Our goal with this redesign is to minimize disruption to what works while modernizing the experience to serve both new users of the platform and the needs of long-standing, tenured members.
OK, but you know, it really seems like the company only focuses on the former and largely ignores the latter. Every time this happens, the long-standing users group gets ignored until the complaints get loud (read: public) enough or your efforts bomb badly enough that you then feel forced to address them or walk them back, like how you are doing right now. What's that apocryphal quote misattributed to Einstein? "The definition of insanity is doing the same thing over and over again and expecting different results."
But we do have to move away from a world where closing questions is essentially the catchall and default option for curation.
I agree as do many curators, but minimizing or doing away with the ability to close questions will not do away with the need to close questions. This has the same bad logic as "if we stop testing for COVID-19, there will be fewer COVID-19 cases!" Sadly, a lot of recent company experiments or site changes on Stack Overflow have followed this same logic instead of actually measuring for useful outcomes.
When we said “no more closing questions,” we failed to take into account that that phrase would mean something very different to you than to us. For instance, we effectively draw a line between closing questions and deleting questions. Deleting questions is definitely still in scope for us.
You talk about doing away with question closure, but not doing away with deletion. So why didn't the announcement discuss how the new site would handle automatic removal of content in place of the lack of closure curation tools. I think we'd all love to discuss ways that the site can automatically handle curation such that we don't need to manually cast close votes, if what you're claiming is true. That would be a very ambitious and interesting technical challenge, the kind of thing programmers love! But why not mention that at all? Personally, I suspect it is because this is a sudden ex post facto justification, not something that was planned from the outset (or did you not run that beta site post by any curator or moderator who has been using Stack Overflow for more than a year or two to see how that bit would go over before posting it?).
Remember the age old adage, "build it, and they will come"? Current top-level company leadership does not seem to understand or remember what was built, or why people came in the first place: that Stack Overflow was built to be a repository for freely accessible, high quality, and eminently reference-able questions & answers about programming languages and tools.
Q: So, why is there such a big disconnect between the company's terminology and the community's terminology?
A: Because the company doesn't use its own product anymore.
Why are your employees not active users on Stack Overflow, asking and answering questions here? Sure, not every role is a programming one, but a lot of them are. Or at least, a lot of them should be, given the industry you're in. And for the non-technical ones, there is no shortage of non-technical network sites. It's important to not just acknowledge that there was a disconnect in language between the staff and the community here, it's critically important to ask why. A company whose staff regularly uses its own product will always be much better positioned to maintain and improve said product for the betterment of all. And with a better product comes happier users and better PR and thus ultimately more customers!
As it happens, we the community also draw a line between closing questions and deleting questions. There's a fair bit of overlap between the two actions, but any seasoned curator will agree that closure does not equate to (even eventual) deletion, and that deletion can/should also happen without closure at times.
Right now, you're hemorrhaging that first group of users you talked about to AI tools because of a perceived accessibility (call it friendliness or elitism or whatever else you want) problem. That makes perfect sense; AI tools always answer your question (except when they can't, or worse, give you a wrong answer and claim it's correct), never present roadblocks to asking a question in the first place (except when you've hit your free quota), and are always polite (except when they encourage or enable abuse, harassment, or death).
The user exodus hasn't been a complete, 100% drop-off because, as it turns out, AI tools are simply not as reliable or trustworthy as real people/experts, and by their nature never will be. But if you get rid of SO's curation tools, or make their usage more difficult, or obfuscate access to them in any way, the latter group you mentioned (experienced, tenured members, or SMEs) will continue to drop and you will have neither incoming new users nor experienced, tenured users.
Specifically, I’d love your thoughts (as answers here) on:
[...]
The purpose of curation tools: Are they meeting your needs? How could they be improved? - What are we missing?
The necessity of all current tools: Are there tools you feel are redundant or no longer useful?
Reconfiguring tools: Are there ways we can adapt existing tools to better serve the needs of the community?
Oh man, today is your lucky day! There are so, so many ways you could address the site's problems. Here are just some of the things often-requested by community, or widely popular with them, that you can do which don't require the Redditification of the site's UI/UX, and that I believe go much further in fixing your core perceived problem of "everything gets closed all the time and new users hate that" (in no particular order):
- Improve the SO Search functions for the first time since, literally ever. Why does SO search still suck massively? Why are there entirely different search functions used for the main search bar vs. the duplicate closure finder?
- Give the new Markdown editor feature parity with the old one. Feature parity should be a hard requirement for any feature overhaul/replacement the company does, FYI.
- Abandon the 2025-era comment redesign and just implement comment threading, even if it's just one-level, under the classic comment design, with a single button to collapse each top-level comment thread. And allow people to @-mention at least two users in a single comment.
- Overhaul site privileges to work more based on related feature usage rather than reputation. This is something Catija was exploring heavily while she was still here. This means not just replacing rep-gated access with usage-gated access, but also increasing the abilities users have with regard to moderation/curation actions based on the number of related feature uses they have which are 'helpful' or 'accepted'.
- Fuzz the displayed score of main/traditional questions once they get below -1; just show "< -1" or something until expanded. That way users with moderation capabilities who might have a use for seeing the actual score can still do so, but the default score displayed is never so off-putting that someone has to read -6 or -15, etc., on their own question, however poor a question for Stack Overflow it might actually be. That's one of the biggest instances of friction for new users: seeing their post get 'downvoted into oblivion'; it's insane to me that this was not addressed like a decade ago when Stack Overflow first started realizing it had an image problem.
- Add higher protection levels to questions, which automatically apply after a question gets more than a certain number of answers and require progressively higher tag scores in the question's tag(s) the more answers there are (if a question has 10+ or 20+ answers, we should not be letting any random user post a new answer anymore).
- Follow through on improving chat with new features/functionality. We got all excited when you said you were improving chat during one of the sprints last year and so far all I can see that changed is now we have a ToS screen nobody reads and 1-rep users can access chat. Oh and they redesigned how chat replies look but rolled half of it back after a few days.
- Update the Roomba to ignore comments on zero-score questions with no answers. After one year, just because there are 2+ comments on a question does not mean it is worth keeping around. At the very least, make it require that it have at least one comment by someone other than OP followed by at least one comment by OP, if you're gonna keep those questions around.
- Reward duplicate closure finders with badges or something. It's only one of the most-requested and upvoted features, given how poorly the search function works.
- Update the Roomba to ignore accepted answers.
- Add parent language tags for SO, make them required and separate from the 5-tag count for questions, and restrict gold-tag closure powers to parent tags only.
- Expand gold-tag closure powers to allow unilateral closure for more than just duplicate reasons. SMEs know when a question needs more details, needs more clarity, or is a subjective matter.
- Allow mods to migrate questions older than 6 months (require 2+ mods if you need some higher threshold for whatever reason), and explicitly encourage that old, off-topic, but well-received questions to be migrated to their appropriate sites. Why should a question with a score in the triple digits with tons of answers and millions of views languish as closed or historically locked on Stack Overflow when it could be migrated and reopened on, say, Superuser or Unix & Linux or Server Fault? Just because it's from 2015 and not 3 weeks ago? What a short-sided restriction! Not all such questions would be good fits for migration, for sure, but many of them would be. It's a great way to grow those other network sites' traffic, too.
- Notify users automatically via site notifications when a question they closed is edited.
- Allow users casting 'Needs details/clarity'/'Needs MRE' close votes (only votes, not close flags) to leave a custom comment posted by the Community user (with mod and staff visibility on who actually authored it) explaining what details are needed, or what is missing to constitute an MRE. A lot of the feedback about SO being unfriendly or 'SO always closes your questions' comes from no direct, clear information (since askers are inherently often lacking understanding in the first place). Sure, people should read the close banners, but those are always generic. Meanwhile curators absolutely don't want to announce that they are the ones during the curating because that often leads to harassment. An anonymous, custom comment option would satisfy both sides' concerns there.
- Add (optional) version tags to answers. Seriously, why the heck don't we have this? It should have been obvious even in 2008 that this would be needed.
- Improve the Ask Question Wizard so it actually moves people and their questions to the appropriate site when they post a question we know is off-topic on Stack Overflow, or tell them straight-up that their question is not suitable for Stack Overflow. Alternatively... just finally do away with all the different sites and instead just have one big site where people can filter questions based on 'topic'. Otherwise those other network sites are never gonna succeed or thrive like Stack Overflow once did.
- Improve the Staging Ground so all new users and a lot more new questions from existing users are sent through that process. This includes applying the Roomba to questions there, and allowing users to edit SG questions that have already been graduated, or overhauling the thing so that it doesn't create a separate question on Main, but rather just 'moves' the SG question itself to Main, and is no longer in the SG.
- AI Assistant might be great for people who want to use it. But a lot of people don't want to use or even see it, and it's insulting that it's given top billing on Stack Overflow, which was founded to be a place for getting answers directly from human experts. You need to let users hide the AI Assistant/turn off all things AI without resorting to user scripts/styles.
- Be more clear to users, especially new users, when posting answers that AI-generated content is not allowed here.
- Make downvotes on Discussions reorder and hide Discussion posts like they do on Main/traditional questions. In other words, Make Discussion Downvotes Matter(TM).
- Bring back Winter Bash, but make sure it has an opt-out option.
- Bring back the original SO Jobs (Joel Test and everything, except add a scoring item: "requires or involves using AI tools" would be an X on the scoring rubric) and the SO Developer Story. The Dev Story especially surely took such little bandwidth and support effort that I can't believe you ever got rid of it, given how popular it was.
- Bring back SO Documentation, just don't award reputation for content contributions there. Do more work to ensure quality (don't make the barrier for entry so low). This means much slower growth/build-out, but it's worth it if the goal is quality and not 'overnight golden performer'. It should have been treated more like Wikipedia and less like gamified Q&A.
If you stop work on UI overhauls that break/remove a bunch of features, and start implementing the things listed above instead, I guarantee you will see people coming back to SO and its reputation in the public start to improve again. You could even still work on native, first-class genAI features, so long as it's alongside and secondary to features like those mentioned the above list.
The list above is just off the top of my head (OK, and maybe the middle of my head, since I started this answer ~2 hours ago)... there are dozens more big ticket changes the community has actively been clamoring for that aren't "yet another site design change", which the community has not been actively clamoring for... ever.