If you think ChatGPT could give a reasonably good answer to your Perl question, just go straight there and ask it. There's no reason to copy that content here.

I have an idea! What if PerlMonks would build ChatGPT into the website so when someone posts something that has a question mark in it, the entire post would be fed into ChatGPT and its response would be posted as the first reply in the thread. This would ensure that people get an immediate response.

So the problem you're trying to solve is that some posts don't get an immediate response? I don't think that's as big a problem as you think it is. Perlmonks is not a support website -- if you need a fast answer, it's obviously the wrong place to go.

Also, Perlmonks doesn't need to adjust itself for the 21st century; it's quite comfortable hanging out at the end of the last century, with a simple interface, but extremely valuable content.

Large language models (LLMs) are a potential substitute for human-generated data and knowledge resources. This substitution, however, can present a significant problem for the training data needed to develop future models if it leads to a reduction of human-generated content.
[…]
In addition, we find significant declines in posting by users of all experience levels, from novice to expert. These results suggest that ChatGPT is displacing various Stack Overflow posts, including high-quality content.

Idiocracy ante portas 😉

And just to make things clear, I think, the answer provided by ChatGPT should be copied verbatim and posted as text on PerlMonks, not just a link or a suggestion or whatever, because links cannot be trusted.

Thanks for that - I really enjoyed having a good laugh at the thought that something as fundamental to the web as "links" cannot be trusted but somehow an LLM can.

FTAOD: no. Just no.

Comedy gold, link obviously untrustworthy.

In that case, could I have a version of PerlMonks of the past century such that it would remain unmolested by any "AI" input of any kind?

If not, I really do want to have all of my posts be removed permanently. Also I would want they not be fed into any AI scheme.

As all the posts are public anyway, my posts might have already been ingested by such a scheme ... hmm ... well, should certainly cease to post anything new as log as SNAFU.

We had a similarly hypothetical discussion only half year ago.

See SO and AI ff

Cheers Rolf
_{(addicted to the Perl Programming Language :)
see Wikisyntax for the Monastery}

It's like cloning our worst troll 1000 times, amputating half their brains and giving them favorite access.

I'm already at the borderline to insanity when search engines send me to legit looking pages which contain only AI BS. (And this I'm afraid will get worse, with AI trained on shitty material generated by other AIs.)

ChatGPT is by far not good enough yet. (A better search engine trained by us might possibly be at some point, but who's going to do that?)

Update

Guys, please don't downvote the OP just because you disagree.

For completeness: I just tried a demo here -> Re^7: AI in the workplace (... in the Monastery) and the results were not convincing.

Much work and experimentation would be needed to automatically create appropriate prompts from the OP's question.°

If that's even possible...

Cheers Rolf
_{(addicted to the Perl Programming Language :)
see Wikisyntax for the Monastery}

°) like "include automatically testable code for your answers/hallucinations". ;-)

From my own experience, that's a bad idea. My boss is currently pushing me to use LLMs for efficiency. So far, all the answers have been suboptimal and either plain wrong or full of bugs and unimplemented error checks. And at least 30% of the "working" code samples would have opened security holes in my software.

If you want to bring PM into the 21st century, i suggest we skip the current AI hype wave (just as we skipped crypto-currency and and NFTs¹).

¹ Uhm, well, technically, we do have both a cryptocurrency simulation and a NFT simulation based on PM XP: XPD - Do more with your PerlMonks XP, because i wanted to play with PM APIs...