The first time I've come across this error message using #ChatGPT #GPT4, I guess it was just a matter of time, as it was a bit of chat deepdive today on many topics.
On the other hand, I've also just noticed that as the usage limits were constantly reducing over the last few months, now they are actually undefined, completely gone from the UI with the most recent model version being released...
And no, no real point going back to the "default model", we are way past that :)
I don't recommend relying on ChatGPT Plus for work: when the #GPT-4 message limit strikes, you'll be divorced from your toolchest for the next several hours. What I use instead is the #OpenAI playground (write-up for a colleague here: https://ndurner.github.io/chatgpt-vs-openai-playground), with this frontend tool that makes it a little more convenient to use: https://ndurner.github.io/geppettos-workbench. The latter also establishes are more suitable-for-work System Prompt:
---
You are a helpful assistant. Answer faithfully and factually correct. Respond with "I don't know" if you are unsure. Answer user inquiry in the context of the conversation that preceded it.
---
At least that's the best I came up with this far. I would love to hear about possible improvements that make it more like ChatGPT, but without the rampant confabulation - aside from the leaked "Sydney" prompt.
Fish in the Percolator
Nils Durner