咖啡机变聪明后，我连咖啡都喝不上了

Core Insights - The article highlights the disparity between expectations of AI capabilities and their actual performance, particularly in executing simple tasks like making coffee or controlling lights [1][5][11]. Group 1: AI Performance Issues - Users have expressed frustration with AI assistants like Alexa, which fail to execute basic commands reliably after upgrades, leading to a perception of decreased functionality [1][2][5]. - Traditional voice assistants operated on a template-matching basis, ensuring predictable outcomes, while newer AI models introduce randomness, resulting in inconsistent responses [7][8]. Group 2: Technical Challenges - The inherent randomness of large language models (LLMs) complicates their ability to perform tasks that require precision and repeatability, such as controlling smart home devices [7][9]. - Despite the potential for LLMs to understand complex commands better, they struggle with generating consistent system calls necessary for reliable device control [8][10]. Group 3: User Experience and Expectations - Users acknowledge that while the new AI systems can handle complex commands more effectively, they still face issues with basic functionalities [14][20]. - There is a growing consensus among users that the challenge lies not in the introduction of AI but in defining its boundaries and ensuring it complements existing reliable systems rather than replacing them [21][22].