Criticizing Current AI Capabilities

Gary Marcus is a professor emeritus of psychology and neural science at NYU, bestselling author, and founder of two AI companies. He has spent years reminding people that we may be overestimating current AI capabilities.

Vision and Contribution: His core criticism targets the fundamental flaws of large language models: “What we actually have is everybody running around with LLMs that are fundamentally unreliable, unpredictable, and can't be aligned with human values.” 🤖

Alternative Vision: Marcus believes in hybrid architectures combining symbolic reasoning with neural networks—not just bigger Transformers. He envisions an AI like the computer in Star Trek: “You ask it a question, and it gives you an answer you can trust.”

Quote: “It's time to rethink the hype narrative.” He misses the AI field of five years ago when people were driven by intellectual curiosity rather than money.