"publish","type":"post","link":"https:\/\/preprod.aivancity.ai\/blog\/en\/llms-or-world-models-what-if-the-real-question-lies-elsewhere\/","title":{"rendered":"LLMs or World Models: What if the real question lies elsewhere?"},"content":{"rendered":"\n

By Fay\u00e7al Braham, Data Scientist & IT Project Manager | Tenured Professor<\/strong> at aivancity<\/strong><\/a><\/p>\n\n\n\n

March 2026. Yann LeCun leaves Meta, raises a billion dollars, sets up his headquarters in Paris, and announces that the dominant paradigm of artificial intelligence is a dead end. Social media erupts. The debate immediately crystallizes into a binary opposition: LLMs versus World Models, ChatGPT versus JEPA, today\u2019s AI versus tomorrow\u2019s AI. I believe this opposition is a misreading of the situation\u2014and that the real question lies elsewhere.<\/p>\n\n\n\n

In this article, when I refer to World Models, I am referring to the approach proposed by LeCun: JEPA and its variants\u2014I-JEPA, V-JEPA, and LeWM\u2014which are the most advanced implementations to date.<\/p>\n\n\n\n

Let’s take a step back. The history of AI isn’t a straight line\u2014it’s a series of conceptual breakthroughs, each of which would have been unthinkable given the one before it. I’m simplifying, but that’s the idea:<\/p>\n\n\n\n

Perceptron \u2192 Backpropagation \u2192 CNN \u2192 Reinforcement Learning \u2192 Transformers \u2192 LLMs<\/strong><\/p>\n\n\n\n

Each new stage did not destroy the previous one. It incorporated it, moved beyond it, and opened up a realm of possibilities that no one had truly anticipated. And each time, the scientific community initially resisted, before eventually coming around. This is the natural rhythm of deep science, which tends to be slow to change.<\/p>\n\n\n\n

What characterizes the current race around LLMs is precisely that it lacks that disruptive quality. More parameters, more computation, more data, upscaling\u2014it\u2019s a process of maximization with a few optimizations here and there, but no real breakthrough. A linear race in a field that has always advanced in leaps and bounds. This doesn\u2019t mean that LLMs are a dead end: they have represented a real, massive breakthrough with significant applications. But they\u2019re hitting a ceiling.<\/p>\n\n\n\n

It is in this context that I read Yann LeCun\u2019s work\u2014not as the revealed truth of a Turing Award winner, but as the patient work of a researcher who refuses to be swept up by commercial pressures. Fifteen years spent on ideas that have struggled to prove themselves and gain acceptance. Publications released on the day he left Meta, serving as a scientific legacy rather than a publicity stunt. In an ecosystem where many confuse computational power with depth of understanding, this profile deserves our attention.<\/p>\n\n\n\n

But LeCun isn\u2019t an oracle, and JEPA isn\u2019t necessarily the next big thing\u2014it may just be one of the most promising candidates. What interests me here is something else: understanding why LLMs and World Models have more in common than we realize\u2014starting with Transformers themselves\u2014and highlighting what the real divide reveals: not which architecture will prevail, but rather whether AI research still has the means and the courage to think disruptively rather than exhausting itself in a linear, short-sighted race.<\/p>\n\n\n\n