Skip to content

Worlds Apart: LeCun's Vision vs. OpenAI's Path to AI's Future

Shelly Palmer has been named LinkedIn’s “Top Voice in Technology,” and writes a popular daily business blog.
shellypalmerwednesday

Greetings from Santa Monica – where it has finally stopped raining.

Yann LeCun, Meta's Chief AI Scientist, has been all over social media lately. He's looking to reboot the AI research echo chamber, challenging the current approach to AI championed by OpenAI, Google, and other big players. His idea is to build holistic, world-modeling systems that mimic human-like understanding and reasoning, which is a stark contrast to the brute-force data processing of today's large language and image models like GPT-4 and DALL-E as well as the trial-and-error of reinforcement learning.

Why should you care? There's so much hype around LLMs and ChatGPT (and other gigantic-scale, energy-hungry, NVIDIA-based, big-tech-owned models) that it's important to consider that that future is not inevitable.

LeCun's vision may not be correct, but his critique isn't just talk – it's backed by an ambitious vision for AI that focuses on abstract prediction and reasoning rather than rote memorization or simplistic reward chasing. He's championing a shift toward AI systems that can understand the nuances of the world without needing to predict every pixel or memorize text patterns. His approach emphasizes the development of "world models'' that can infer the gist of future events, akin to the way humans use common sense to navigate daily life.

This philosophy diverges sharply from the prevailing winds in AI research, which have largely been propelled by the success of generative models and reinforcement learning. According to LeCun, these methods (while technologically impressive) fall short of bridging the gap to true artificial intelligence because they lack an innate understanding of the world.

Which approach is right? It's too early to tell. They may both be wrong. What's important is that the path to better AI or even AGI (artificial general intelligence, akin to human intelligence) is not known or agreed upon… so adjust your investment theses accordingly.

As always your thoughts and comments are both welcome and encouraged. Just reply to this email. -s

sp@shellypalmer.com

ABOUT SHELLY PALMER

Shelly Palmer is the Professor of Advanced Media in Residence at Syracuse University’s S.I. Newhouse School of Public Communications and CEO of The Palmer Group, a consulting practice that helps Fortune 500 companies with technology, media and marketing. Named LinkedIn’s “Top Voice in Technology,” he covers tech and business for Good Day New York, is a regular commentator on CNN and writes a popular daily business blog. He's a bestselling author, and the creator of the popular, free online course, Generative AI for Execs. Follow @shellypalmer or visit shellypalmer.com