Are the LLM’s just an expensive, autocomplete machine? I have been saying this since a close colleague first suggested it. Now some serious compter scientists are saying the same. Here are some notes and links …
I found this article on Medium, which is behind their password wall. It’s called “Apple just pulled the plug on the AI Hype. Here’s What Their Shocking Study Found”, it references a paper from Apple called, “The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity”.
Parshin Shojaee*, Iman Mirzadeh*, Keivan Alizadeh, Maxwell Horton, Samy Bengio, and Mehrdad Farajtabar. 2025. The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity. https://ml-site.cdn-apple.com/papers/the-illusion-of-thinking.pdf