Bug Report: ChatGPT poor integration of context with search results
`Steps to Reproduce`
Edited: Corrected invalid text about automatic searches.
For the last several weeks the GPT-4o model has seemed really bad. Ref these notes and the notes linked from there. #chatgpt-discussions message
Open ChatGPT and ask a detailed question. When search is active, for the session or via perferences, the results are really bad. See links for notes about issues in response patterns.
`Expected Result`
Responses should be coherent, succinct, accurate, in-context.
`Actual Result`
Responses contradict others. Some are inaccurate right after accurate responses. Prompt details are ignored. Responses duplicate search results (viewed from linked Sources), sometimes reduplicating these results in subsequent responses without noticing that the text was already in a prior responses.
`Environment`
All ChatGPT platforms including Windows 10, FireFox 135
serene anvil
