AI: Yann LeCun, Meta's long-time AI Chief Scientist leaves. RTZ #903

The core debate underlying LLM AIs (large language models) to date, is whether they can by thsmselves get us to AGI (artificial general intelligence). Or are they just stochastic parrots that don’t yet understand enough of the broader, physical world to get the job done. That debate is at the crux of the current unprecedented global investment bet on AI. Particularly the implicitly discounted time frames. More in a few years vs a few decades.

And one of the key moments this year in these early AI Tech Wave days has of course been the dramatic shift by Meta in its broader AI ambitions. Founder/CEO Mark Zuckerberg has made some breathtaking shifts at Meta in terms of ramping up its AI data center, infrastructure, power AND talent investments to over $600 billion in the coming years. And that shift on the AI talent side has meaningfully impacted the other big tech/AI companies in their trajectories and plans as well.

One of the long-awaited impact was on how Meta chief AI scientist to date, Yann LeCun would react to all these changes. I’ve written about Yann at length in these pages, given that he’s one of the trio of ‘AI Godfathers’ that got this LLM AI thing going over the last decade and more.

Yann in particular has been more vocal of late that LLM AIs alone were NOT going to get us to the AGI, (aka Superintelligence) aspirations of most of the current AI leaders. That we would need additional AI research breakthroughs in AI ‘World Models’ that would provide far better AI reasoning, accuracy, reliability and safety features that are so needed on this breakneck AI ramp that is accelerating globally.

I wrote about this in detail recently, highlighting another leading AI scientist Andrey Karpathy, of ‘vibe coding’ fame, who also views the current LLM AI industry path as taking longer than current consensus to AGI/Superintelligence.

It looks like Yann LeCun is ready to try some new directions.

The Financial Times provides additional details in “Meta chief AI scientist Yann LeCun plans to exit and launch own start-up”:

”Meta’s chief artificial intelligence scientist Yann LeCun is planning to leave the social media giant to found his own start-up, as Mark Zuckerberg seeks to radically overhaul the company’s AI operations.”

“LeCun, a Turing Award winner who is considered one of the pioneers of modern AI, has told associates he will leave the Silicon Valley group in the coming months, according to people familiar with the conversations. The French-US scientist is also in early talks to raise funds for a new venture, one of the people said.”

This is of course not occurring in a vacuum, as I described above:

“The impending departure comes as Meta’s founder shakes up its AI strategy in order to challenge rivals such as OpenAI and Google in developing more powerful forms of AI. Zuckerberg has pivoted away from the longer-term research work of Meta’s Fundamental AI Research Lab (Fair), which LeCun has headed since 2013, to focus on more rapidly rolling out models and AI products after deciding that Meta had fallen behind the competition.”

“Over the summer, Zuckerberg hired Alexandr Wang to lead a new “superintelligence” team at Meta, paying $14.3bn to hire the 28-year-old founder of data-labelling start-up Scale AI and acquire a 49 per cent interest in his company. Within those wider AI efforts, Zuckerberg also personally handpicked an exclusive team, called TBD Lab, to propel development of the next iteration of its large language models, luring staff from rivals such as OpenAI and Google with $100mn pay packages.”

This has of course, meant some changes for Yann LeCun at Meta:

“As a result, LeCun, who had previously reported to chief product officer Chris Cox, is now reporting to Wang. Zuckerberg’s pivot followed the botched release of Meta’s most recent Llama 4 model, which performed worse than the most advanced offerings from Google, OpenAI and Anthropic, while its Meta AI chatbot has failed to gain traction with consumers. LeCun, however, has long argued that the LLMs that Zuckerberg has put at the centre of his strategy are “useful” but will never be able to reason and plan like humans, increasingly appearing at odds with his boss’s AI vision.”

And Yann has been focusing on new directions for AI Research:

’Within Fair, LeCun has instead focused on developing an entirely new generation of AI systems that he hopes will power machines with human-level intelligence, known as “world models”. These systems aim to understand the physical world by learning from videos and spatial data rather than just language, though LeCun has said it could take a decade to fully develop the architecture. LeCun’s next endeavour is focused on furthering his work on world models, according to two people familiar with the matter.”

Meta under Mark’s direction has of course been more than doubling down on LLM AIs as they’re scaling today.

”Zuckerberg has come under growing pressure from Wall Street to show that his multibillion-dollar investment in becoming an “AI leader” will pay off and boost revenue. Meta’s shares plunged 12.6 per cent — wiping out almost $240bn from its valuation — in late October after the chief executive signalled higher AI spending ahead, which could top $100bn next year.”

“LeCun’s departure marks the latest in a string of exits and leadership and organisational reshuffles at Meta in what has been a tumultuous year for the $1.6tn company. In May, vice-president of AI research Joelle Pineau left and recently joined Canadian AI start-up Cohere. Last month, the company also laid off about 600 people from its AI research unit in a bid to cut costs, eliminate bureaucracy and release products more quickly. Zuckerberg has also brought in new AI leaders on high salaries of hundreds of millions of dollars, irking some of the old guard. In July, Shengjia Zhao, co-creator of OpenAI’s ChatGPT, was hired as the chief scientist of Meta’s Superintelligence Lab.”

Yann’s impending move marks another notable AI event in these early days of a frenetic AI Tech Wave.

Much of today’s accelerating AI investments are based on bottoms up technical debates that themselves are being debated sharply by the top of the top AI Researchers in the industry.

And that is something to note and track intently going forward. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)