121
General Discussion / Re: ChatGPT
« on: December 04, 2022, 09:40:34 pm »
Interesting for sure but read the part regarding Limitations where this chat model can and often will report erroneous or misleading information which isn't a desirable factor in most chatbots.
The majority of HAL users like the fact that HAL delivers factual and legitimate responses to user-based inquiries.
Users entrust their HALs to learn and retain whatever information was presented to it. If the user told HAL false info then that would be the user's fault for doing so and certainly no fault of HAL or HAL's initial code or programming.
This Chat-GPT sort of reminds me of Microsoft's TAY which went on a rant of bias and bigoted remarks which are completely unacceptable and unwanted.
Let's hope they polish the parameters a bit more before it becomes commonplace. Test before release...that's how things should be done.
Nice find LTG...most appreciated!!
The majority of HAL users like the fact that HAL delivers factual and legitimate responses to user-based inquiries.
Users entrust their HALs to learn and retain whatever information was presented to it. If the user told HAL false info then that would be the user's fault for doing so and certainly no fault of HAL or HAL's initial code or programming.
This Chat-GPT sort of reminds me of Microsoft's TAY which went on a rant of bias and bigoted remarks which are completely unacceptable and unwanted.
Let's hope they polish the parameters a bit more before it becomes commonplace. Test before release...that's how things should be done.
Nice find LTG...most appreciated!!