Reported by @wise ferry
Bug Report: GPT-4o made worse again?
`Steps to Reproduce`
- use GPT-4o and send a prompt
- answer seems different, sloppier, cheaper and worse, it seems that ChatGPT does not remember the context of the chat
`Expected Result`
- same quality of GPT-4o and other models as before
- ChatGPT remembers the entire context of a chat as usual
`Actual Result`
- a different, sloppier, cheaper and worse answer
- context of chat seems forgotten
`Environment`
ChatGPT Android app, all browsers