I want to express my gratitude to the OpenAI team for their outstanding work. ChatGPT is a powerful tool that I use for analysis, work, and communication. I truly appreciate its convenience and the quality of responses.
However, I believe there are a few areas that could be improved:
Lack of information about conversation time and date в - It would be helpful to see when a chat session started and ended. This would allow for better tracking and analysis of interactions.
Time spent using ChatGPT - Understanding how much time has been actively spent in chat would be useful, especially for those using ChatGPT for work or research.
More transparency about how my data is used to improve the model - Currently, there is an option to «Improve the model for everyone» but there is little detailed information on how exactly my conversations contribute to AI training. It would be great to have more clarity, or even a personalized analytics feature.
I would be happy if these suggestions could be considered in future updates. Thank you for your hard work!
I like particularly your first suggestion.
I’ve used since beginning a custom instruction for it to print a timestamp in the end of every message, however:
it never got the time right. Date is fine, at least.
sometimes it will forget this instruction, I only notice it later when i need a timestamp
Thanks for sharing your experience, @qsq! That’s really interesting - I didn’t know ChatGPT could be inconsistent with timestamps. It seems like an integrated feature would be much more reliable. Have you found any workarounds for this issue?
GH Copilot, it shows timestamps, can’t remember if i had instructed it to do it, looked to me like AI has some central processing unit that carries user info across all services
ChatGPT has literally been a live changer for me.
It’s the first instance of the fulfillment of what I always wanted from the Internet … a wisdom machine.
That said, the new version is utterly slow in comparison.
It says:
Reasoning, clarifying, responding to user feedback, addressing user concerns … I don’t need to see a ‘log output’ of what is happening behind the scenes, I want speed over details.
It now takes 10-15 seconds to answer a simple question that required only 2-3 seconds before.
Also after receiving an answer I want follow up questions to first be applied to the discussion we just had instead of assuming it’s a new question.
For example if I ask about baseball stats and you give me the name of a player and I then ask “MVP?” I want you to assume I’m asking about the player you just told me about instead of coming back with a definition of MVP.
That’s s an interesting theory! I wonder if AI tools like GitHub Copilot and ChatGPT share certain user preferences when using the same OpenAI account. Or maybe it’s s just a coincidence based on past interactions. Have you noticed similar behavior in other AI services?
ChatGPT has literally become a living variable for me.
This is the first instance of the internet fulfilling what I’ve always wanted… A machine of wisdom.
However, the new version is extremely slow in comparison.
It states:
Reasoning, explanation, responding to user feedback, solving user problems…
I don’t need to see a “log output” of what’s happening behind the scenes—I need speed, not details.
Now, a simple question takes 10–15 seconds to answer, whereas before, it only took 2–3 seconds.
Also, after receiving an answer, I want follow-up questions to apply to the discussion we just had, rather than assuming it’s a brand-new question.
For example, if I ask about baseball statistics and you give me a player’s name, and then I ask, “MVP?” I want you to assume I’m asking about the player you just mentioned, rather than returning with a definition of MVP.
Thanks for the clarification! So it’s s only happening on GH Copilot. Do you think Copilot automatically detects when timestamps are useful, or could it be related to a specific setting?
I recall that on my very first interaction after signing up for Pro plan, it gave me the timestamps. I hadn’t entered any custom instructions, so perhaps they have some customer unique ID.
It looks like OpenAI assigns a unique client ID (client_id) to each user or service. This ID is recorded in the logs and indicates which application or platform is being used (e.g., ChatGPT iOS, OpenAI Platform, etc.). If GitHub Copilot Pro has a specific client ID, it might enable certain features like timestamps automatically. So, it’s s possible that timestamps appeared not because of a manual setting, but because your client ID triggered them.
