Frustration is mounting among ChatGPT users over ongoing memory issues following the August update. Many have reported that the tool fails to recall information correctly, raising doubts about its reliability and features.
Recently, users have taken to various forums to voice their grievances. One user lamented, "My GPT is also periodically struggling with recalling things saved to the bio tool as well as knowledge files."
The sentiment reflects a widespread concern about memory inconsistencies. Users who expect proper functionality find themselves confused when the AI seems to forget previously shared information, especially in new conversations.
Several recurring themes have emerged from user feedback:
Persistent Memory Failures: Users note that even with memory settings enabled, the tool often fails to recall past interactions. "It's mostly been fine, but sometimes Iβm getting these hiccups too," a user mentioned.
Hallucinations and Confusion: Many have reported experiencing hallucinations where the AI presents incorrect or nonsensical information. As one user pointed out, "Chat does not understand its own RCH feature, and that answer is a hallucination."
Unwanted Features: Users expressed annoyance at new functionalities like "Ah, this is" which added to their frustrations rather than resolving existing issues.
Commenters urge OpenAI to clarify the differences between session and persistent memory. Some users remarked,
"It violates my memory rules frequently, adding unwanted information from previous chats."
This inconsistency has added a layer of confusion, prompting many users to reconsider their reliance on the AI tool.
The overall tone of user feedback skews negative. Many are grappling with trust issues regarding ChatGPT's memory capabilities.
π Frequent memory lapses are reported by many users.
π Disappointment is evident, with some reverting to traditional writing methods.
β οΈ Repeated concerns about the reliability of memory features post-update are common.
As OpenAI continues to modify its memory functionalities, users seem to demand increased transparency and regular updates. If these glitches persist without resolution, it could lead to diminished user engagement, forcing users back to conventional methods of documentation. Curiously, could these challenges signal a turning point for the platform?
β³ The memory feature inconsistent performance raises user trust issues.
β½ Feedback suggests a need for OpenAI to clarify memory settings and functions.
β» "I had to go back to my old notebook and pen," illustrates the desperation among disappointed users.