I really like that this proposal seems to come more early in the development process compared to other such changes lately and it seems to explain more than the average initiative. On the other hand I really don't like that the solution seems to be to put so much extra burden on curators, demolish the Q&A model over Reddit style threaded replies and in general doesn't give that much emphasis on content quality.
It seems to me to be something like: hey let's be like Reddit and additionally we can also be like Wikipedia in a two level system.
But I believe that:
- Reddit like threaded replies are not an efficient way for gathering knowledge
- Distillation of the knowledge from such conversations is a non-trivial, extremely time-consuming act
- While it's maybe optimal in the sense of (no downvotes, no close votes, I get my half-baked question answered and improved and elevated for free) no friction experience for the asker, it's a far worse experience for the curators and answerers. It would quickly overload the experts.
This would only make sense if the shortage is really only on the side of question askers and not also on the side of answerers or curators.
Why can't the elevation vote not simply be traditional upvotes? Why can't the workshop not simply be the staging ground for everyone (possibly extended)?
You basically ask us to do everything twice: like please give everyone a quick answer to whatever they write, and then polish it and answer it again, but even better. Be a relentless live debugger/help desk and a book writer. But who actually has time for that? I may be wrong, but I think the one cardinal problem is simply that we don't have the amount of free labor needed to make this paradise a reality. (AI has, we haven't. That's simply how it is.)
I like Shog9's wisdom from the past, but I don't like his analogies. Analogies always leave important parts of reality out and can easily be changed into anything else. Just replace water with garbage and the picture might become more clear.
I personally would have done this instead: I would have decoupled questions from answers and focused on finding and linking existing content to new questions (maybe without much voting). My premise would be that most new questions are duplicates, some aren't and curation/expert time is limited. This path you explicitly do not seem to follow.
It seems that the answer to AI chatbots seems to be that you can write anything and human experts will give you tons of immediate help. I'm really not sure we can beat AI at its own game, especially not without high quality content.
We will surely learn from this big experiment if Reddit like threaded replies are useful for knowledge generation. Unfortunately that is your weakest point. You don't measure quality/knowledge generation. You only measure engagement. This engagement might just be meaningless for the purpose of the library creation. It's absolutely unclear and my impression was that people are not getting help with the opinion questions (and there wasn't a single positive example in the Q&A asking for positive or negative examples). That means if, if you try that (and I think you will), you will have no chance of seeing if it goes wrong. You go into it blind.
In the end we might just only collect a pile of garbage (questions didn't get closed for fun in the past). But it's worthwhile to find out. New usage of SO is so low now, one can as well take more extreme measures and bet the whole house, there isn't much to lose.
P.S.: I slept over it and now I think that the vision is basically a human super chatbot knowledge machine. A chatbot that can answer every question without any problems or reservations, extracts knowledge on the go and is powered by an endless supply of free human labor. Problem is that it still competes with AI chatbots, they might still be faster and cheaper and this endeavor might be futile. Additionally it's not clear if the knowledge base really needs the chatbot output to distil knowledge. Instead it might be more efficient to completely ignore that and start generating knowledge by other means, whatever they are.