Page 4 of 4
Re: Using AI models ? (deepseek and ...)
Posted: Tue Feb 18, 2025 1:49 am
by BarryG
Do you guys have an account with ChatGPT? Maybe that makes a difference. I use the free account that it offers.
Re: Using AI models ? (deepseek and ...)
Posted: Tue Feb 18, 2025 1:55 am
by Quin
BarryG wrote: Tue Feb 18, 2025 1:49 am
Do you guys have an account with ChatGPT? Maybe that makes a difference. I use the free account that it offers.
I do, yeah. The results its given me on PB code has been mixed, but it at least seems to have learned that I love the language, when I ask it questions about programming languages PB very often gets brought up now.
Re: Using AI models ? (deepseek and ...)
Posted: Tue Feb 18, 2025 11:17 am
by Skipper
BarryG wrote: Tue Feb 18, 2025 1:49 am
Do you guys have an account with ChatGPT? Maybe that makes a difference. I use the free account that it offers.
So do I, but my results are mixed and vary wildly.
Sometimes, changing the LLM version does improve consistency somewhat, but never for very long.... This is my experience on the
implementation side when using PureBasic as the output language; when talking about possible algorithms for a task, my results are better, but this clearly is a more generic topic.
Re: Using AI models ? (deepseek and ...)
Posted: Sun Aug 17, 2025 11:16 am
by BarryG
STARGÅTE wrote: Wed Feb 12, 2025 8:12 amwhat else can we (as the community of Pure Basic) do, to correct such wrong answers, for ALL?
I notice ChatGPT now allows your corrections and content to be used by all (it's in the settings for your account, called "Model improvement"):
