Apple recently released a paper titled "The Illusion of Thinking: Insights into the Strengths and Constraints of Reasoning Models through the Prism of Problem Complexity." This paper rigorously examines the reasoning capabilities of large reasoning models (LRMs), including DeepSeek R1, OpenAI GPT-3, and Anthropic Claude, through a series of meticulously crafted experiments. Apple's research highlights that while these models excel in tasks of moderate complexity, their accuracy plummets and even collapses entirely as problem complexity escalates to a critical threshold. Furthermore, the paper reveals that even when equipped with explicit problem-solving algorithms, these models' performance fails to improve, implying fundamental limitations in the current LRMs' reasoning abilities. This revelation challenges the industry's optimistic outlook on LRMs' reasoning capabilities and outlines avenues for future research.