辛顿敷衍走场，是对科学的败坏

Core Viewpoint - The article discusses the contrasting perspectives on artificial intelligence (AI), highlighting the divide between optimistic and pessimistic views regarding AI's capabilities and understanding [1][2][4]. Group 1: Perspectives on AI - There is a debate between AI optimists, like Geoffrey Hinton, and pessimists who argue that AI lacks true understanding and relies on statistical methods [1][2]. - The notion that "intelligence is based on reasoning" is criticized as overly simplistic and reflective of Western rationalism, which may overlook the complexities of human understanding [2][7]. - The article emphasizes the need for a scientific approach to AI, advocating for a realistic assessment of its capabilities rather than subjective interpretations [2][4]. Group 2: Scientific Foundations of AI - AI currently lacks a foundational scientific theory, functioning more as a craft based on empirical methods rather than established scientific principles [8][9]. - The historical context of scientific breakthroughs is discussed, noting that modern science has faced stagnation in foundational theories since the mid-20th century [8]. - The article argues that the recognition of AI researchers with Nobel Prizes does not signify a theoretical breakthrough in AI but rather highlights a broader stagnation in scientific understanding [8]. Group 3: Technical Principles and Applications - AI is described as a statistical method that has gained prominence due to its practical applications, rather than theoretical advancements [9][12]. - The relationship between AI and information technology is outlined, indicating that AI is a subset of broader technological applications aimed at enhancing human capabilities [12][14]. - The article posits that while AI can surpass human performance in specific tasks, it does not equate to achieving human-like consciousness or understanding [14][16]. Group 4: Historical Context and Future Implications - The evolution of technology from craft-based methods to modern scientific approaches is discussed, emphasizing the limitations of purely empirical methods [15]. - The article warns against the potential for misinformation and exaggerated claims about AI's capabilities, suggesting that such narratives may distract from genuine scientific inquiry [16][18]. - It concludes with a cautionary note about entering a "post-scientific" era, where the integrity of scientific discourse may be compromised by unsubstantiated claims [18].