The primary aim of these experiments was to encourage more users to comment, which would increase the number of comments, while avoiding any overall negative impact.

Having more comments on the site is not a positive outcome!

Stack Overflow is a site for questions and answers. Comments are useful for requesting clarifications or suggesting improvements to questions or answers. But simply having more comments does not mean that a change is positive for the site or the community:

• Comments just saying "thanks", "+1", etc. are noise. The new design requires more clicks to flag unneeded comments, leading to more useless comments staying on the site.
• Having the "add a comment" box at the top instead of the bottom, encourages people to write comments before reading the existing ones, leading to more superfluous comments saying things that others have already said.
• Threaded comments encourage people to have multiple separate discussions on diverging topics, leading to more irrelevant comments which don't request clarity or propose improvements to the post they appear under.
• The old design used colour to highlight comments with more upvotes. The new design makes it harder to see the most useful comments, so when people go to check if someone has already written the comment they want to write, they are more likely to miss it, leading to more duplicate comments.
• Allowing code blocks in comments encourages post authors to add clarifying details in comments, instead of improving their post by adding those details with an edit.
• The old design had some instructions: 'Use comments to ask for more information or suggest improvements. Avoid comments like "+1" or "thanks".' The new design lacks these instructions, presumably leading to more of these comments we don't want.
• Displaying comment ages capped at "a year ago" encourages more responses to out-of-date comments, by hiding how out-of-date they are. (An SO staff member has confirmed, twice, that this is the intended consequence of this misfeature.)

OK, so you said "while avoiding any overall negative impact". But how have you measured that negative impact? Have you measured it? The way your post is written, it sounds like you just went with the design which maximised the number of comments, without subtracting any metric for the negative impact.

Personally, I would say that the raw number of comments is actually a quite reasonable rough metric for the negative impact of the change. Most posts only need a couple comments to iron out whatever improvements can be made to them, and any comments beyond that have diminishing returns and then probably negative value. Moreover, as mentioned above, the mechanics of the new design encourage more bad comments, but it's hard to see how any of the changes encourage people to write more good comments.

So, sorry to be a downer, but my belief is that you have optimised for the change which will cause the most harm out of all the designs you tried. This harm will be compounded in the longer term, as the presence of more useless, superfluous and irrelevant comments will teach new users that such comments are acceptable.

I suggest you scrap the harmful changes entirely, and for your next experiment, adopt a metric which measures the quality of Q&As. As @user56reinstatemonica8 notes, when you announced this 'experiment' you said the metric would be constructive comments, not just the total number of comments. One way to count good comments would be like this:

• A comment is made by someone other than the author of a post,
• After the author sees the comment, they make an edit to their post,
• After the edit, some other user upvotes and/or reverts a downvote on the post. Bonus points if this is done by the user who wrote the comment, and bonus points if the commenter also deletes their own comment (indicating it was resolved).

Or this:

• A comment is made by someone other than the author of the post.
• The author of the post responds with another comment.
• The first commenter sees the response and makes an edit to the post. Bonus points if they delete their own comment and/or flag the author's comment as "no longer needed".

These would be much better metrics for judging whether a change to the comments system is actually beneficial, because they would directly measure whether the additional comments are helping to improve the actual content that people come here for.

This is textbook Enshittification

You've started by choosing a terrible metric.

Comments have until now been seen basically as a necessary evil. A huge amount of effort has been spent on features to make comments less necessary, more transient, and easier to clean up.

It's like a supermarket which is losing sales measuring "number of people using the customer toilets".

Let's be clear: supermarkets should have customer toilets. They should be functional, and reasonably pleasant to use. They should not be the most attractive thing in the shop. And if every customer is needing the toilet before they leave, maybe you need to reduce your queues or something.

Then, in order to improve that metric, you've deliberately made things worse.

The most obvious example is the "more than one year ago" label - taking information away from users, to trick them into "engaging" in a completely meaningless way.

It's like the supermarket tricking people into walking into the toilets by making them look like the exit. Who cares if everyone trying to exit is now really annoyed, the metric went up!

The same goes for the outsized vote and reply buttons, and the removal of help text about what comments are for.

Maybe threading is a good idea, I don't know. But neither do you, because you measured completely the wrong thing, and mixed in deliberately bad UX.

Enjoy your busy toilets.

Is this release coming with the anti-accessiblity feature that all dates are displayed as "over a year ago"? Please, I hope not, it's a terrible experience for those that can't access information hidden behind tooltips. This was requested, but like a lot of stuff around this "improvement", has been ignored.

Are keyboard shortcuts fixed? Last I checked, they weren't.

Is this release going out across network, and to questions, so that there's a consistent experience, and so that other sites can feedback on (how poor) the UX this has been for long term users, so that maybe you'll fix some of the issues that have long been ignored and forced many to turn experiments off?

Please explain how celebrating

The new commenting UI saw a statistically significant higher conversion, estimated to increase monthly comments by 5-8% and introduce 2,000+ new commenters.

is consistent with the Tour page's

Why is there a giant header "3 comments"?

Why is the button to create a new comment as wide as the post body?

Why are the timestamps crap? (I think enough people have commented on the details here)

Why are comment author and the comment action buttons on separate lines, using up more space than necessary?

Okay, the last point is probably a bit tricky to implement. But I think it's worth it to optimize for space here a bit as otherwise comments get excessively noisy and large.

Just throwing a bunch of flawed designs into action and choosing the one with the best arbitrary engagement numbers is not a particularly good way to develop features like this.

Another specific valuable feature of the old comment UI that has been lost and makes using SO worse: good comments no longer stand out.

There used to be a really nice visual hierarchy: answers stand out from comments, then, individual good comments stand out from the others, through the bolder colouring of the upvote number. The whole comment section took a step back, then valuable comments took a small step forward.

Here's a real-life issue I just hit on reading an answer. In the old UI:

The important +14 comment about wss: really stands out and is hard to miss, even when skim-reading, thanks to the deeper reddish colour of the number in a calm-low-contrast area. This is good. It's part of one of the things that really made StackOverflow stand out from the competitors - the important information stands out, the rest is present but easy to skim over.

In the new UI:

...everything is shouting at the same volume. I'm posting this here now because I completely missed that important comment on the first read, until I returned to figure out why the answer hadn't worked. Nothing about the comment block stood out from the rest (and the whole block looks like the comments under a blog or news site, so I think I just instinctively tuned it out).

That's on an answer with only 3 comments. On an answer with lots of comments it's worse:

The +33 one doesn't stand out at all unless you've learned to specifically look for such things. In the old UI, it did stand out:

It was a good design feature that helped first-time users and regular users alike to not miss important information.

Side note... I also just noticed you have not one but two "Add a comment" buttons and you show them even when the comment section is collapsed. What possible justification could you have for encouraging people to add new comment before they can even see if their comment has already been posted, other than to game the metrics?

D O N ' T    W A S T E    M Y    S C R E E N    S P A C E   ! !

I like to think of myself as pretty easy-going. I skim meta, usually just clicking on one of the "featured" posts.

The previous concerns raised repeatedly by everyone on SO are all valid points; but they are just esoteric enough that I let them slide off my back. I'm not a mod, so I can deal with an extra comment to one of my questions or answers, ok fine. Crappy timestamps? I've gotten used to that in so many other places that wouldn't know UX even if they could spell it, in the name of user "friendliness" - Bleh, I'll choose to ignore it   :- |

But this change ticks me off!!!

I hit a cached SO page that I'd looked at a day or two ago for... something - you derailed my train of thought frankly. DON'T DO THAT. It's really annoying

So I was going to vote up a comment, but it told me I wasn't logged in. So, OK, I log back in, and

BOOM

after the refresh all of a sudden I have a screen full of comments. They're taking up so much space I don't see the question anymore. They're taking up twice as much space as before.

When I scroll to find the question, I scroll over it. Between the amount of space they take up and their formatting and the prominence of all the avatars with their dazzling colors -- that's all I see.

Yeah, sure, I can see the question, but why do I have to dig for it? Why do I have to buy a larger trackpad now to practice long jumps over the comments to get to the next answer?

The site is barely usable to me now. And it's SO irritating (pun intended) that it's ruined my evening when I was actually making good progress on debugging a complicated issue. (Yeah, OK, working on the Friday evening of a holiday weekend. I've got no life. That's beside the point)

This is NOT Reddit Overflow

This is a Q&A site - NOT a comment thread discussion site. If I want to sort through a lot of cruft to find something useful, I'll go over to Reddit and do that.

Maybe, just maybe this kind of change would be ok on the "Arqade" or "Science Fiction and Fantasy" or "Movies and TV" communities.

But this is a tech community. We're NO NONSENSE. Don't give me color unless it means something. I don't want sounds of random object from marbles to bowling balls falling on a tin roof, just to distract me with something I don't care about right now. I don't want stupid auto-replacement of emojis when I send email about Date::Parse and get Date:😋arse! I don't want emoji's at all - An ASCII face is plenty >:-(   I program in Vim. It just stays out of my way.

That's what I want from SO - STAY OUT OF MY WAY!!!

OR, put another way, just give me The Truth, the Whole Truth, and Nothing But the Truth Anything else, and that's when you hear the gavel crack and a command for ORDER!

Have you enjoyed the waste of your screen space here? Good! Because that's what we are all experiencing now, too, on top of all the other effects this change causes.

pssst - did you know you have a UX community? Did you ever think of asking them?

Is there going to be an option to turn off the new and improved worsened UI. Especially since it:

• hides information some want to have
• massively distracts from actual answers
• massively increases wasted space

Up until graduation, one can at least opt out of experiments; now we lose that option.

The design below was the variation that performed the best.

Well, amongst the variations you've tried; there were many suggestions to improve this design. I will put aside all the functionalities that broke due to this. Keyboard shortcuts and userscripts are not akin to CPU overheating, but let's brush past all the bugs as they've been already brought up in the other answer and elsewhere.

However, what about suggestions that already bowed to the idea that we need more noise on the page? I did bring up balpha's userscript:

Or even a simpler change of moving the vote/reply/... to the side to save a little bit of space, so we can see a little more of the actual content (i.e. answers):

The parallel experiment focused on lowering reputation required for commenting remains ongoing at this time, and we’ll be following up with more about that in the coming weeks.

On another note, have you thought about the effects of making flagging harder (hiding it) while experimenting with lowering rep requirement for commenting?

Let me go on a bit of a tangent here: every time I see an experiment and I say that it is not a viable product, I get this almost boilerplate response that this is just an experiment. Yet, every time nothing meaningful changes at the end. We get some line like this performed the best and the sub-optimal design stays around... touché.

There's a lot of concern about what metrics you used to judge the experiment a success, and I think a feeling you moved the goalposts, replacing goals that are positive but hard to measure with ones that are negative but easy to measure.

In the initial post announcing the experiment, you said:

Our primary metric will be an increase in constructive engagement within comments. We will also closely monitor comment flags and deletion rates to understand the impact on content quality and moderation load.

The response was mixed: many specific disliked details and general distrust, mixed with some cautious positivity and agreement that this was a worthy goal. Then, here, you said:

The primary aim of these experiments was to encourage more users to comment, which would increase the number of comments, while avoiding any overall negative impact.

This is very different:

1. "constructive engagement within comments" (a good thing that is hard to measure) became "increase the number of comments" (a bad thing that is easy to measure).

   • Comment volume might be (usually?) a good thing on a social networking site but is not a good thing on a site where comments are more about problem-reporting and user-led moderation than social interaction or content creation. It would be like a social network rebranding their "Report a problem" button as "Engagement".
   • The initial post seemed to understand this, and treated comments as a means to an end not an end in themselves. This post and the staff replies seem to treat all comments as positive "engagement" which must be increased at all costs.

2. "closely monitor comment flags and deletion rates to understand the impact" was going to be a key part of the evaluation, but was relegated to a vague claim that the changes are "avoiding any overall negative impact" without any evidence cited.

   • What evidence is the claim that there is no negative impact based on? A lot of posts have asked but I don't see direct replies to this.
   • In particular how are you addressing the concern that overall increases in the number and prominence of comments will cause damage that builds slowly and is hard to measure (and hard to undo)? We'd expect slowly lowering satisfaction among users who like solving problems, obscured by quickly increasing satisfaction among users who enjoy arguing with strangers, leading to a slowly-accelerating long-term decline. How do you measure that in a short experiment?
   • If your metrics included things like user satisfaction, did you include users who opted out of the experiment? Or did you filter it to only include users who liked the experiment enough to not disable it?

Apart from the lost functionality with time stamps the biggest issue for me is the dilution of useful content by giving comments so much more importance and screen space now compared to before. This makes the site much less informative and lowers the value for almost everyone considerably.

From the more than 90 million comments on SO, surely 80 million are purely rubbish. And you decided to display them more prominently. Hey, look at our rubbish, we have the best one in town. In comparison, current LLM interfaces often look much more focused and clean. It's sad but it's true: you want us to go. You hide the real value, which is predominantly in questions and answers.

How could this happen?

Your metric is broken in multiple ways. Not only is not every engagement equal but there is more and less useful engagement. And while number of comments are one part, you probably didn't check if there are more or less questions in the long run. My guess would be that the higher number of comments will come at the expense of lower number of questions and answers, overall higher dilution of content.

You seem to have given up competing with LLMs, which curiously enough offer much higher information density, and instead see your future as a social network. That's fine by me, although I personally have no interest in tons of programmers chitchat. But others may be interested in that. I can easily go somewhere else and you can continue with the development here. No problem. I'm just interested in understanding the directions you take.

Time to take out the graphics I used when you started to discuss the topic on Meta.SE. Do you see the problem?

No solution for the dilution.

P.S. I remember that I liked the three lanes model introduced in the community AMA in March this year. I thought it allows for width (different kinds of content) and depth (high quality) at the same time. But this design change now is the complete opposite and instead mingles everything. Who needs chat if you can chat in comments? Who will find interesting content if it's hidden by chats? Obviously engagement as metric has no sense of usefulness of the engagement nor long term effects.

Not even mentioning the waste of screenspace by any other of the improvements, we also lost information to meaningless screen clutter, that is even longer than the original information. strlen("Over a year ago") == 15 vs. strlen("2014-04-09") == 10. Working with long running frameworks it is VERY relevant if an answer is (in the case of Qt for example) from the Qt4, Qt5 or Qt6 era, since not everybody specifies their used version.

bug

The UI truncates comment-replies after telling the user it will fit. I think it's including an invisible @user plus 2 more characters, but not counting that against the 600 character limit that's shown to the user. (The user I was replying to has a space in their username; without the space it's 12 characters; with a leading @ and a trailing space it's 14. I haven't tried this when replying to a user without a space in their name.)

After editing my comment down to not exceed the available length the UI told me, so I had "0 characters left", I posted. The end of the last sentence was missing. Clicking edit (which now takes 2 clicks with mouse movement between), it was still gone, but the UI told me I had 15 characters left. Typing them back in, they got truncated again.

(This was already reported in early July on Experimenting with Threaded Replies I hadn't checked that thread.)

Other functionality regressions that haven't been mentioned yet:

The UI only shows you how many characters are left when you're within 120 of the limit. The old UI shows the count all the time, so I can more easily estimate if I have room to start on a thought which would take 3 sentences, or if I should try to phrase it more compactly and leave something out. It sucks to look for something that's sometimes not there; takes longer to be sure you didn't miss it than to just glance at a number that's always there. Starting gray and turning faint orange at <360 left, brighter at <240, and full orange at <120 was much nicer. (And red when over the limit, like the current UI). Maybe other people that don't type as much will like having one less thing on the screen? I hate it.

Also, undoing a vote doesn't ask for confirmation. If you mis-click, you could undo a vote and not be able to redo it. When I was first trying the new UI, I was puzzled at how long it was spinning for after I clicked, and tried clicking again since I thought it hadn't gone through. This ended up undoing a vote since I was still clicking when it finally did go through.

(I've been using the old voting system for over a decade, unlike potential new users of SO. So I'm used to the old asynchronous system which didn't wait confirmation, just popped up a message later if there was a problem.) But the old comment-vote system still exists on questions, and other non-SO stackexchange sites, so it's potentially confusing to have it different, but I assume having both is intended to be temporary. I liked the old system. Having a spinning UI element for a second or two is distracting when you already see another comment you want to reply to, or you want to edit the comment you just posted but can't because it's still spinning.)

Editing feels slow. You can't get back into the editor after noticing a second mistake for several seconds after saving the first time. I'm usually still proofreading as I hit enter on an edit. (To make sure at least the first edit gets in under the grace period, but mostly just to optimize for the common case where there isn't anything else I want to edit.) With the old UI, there's an [edit] link directly clickable, and there's negligible latency for either editing or saving. (If you finish your second edit too fast, you run into the 1 per 5 second rate limit on comment edits, but being able to open the textbox and be editing during those 5 seconds is a big advantage.)

bug

When you have two "reply" textboxes open under different comments on the same answer, posting one of them closes the other, discarding any text you typed. (I think I'm remembering that correctly; I think I just had it open to play around with the displayed length limit, I don't think I actually lost anything, but I would have if my ADHD brain had started another reply before coming back to finish one I'd started earlier.) The previous comment UI didn't let you have multiple comment boxes open at all, of course, which was also annoying when composing a reply to one comment when a new comment came in and you wanted to give a quick reply to that first while the other human is still right there, but at least it didn't tempt you into data loss.

bug?

The nesting limit seems arbitrary. After a couple levels, there's no "reply" button and you have to click "reply" on the parent of the comment you actually want to reply to. But do they get notified? Super unclear if that happens automatically, like if all previous commenters in a subthread will get notified. I've been typing @user to manually notify them in that case, but there isn't auto-complete for @user names in the new UI so this sucks a lot.

(Something possibly related to this was mentioned in Experimenting with Threaded Replies , but at that point the @user stuff was still visible and/or you still got auto-completion.)

Are we supposed to be able to notify other users who have commented in different subthreads of the same comment section? Surely by typing @username, which should autocomplete.

Update: Replying to the parent of the comment you're actually replying to does notify that person, at least in the case I tested. But their comment might have been the grandparent of the one I clicked "reply" on, so I'm still not sure exactly who gets notified if a third person jumps in at the deepest level. Is there a help page about this? https://stackoverflow.com/help/search?q=comments and https://stackoverflow.com/help/search?q=notification don't have anything.

Being sure that the person you're replying to will get notified is an essential feature for chat, and now we don't have that without manually typing a username.

My thoughts on the feature and layout

Threaded comments would be very very good, something that could hide clutter and make it possible to not arbitrarily delete even useful comments. The fact that useful comments (not just clutter and noise) can and do get deleted or moved to chat is one of the worst things about the comment system and how the community treats them. (Especially in lower-traffic tags like [assembly] and [cpu-architecture] where real clutter is rare and there is a lot of interesting stuff for people to add as comments on each other's answers, or to discuss possible performance experiments we can try to better understand something, for example.)

So I'm very much in favour of a better comment system that makes them less of a second-class citizen. That seems to me like it was partly just a necessity born out of the limited flat comment system that SO has always had. (But also because it's even better if the good stuff is in answers, not just comments. But that's sometimes a lot of work, and IMO it's better to have the discussion between experts kept around in comments until it makes it into an answer.) Other than arguments that comments shouldn't matter, I agree with 99% of what everyone else has been saying.

(Comments are great and have lots of value. But they should still stay out of the way by default, not take up lots of space and visual attention between answers for people just looking for answers.)

But this new UI for them is so bad that I'd rather have the old system. Apart from the bugs and worse usability when actually commenting, takes up way too much space and isn't collapsed by default. And most importantly, doesn't highlight the highly-voted comments in a partially-collapsed state. That was the most important feature that made comments valuable to future readers. It seems the new UI only has fully collapsed and fully expanded states. And even when expanded, the votes aren't visually scannable for highly-voted comments, so any important problems with an answer pointed out by comments will no longer be visible.

Do you want us to just edit our own opinions into other people's answers now, since comments are now much less useful to future readers for critiquing answers? They're mostly only useful for the people involved in the chat, or for capturing whole discussions. (Wow, I just used the word "chat" without thinking, but SO does have chat rooms. That's a separate feature.)

And threading isn't enabled under questions, where it would often be more useful than under answers since there are more often multiple different things address about a new question. (But please don't do that until at least changing the layout to something more like the old system.)

Going to add myself to the list, but disagree with many and say they measured the right primary metric, but then completely drew the wrong conclusion. Measuring comments is the right thing to do, but more comments is by itself bad.

You measured that more comments were made, but based on what you're saying you didn't measure any positive effect.

• Effect on new users: Negative, because more users will write chatty comments and then be told off. Plus from first hand experience when teaching at a coding bootcamp new users were already struggling with getting 'distracted' by comments on the Question. I literally had to 'teach' students to skip over the comments to the answers... And I have seen this happen to junior devs on my team as well where they copied a bad answer from the comments (the first answer they saw reading from the top).
• Effect on active users: Probably positive... maybe? Not sure of that though, because reading through the comments has gotten harder, but arguably there are some positives. If there is any group that might benefit it's maybe this group, because this is the group that should be engaging with comments, and they are the group that understand that comments should be ignored most of the time. But then again, ignoring comments has gotten marginally harder.
• Effect on power users: Negative, because they will have to spend more time cleaning up comments.
• Effect on ad revenue: Probably zero or negative, as the quality of the information has become worse, thus probably causing Stack Overflow pages to rank more negatively eventually on Google compared to more focused resources like blogs.

Is there an indication that the quality of posts went up? If it's shown that e.g. the number of edits to posts went up by the original author of the post then I would happily accept that as a proxy for the posts improving. And then maybe the negative 'comment count went up' metric might be outbalanced by 'post quality'.

The claim:

increase the number of comments, while avoiding any overall negative impact

Is just completely absurd. An increase in the number of comments is a measured negative impact. It might be a worthy sacrifice to make, but by itself it's a negative impact.

we intend to add the updated UI to questions.

Please don't do this. The clutter under questions is too much as is; please don't encourage users to add more to it. The great thing about Stack Overflow is that you can identify if a question is similar to yours very quickly and find its solution very quickly. More comments means the good stuff is hidden under clutter.

Also, the metric used to evaluate this experiment is backwards. Increased comments is not desirable. I would actually argue that if the experiment is showing that the new UI is increasing new comments, then it should be scrapped.

The new commenting UI saw a statistically significant higher conversion, estimated to increase monthly comments by 5-8% and introduce 2,000+ new commenters. First time commenters were more likely to comment again using the new UI, a trend that continued even as lower-rep users gained the ability to comment as a result of the parallel experiment focused on lowering reputation required for commenting

The metrics you pick are very important - and in this case a difficult but important one to gauge is - did the comments help improve the posts. In addition, how much of an effect did lower rep users having the ability to comment colour the data, and which of these repeat comments were reply to a older comment thread?

The UI is painfully busy, on dark mode at least:

Those grey borders around the upvote button, the reply button and the ... button are really not necessary.

Compared to the UI for the comments on questions:

(I haven't really followed this closely, are comments on questions getting the same UI too? Yuck.)

If the goal is to encourage newbies to comment, could us oldbies retain the ability to opt out? Thanks.

EDIT Yep, simply removing the borders makes it much easier on the eyes:

Designing a proper comment system should not be rocket science.

• On the main site(s), comments are generally not important so they need to be minimized as default and given a fixed space below each post. Those interested in reading the comments may then actively expand the comment field as they want but it should not be expanded by default.

  A huge flaw in the current system is that the comment field just keeps growing and growing and you have to actively scroll past it. Very distracting and really messy. It neither helps those not interested in comments nor those who are - the latter can't find anything because there are no threaded comments, just a big blob of random crap, possibly with some interesting comments in there somewhere.

• Comments should be threaded. Although having multiple discussions in comment thread mostly makes sense for the meta sites where there should actually be discussions present. Not so much for the main sites as of today, but it could add extra value if allowed, see "Are more comments a bad thing?" at the bottom of this post.

  This means that each thread can set a topic and then comments below that thread need to follow that topic or otherwise get deleted. This means that moderators can easier get rid of off-topic debates instead of as of now with the current dysfunctional system: migrate everything to chat & throw the baby out with the bath water.

• This isn't social media, so things like user avatars, voting on comments, "likes" etc is just bloat. It shouldn't be there. The size of buttons etc must be minimized so that each comment takes up as little space as possible. Meaning you need to place such things at the side of each comment, not above/below it. The current system is fine: "John Doe Commented Aug 28 at 18:22" where the name John Doe is a hyperlink to that account. And then buttons like expand/reply/flag/link etc can be implemented to the left of the comment just like in the current system.

• Obviously comments should be labelled with correct and exact dates, not just "John Doe posted this two days ago", that's unhelpful.

• Making it possible to link to individual comments is necessary.

• Code formatting inside comments would be a nice to have feature, for very minimal snippets, just 5 LoC or so.

And that's it.

I also wish to give some emphasis on this:

Are more comments a bad thing?
Not necessarily, if that the UI is designed so that comments don't distract from the actual post. It can often be very interesting to follow an on-topic technical discussion between several experts discussing the contents of a post. That is something that adds a lot of extra value to the site! It adds value for newbies and experts all alike and attracts them to the site. Sometimes comments can be more valuable than the post itself, depending on the reader's background.

But again, it needs to be implemented so that it doesn't distract from the post itself or worse, take over the post completely.

The new commenting UI saw a statistically significant higher conversion, estimated to increase monthly comments by 5-8% and introduce 2,000+ new commenters. First time commenters were more likely to comment again using the new UI, a trend that continued even as lower-rep users gained the ability to comment as a result of the parallel experiment focused on lowering reputation required for commenting.

We don't need that much more comments, we need clarifications about the answer/question, not spams! e.g. "+1", "-1" (the privilege to comment everywhere)

• Bringing a complete editor to comments

Noooo. I don't prefer that much space. There's already a markdown viewer for the users to preview their comments before they publish. Also it takes too much space for mobiles.

Separating comment voting privilege from comment flagging privilege. This means that commenting voting will be a standard feature regardless of rep.

biased voting will happen there

Also, the new UI seems way larger than it should be (How can I revert the style/layout changes to comments?)

Here's a UserScript to remove the comments from the answer to reduce noise on the screen.

// ==UserScript==
// @name         Remove comments from containers
// @match        *://stackoverflow.com/*
// @version       1.0
// ==/UserScript==
(() => {
  const idPattern = /^follow-ups-container-\d+$/;

  const purge = (root = document) => {
    root.querySelectorAll('div[id^="follow-ups-container-"]')
      .forEach(el => {
        if (idPattern.test(el.id)) {el.remove();}
      });
  }

  purge();
})();