hg是什么意思

Question

We have this issue where we have a site named Generative AI, but a scope focused on prompt design. So...

Question: Do we want a prompt design site or a generative AI site?

Before we can answer many Is XYZ on topic? questions, we'll probably need some idea of where we want to go with the site (maybe XYZ is off-topic for the users who think we have a prompt design site, and on-topic for the users who think we have a generative AI site).

Update: I'll add my thoughts:

1. If we stick with "prompt design" (experts = genAI power users), when the site's launched many people are going to be confused as to why questions about, say, training GenAIs are closed as off-topic at a site called Generative AI. We could try to force the square peg into the round hole, but I think the results will be as tenable as Health.SE dealing with personal health questions. I certainly wouldn't want to be the moderator explaining to a genAI developer (who knows vastly more about genAI than me) why their questions at GenAI.SE are off-topic. And I'm fairly sure future users and moderators will just override us, and choose a more logical scope. The only way I see this working is if we change the site's name to PromptAI or something more indicative of "using GenAI".

2. If we instead aim for "generative AI" (experts = genAI developers, etc.), we should rewrite the scope almost from scratch, ideally with the help of genAI experts. Generative AI Stack Exchange could work, but only if generative AI (not prompt design) is the central topic; maybe prompt design should be considered borderline off-topic, like at AI.SE. I think multiple stakeholders are like me: avid users of range of genAIs, but far from experts when it comes to designing them. This may mean we are of limited help to such a site.

3. Or we could just let the ship crash into the iceberg when the site's launched the 26th (~5 days). We can battle for supremacy through upvotes, downvotes, votes to close, flags, and arguing (which we're already seeing signs of), and see who wins. But at a site called Generative AI, generative AI (no. 2 above) is inevitably going to win, so the "experts = power users" side should bow out gracefully and wish the "experts = genAI developers" side a successful site.

Answer 1

A generative AI site for users of the tools. I'd roughly see the topics as:

(*: Possibly off-topic for other reasons)

I don't think the site's scope should be developement of generative AI tools, as (for the most part) it's far outside the domain of existing questions and user expertise, and would likely have significant overlap with AI.SE and DataScience.SE.

I also don't think the site's scope or name should be limited to prompt design. Most questions currently on the site aren't about designing prompts, and questions about NVIDIA Canvas or use of AI upscalers for diagrams/charts might not even be related to the tools that accept prompts.

If we instead aim for "generative AI" (experts = genAI developers, etc.)

Many sites on the network named/about "X" are for (power) users of X, rather than the developers of X. Although, that'll be more obvious for sites like TeX.SE (which isn't solely for Donald Knuth) than it will be for GenAI.SE or WebApps.SE.

But at a site called Generative AI, generative AI (no. 2 above) is inevitably going to win

I'd personally predict the opposite. There are far more users of generative AI than there are developers of generative AI - including massive active communities dedicated to various forms of media creation using generative tools.

That said, while I don't think they'd outnumber generative AI usage questions, it does seem likely that some people will still ask questions about AI developement. It could make sense to consider changing the long-form name, like:

Android.SE == "Android Enthusiasts"
GenAI.SE == "Generative AI Users"

Answer 2

When selecting a topic, I suggest a narrow focus due to the lack of clarity and consensus regarding question-asking procedures in the private beta (rel. "How to Ask Questions in Private Beta" in GenAI SE).

Having a narrow scope will alleviate the burden that might be derived from having to answer "Is XYZ on topic? questions" as most questions might be closed as duplicates of FAQ explaining the site scope and why it's very narrow during the current stage and how the site scope changes will be announced.

There should be a consensus about the quality standards and site workings, like how votes should be used in Meta / how it will be determined what decisions were consensual and what decisions for practical matters won't, and how they will be enforced before expanding the scope.