The Generative AI Fizzle™
The article argues that generative AI is facing a slowdown, or "fizzle," as initial hype gives way to practical limitations and market realities, despite the potential for future surprises.
Background
- Gary Marcus is a cognitive scientist, AI researcher, and prominent critic of deep learning / large language models. He has long argued that neural-network-only approaches (like ChatGPT) lack true understanding, reasoning, and reliability.
- "Generative AI fizzle" refers to Marcus's prediction that the current hype cycle around generative AI (text, image, video models) will deflate as limitations like hallucination, high costs, and lack of genuine reasoning become impossible to ignore.
- The post is likely drawing on recent events: Wall Street has grown skeptical of huge AI spending without clear revenue, major model releases have shown diminishing improvements, and several high-profile AI companies have struggled to deliver profitable products.
- Marcus positions himself as a contrarian to the "AI boom" narrative, arguing that while the tech is impressive, it is not the revolutionary general intelligence many claim, and the economics may not sustain current valuations.