ChatGPT finally knows how many ‘R’s are in ‘strawberry,’ but confident mistakes remain

OpenAI says ChatGPT now counts the R’s in “strawberry” correctly and handles a simple car-wash logic prompt.

But users quickly pointed out the model still makes other obvious mistakes, raising doubts about how broad the fix really is.

The update may reflect a narrow or hardcoded improvement rather than a major step toward reliability.

The episode underscores a bigger trust problem for AI chatbots: they can still sound confident while being wrong.