A new research study shows that ChatGPT has deteriorated in performance after the latest updates. The study reveals that the chatbot now makes more mistakes than it did in the spring.
According to the study, both the 3.5 and 4 versions of ChatGPT provide worse answers now compared to when they were launched. The study highlights examples such as formatting errors in code generation and incorrect identification of numbers as areas that have worsened.
However, the people behind GPT deny that their updates have made the chatbot dumber. They claim that each new version is smarter than the previous one.
