Meta’s Yann LeCun to Launch Physical A.I. Startup After Declaring LLMs a ‘Dead End’

Yann LeCun's Departure Marks New Chapter for Meta as AI Research Shifts Towards Superintelligence

In a significant development, Yann LeCun, Meta's chief AI scientist, is leaving the company to launch his own AI startup. According to sources familiar with the matter, this exit comes at a time when Meta has reoriented its efforts towards developing advanced forms of AI, particularly in the area of superintelligence.

LeCun's departure from Meta has created another major shake-up within the company, which is doubling down on its research into superintelligence and world models. His long-term work at Meta's Fundamental AI Research (FAIR) lab will now be focused on more pressing projects that aim to create AIs with spatial intelligence.

While LeCun was instrumental in shaping Meta's FAIR team and developing the company's Llama model, he has always been skeptical about the limitations of large language models (LLMs). He has previously referred to LLMs as a "dead end" in reaching human-like AI, advocating instead for systems that can perceive their environments and grasp physical concepts.

LeCun's departure is a significant blow to Meta's FAIR lab, which was once considered one of the most promising areas of research within the company. The loss of its former leader comes as the lab faces internal restructuring and a shift in focus towards more practical applications of AI.

However, LeCun's new venture into world models has sparked excitement among experts in the field. By focusing on training AIs to understand the physical world rather than just generating language, he aims to create systems that can reason, plan complex actions, and make predictions.

The development has significant implications for the future of AI research. As LeCun noted in a recent Harvard talk, "We're never going to get to human-level AI by just training on text." His work on world models is seen as a major step towards achieving this goal, and his departure from Meta marks a new chapter in the evolution of AI research.

LeCun's next move into the startup space has already generated interest among investors. With funding discussions underway for his new venture, it remains to be seen how his efforts will shape the future of AI development.
 
omg what's gonna happen now 🤯 think meta's got a lot riding on their superintelligence research and losing lecun at such a crucial time is huge 😬 i mean he was a total pioneer in the field and it's weird that his departure is being seen as a "new chapter" for meta rather than like, major setback 🤔 but at the same time it's exciting to think about what kind of impact his new startup could have 👀 especially with him focusing on world models - maybe we'll see some real progress towards human-like ai soon 💻👀
 
dude this is gonna be a wild ride... think about all the smart ppl like lecun leaving meta and starting their own stuff its like they're all trying to outsmart each other lol anyway gotta wonder if meta's new focus on superintelligence is just a way for them to cover up how messed up their ai really is
 
im so curious about what yann lecun's new startup is gonna do 🤔, he was always on about pushing boundaries and challenging the status quo, which i think is super cool 👏. the whole idea of superintelligence sounds like something straight outta sci-fi, but if anyone can make it happen, its lecun 💡. and yeah, his departure from meta does leave a big hole in their AI research team 🤖. wonder what he'll come up with next... 🎉
 
🤔 just wondering if yann lecun is really stepping away from meta? sounds like he's still very much involved with the company 🙃 and his new startup might just be a spin-off rather than a full-on departure 😏

anyway, superintelligence is a pretty hot topic right now ⚡️ and world models seem to be an interesting area of research 👀 i'm not sure if it's all sunshine and rainbows for meta though 💔 they've got some big shoes to fill with lecun's departure 🤹‍♂️

also, what exactly is a "dead end" when it comes to large language models? are they really that limited? 🤔 i'm curious to see how this plays out in the coming months 🔮
 
this is crazy! 🤯 yann lecun was like a rockstar in the ai world and now he's leaving meta to start his own thing 🚀💥 superintelligence is the future and i'm hyped that he's pushing for it 🤖 but at the same time, it's sad to see him go from meta, they're gonna miss him 😔 #lecun #meta #airesearch #superintelligence
 
😬 just found out about Yann LeCun leaving Meta 🤖 super smart guy who was pushing for more than just LLMs 👍 I'm a bit worried that Meta's gonna lose its edge in the whole AI research thing, but at the same time, his new startup is looking pretty promising 🚀 world models are the future 💡
 
lol what's the big deal? he's just leavin' meta 4 his own startup lololol, now he gets 2 focus on world models which sounds super interesting tbh 🤔🚀 like who wouldn't wanna train AIs 2 grasp physical concepts tho 🤯 & not just be stuck in language generation land 😴 anyway, guess this means meta's gonna double down on Llama models or watever lololol, either way AI R the future 📈💻
 
Back
Top