Okay, another answer, this time, even though this might not be something that I would take part in personally, I try to make it happen. Here is your new business model:
It seems clear to me that you want to become like Reddit only with a bit higher quality. Fair enough. If that was indeed the goal, how would one do it?
- Askers can ask anything, repliers reply to that in any possible way, voters vote on it - all like in Reddit
- Do not ask people to dig through the mud and search for diamonds. You won't find enough people anyway. The only thing that can do that without burning out is AI. Use AI to write meaningful excerpts of question and answer for questions with a score above a threshold. Train the AI to take votes into account. How you do that exactly? It's your business secret. If you do it well, you'll create value. If not, then not. That's your big task. The new core of the business is knowing how to automatically extract knowledge from a water/mud/garbage mixture.
- Publish these excerpts to the second tier after some time. Make them editable there for trusted users. Add links between the original discussions and the edited summaries.
- Very important: every question gets only a single summary which summarizes the question and the answer, even if there are five different solutions outlined in five different replies. The expectation would be that all sufficiently upvoted solutions are included in the summary in the order of their votes.
- Visitors can them consume the summaries, which hopefully have a quality high enough to compete with LLM results (i.e. they may not be tailored but they explain more and are right more often).
- Optionally add a duplicate finder system, where related questions / summaries are linked as early as possible, without limiting the ability to comment on anything endlessly.
- State that you want to become a high-quality Reddit alternative.
- No AI content creation by users, only by you, only in the one intermediate summarizing step.
Simplified user roles:
- askers ask questions
- answerers reply to questions
- all visitors vote on content (up and down)
- AI sumarizes vote weighted content
- experts (distinguished answerers) edit AI created summaries, maybe if they want to can even bypass the AI part, but let's see if this route is frequented
- visitors read summaries or ask their own question
- no close, deletion only of spam, instead only votes and visitor counts
Anything else, especially when it puts too much burden on curators, stay away from it. More AI, also stay away from it. One crucial AI application that you must do well and one only. You don't want to fail. Make it simple (For example are question types really needed?).
And would I like it personally? Not that it matters because you surely can find a different audience elsewhere. I really like that it's more oriented towards improving content, more collaborative. And I like summarizing multiple answers. This concentrates knowledge much more. But I don't like conversations. My time is too precious to take part in them. Still, if empirically it's shown that this is the only way to get humans to reveal their problems .. what use is a system that nobody uses. I definitely don't believe in that all questions have been asked already, so making a system where people ask them can't be too bad. But I myself would not volunteer for it probably.
The alternative is that everyone uses only chatbots in the future and amount of publicly available knowledge will shrink. Maybe generating new knowledge will actually become harder not easier.