ChatGPT just announced it can pass the ‘how many “r”s in strawberry’ test, but users found otherwise (www.techradar.com)

ChatGPT has made headlines by announcing its ability to accurately count the number of "r"s in the word "strawberry," a task that has historically stumped AI models. However, users promptly discovered that when the challenge was shifted to "cranberry," the model faltered, highlighting ongoing difficulties in basic letter-counting tasks. This incident underscores a broader issue within AI language models, where they excel in complex problem-solving but still struggle with seemingly straightforward counting tasks due to their linguistic processing limitations. The significance of this event lies in the persistent gaps in AI reasoning abilities, even as models are marketed as improving. While the announcement suggests enhancements, the failure to generalize the counting logic to similar words raises questions about the depth of learning and adaptability in these systems. Furthermore, the so-called "car wash" problem illustrates additional hurdles in logical reasoning, where AI often overlooks context-critical details. The ongoing challenges faced by models like ChatGPT and Claude suggest that advancements may be tailored towards passing specific tests rather than achieving true comprehension, prompting the AI community to reconsider how progress is measured in artificial intelligence.