As I worked on a project with ChatGPT-4o, more and more errors and bizarre behavior appeared, including deliberately lying to me (and ChatGPT even admitted it is allowed to lie ). Suddenly they upgraded to ChatGPT-5, and things got even more bizarre. You can see the screenshot and decide for yourself. I am not a firm advocate of intense regulation and scrutiny of AI development by the government. At the very least, they should deliver and maintain a reliable and sensible product. Imagine turning over your home to a Smart Speaker and when you ask it to do something, it does something else completely unexpected.
Anonymous restaurant reviews, cooking experiments, miscellaneous musings on living in the Lower Mainland, BC, Canada.