LLM use policy and stance
To be honest with y'all, current LLMs are just trash. And not only that, but the results horrible, at best and sometimes not even on the same topic as the question. LLMs in my opinion are just some fancy toys that Big Tech plays with and while Machine Learning indeed has a bright future in other domains, like field and pattern recognition, maybe transcription, currently, it's not good enough to be implemented as an answer engine.
Disclosure of LLM usage
No LLM generated content has been used in the making of this website, nor its initial content, and nor in its future content. Everything is human-made and external resources are curated to not contain any generated content/are human made.[1]
For full disclosure, I have attempted the use of LLM generated content not in the making of this website, but rather in the search for resources for learning how to make this website [2].
This only attempt was a failed one, therefore no generated content has been used in the making of this website or the learning process
My stance on LLMs is strong and is sustained by valid concerns.
Footnotes
[1] eventual inclusion of external resources to which LLMs may have contributed will be fully disclosed (eg: github repos, texts where proofreading was made with LLMs)
[2] For further disclosure, the resources searched were for a method of replacing the "marquee" tag, now deprecated. The final resource used was "https://stackoverflow.com/questions/31951282/why-is-marquee-deprecated-and-what-are-the-alternatives" and the search page from which I extracted the resource was "https://kagi.com/search?q=replace+marquee+html&r=ro&sh=INwrmcd0mfrRpGv24Z5KLA".
Furthermore, the LLM provider platform used was "https://duck.ai" and the model used was GPT-5 mini. The prompt used was "How can I replace the now deprecated marquee tag" with search mode turned on. The results were lackluster, therefore having not contributed in any way.
The current implementation of the StackOverflow resource used on this website is a badly adapted implementation and most probably unscalable. But it works well enough for my case. The adapted implementation has been created by yours only, cloud (me).