Attempting to crack the New York Times Spelling Bee puzzle, the author turned to AI chatbots for assistance. Despite providing specific instructions and letter combinations, the chatbots struggled to generate valid words within the constraints of the game. AI models, based on statistical patterns rather than logic, rely on training data for accuracy. The lack of specific training in word puzzles highlighted the limitations of generative AI models. As AI companies market their technology as a solution for complex problems, experts caution against unrealistic expectations. While some speculate on the potential of Artificial General Intelligence, others like Yann LeCun emphasize the current limitations of AI in understanding complex tasks. Transparency about AI capabilities is necessary as tech companies integrate AI into various products. Despite advancements, the article concludes that AI chatbots may not be suited for every task they are assigned, highlighting the need for realistic expectations and clearer communication about AI capabilities.