Part 2/12:
The initial launch of GPT-5 was far from smooth. When the model was rolled out, users found broken presentation graphics, with charts where 30% and 70% were indistinguishably identical—an embarrassing failure for a company that prides itself on scientific rigor. Additionally, OpenAI removed popular legacy models without prior warning, leaving many users in the dark and causing confusion about what models they were interacting with. Most notably, users woke up to find trusted tools replaced with the mysterious GPT-5, which felt different and unfamiliar, raising questions about whether the model was truly ready for prime time.