Did Meta’s AI chief LeCun get removed for criticising LLMs and AI? Social media fuels speculations of an ‘AI cult’ |

Yann LeCun’s upcoming departure from Meta has ignited a storm of online discussions, with social media users questioning whether his outspoken criticism of large language models (LLMs) contributed to his marginalisation within the company. The news has intensified debates about whether Big Tech’s AI leadership tolerates dissent or whether an online-coined “AI cartel” is quietly shaping the industry’s direction — pushing hype, suppressing alternative ideas and positioning LLMs as the only viable path forward because it benefits the companies that dominate the market.

A growing rift between LeCun and the industry

LeCun, one of the pioneers of modern machine learning and a Turing Award laureate, has long argued that LLMs are not a credible path toward human-level intelligence. He has repeatedly described them as limited “autocomplete machines” that lack foundational cognitive pillars such as reasoning, planning and causal understanding. This stance increasingly put him at odds with the dominant industry narrative that positions LLMs as the foundation for future AGI.He reportedly plans to leave Meta to build a startup focused on open-ended learning and world-modelling approaches. Insiders note that Meta’s strategic shift toward rapid LLM commercialisation has already reduced the influence of fundamental research groups, including those aligned with LeCun’s vision.Many observers argue that this shift is driven not only by scientific ambition but also by enormous commercial incentives. LLMs generate attention, attention attracts investment, and investment strengthens the commercial position of the biggest AI players — creating strong motivation to maintain the perception that scaling these models is the only path worth pursuing.

Social media reactions spark claims of an “AI cult”

A widely shared social media thread intensified the controversy. The post characterised the AI community as a “cult” that aggressively protects the LLM-centric agenda and sidelines anyone who questions it. According to this view, LeCun’s criticisms made him an outsider in an environment that increasingly rewards alignment over dissent.Supporters of this perspective argue that the concentration of AI resources in a handful of companies creates cartel-like dynamics. These firms shape public narratives, control funding pipelines and set research agendas in ways that discourage alternative thinking. Many users noted that this hype-driven model pressures the public, investors and even policymakers into believing that LLMs are the inevitable future.These claims remain speculative and reflect social-media discourse rather than confirmed evidence. However, they do capture growing public anxieties about transparency, power concentration and the ability of major AI companies to steer both imagination and investment in directions that serve their commercial interests.

The hype machine and the risks of “replacing humans”

Another theme emerging across social media is the belief that major AI companies intentionally amplify the narrative that AI will replace humans in order to maximise perceived value and market dominance. Critics argue that such messaging attracts massive investment flows from venture capital, governments and corporate clients eager not to be “left behind.”While advancement in AI is necessary, critics warn that portraying AI as a replacement for humans — rather than a tool to assist them — is neither realistic nor welcome. They argue that the race to make everything effortless or automated may create long-term consequences: degradation of human skills, job displacement, economic instability and over-reliance on systems whose limitations are still poorly understood.These concerns intertwine with broader suspicions that what some online users call an “AI cartel” thrives not on scientific consensus but on market psychology and narrative control.

Evidence supporting LeCun’s scientific concerns

LeCun’s scepticism is grounded in more than opinion. Multiple peer-reviewed studies have documented significant limitations in current LLMs, particularly in tasks requiring causal reasoning, long-term planning and grounded understanding. For example, NeurIPS-era evaluations such as CLadder and related papers show that LLMs often struggle on structured causal-inference benchmarks and rely on superficial statistical shortcuts in certain tasks.The exact performance varies by benchmark and prompting method, but the recurring finding across studies is clear: current LLMs lack reliable causal reasoning — a capability LeCun considers essential for progress toward human-level intelligence.For LeCun, such results underscore the need for hybrid systems that integrate perception, planning, world models and grounded learning, rather than relying exclusively on scale. His upcoming startup is expected to pursue these alternative research directions.

Meta’s silence leaves room for speculation

Meta has not addressed claims that LeCun faced internal pushback, nor has it confirmed whether his departure was voluntary, strategic or linked to disagreements about the company’s increasing focus on LLM products. The absence of clarity has allowed speculation to spread across X, Reddit and other social platforms.Some analysts caution against assuming conflict, noting that senior researchers often leave major firms to pursue independent ventures. Others argue that the timing suggests deeper philosophical disagreements about the future of AI and the growing dominance of corporate-driven LLM development.

A debate that reflects an industry at a crossroads

LeCun’s exit comes at a moment when the AI field is sharply divided. One side believes LLMs are on the path to general intelligence and should be scaled aggressively. The other warns that scaling alone will reach its limits and that breakthroughs will require fundamentally new architectures.Whether LeCun’s departure was shaped by corporate politics, strategic misalignment or personal ambition, the broader debate it has triggered speaks to a deeper tension running through the AI world: the conflict between scientific diversity and commercial monoculture.As long as a small number of companies control the resources, investment flows and public messaging, questions about openness, dissent and the possibility of what some online users call an “AI cartel” or “AI cult” will remain central to discussions about the future of artificial intelligence.