15

SE announced OverflowAI today, another attempt to incorporate GenAI into the sites. From reading this post, the uses actually sound like they could be helpful; I certainly hope they are. If they are not, though, and they are poorly received by the community, does SE plan to listen to and act on feedback? Of course, there is likely to be disagreement within the community, but if the clear community consensus is that a feature is harmful, will management at least strongly consider disabling it until all major issues have been fixed?

Improve this question
8
  • 14
    I mean... the answer to this wouldn't be all that... useful. Will they seek input? of course. They've already made posts on mso where we can leave feedback. Will they listen to feedback? Of course... though past cases of "listening" to feedback has had mixed results. Some people will feel heard and others won't.
    – Kevin B
    Jul 27 at 15:55
  • 4
    They listened to feedback with the formatting assistant.
    – Rebecca J. Stones
    Jul 27 at 16:05
  • I'll be surprised if these AI features fare much better than the formatting assistant. The underlying technology is too unreliable.
    – ggorlen
    Jul 27 at 16:09
  • 4
    Rosie's comment on another discussion says yes, the company is actively soliciting community feedback and I'm sure they will use that input since they went through all the trouble to get it. I doubt any company would commit to blindly doing whatever the community consensus says they should do.
    – ColleenV
    Jul 27 at 17:14
  • 4
    All questions about the future are somewhat problematic but judging from the past, I expect something in the middle. They might listen sometimes to the community but that might also be just a coincidence.
    – NoDataDumpNoContribution
    Jul 27 at 17:56
  • 2
    May I have misconceived, but this announcement seems to suggest that the use of AI would interpose with the audience that perceives, or seeks, the SE sites only as a helpdesk. Of course, I am making a personal interpretation of the announcement, and I know that SE Office is not run by angels.
    – Augusto Vasques
    Jul 28 at 4:23
  • 4
    @RebeccaJ.Stones You can’t use an abomination like the formatting assistant as an example of listening to feedback… That trash can of theirs was an interdimensional earth quake generator so strong that it sucked itself right into its own earthquake portal thing whatever you wanna call it.
    – Andreas moved to Codidact
    Jul 28 at 17:25
  • 1
    @ruben, I don't think it's necessary to tag with stackoverflow-labs if it already has overflowai. My understanding is that the project "belongs" to SO Labs.
    – starball
    Aug 28 at 0:38

4 Answers 4

Reset to default
40

I wouldn't hold my breath.

This push to AI has been pockmarked with the declaration of "community inclusion" and acceptance, ever since the first blog post the CEO put up about this went live, but has been marred by a history of simply charging forward in the face of feedback that questions the worth of these kinds of things, with limited changes being made to shape their vision of the service or product.

It's kind of an unfortunate circumstance to be in. This would be an ideal opportunity for the company to really come together to work on what it is they want, or if they want something at all. However, the fist on the scales here says that "AI is the future", and that much really doesn't seem to be up for debate.

So if you want them to listen, well, I wouldn't expect much. But you're more than welcome to give them feedback. Who knows - they might read it.

Improve this answer
1
  • "what it is they want" - money!
    – OrangeDog
    Aug 28 at 12:09
15

If by community input you mean to use Meta Stack Exchange input, it's possible but might not have the weight we might be expecting.

I have learned that we are no longer "The Community". We might not be that in a long time. The Meta Stack Exchange users that actively participate here and in other places look to be just an "affinity group" that is part of whatever Prashanth, the Stack Overflow, Inc. current CEO, might mean when he mentions "community" in their posts.

His last post here was on October 2020. While he might have decided to use his time and "powers" better, apparently, he is looking for direct community input elsewhere. I don't blame Prashanth; it just looks that Meta Stack Exchange is not well suited for the CEO having direct input from an "affinity group". Instead, they have preferred to use other means like Twitter, LinkedIn, and surveys, among other ways, to get community feedback, probably giving high weight to the Stack Overflow annual survey despite the feedback about the bias, among other problems shown on it.

Note: Prashanth finished his WeAreDevelopers talk on 2023-07-28 by asking the audience to reach him on Twitter and LinkedIn.

Reference

Community

Surveys

Related

Improve this answer
2
  • 6
    In fairness, the CEO doesn't have to post his own questions here on Meta to consider feedback given here (on questions posted by staff - perhaps sometimes at his request - and even on questions not posted by staff). It's quite a leap IMHO to say they prefer surveys or to conclude that he must then ignore all feedback posted here (a fair conclusion based on actions, probably, but not based on the absence or presence of his own questions).
    – Stuck at 1337
    Jul 27 at 17:21
  • 1
    @Stuckat1337 Thanks for your comment. I agree that there there are things to improve on my answer. Regarding "fair conclusion" they can also be made based on the "lack of actions". Music has silence; without correctly placing silence in music, it will not be pleasant; it will be just like an airplane motor noise, annoying. While we still use airplanes despite the unpleasant noise, that doesn't mean we can't draw conclusions from airplane motors performance.
    – Rubén
    Jul 27 at 17:29
3

If the input is

we don't want genAI at Stack Exchange

then no, they won't seek it, nor act upon it---the conversation has moved on. The decisions to incorporate genAI into Stack Overflow and subsequently Stack Exchange were made months ago (irrespective of our feelings towards it).

Useful feedback has been repeatedly solicited by Stack Exchange staff (meta posts; focus groups; and maybe more I haven't heard of). Rosie commented:

Features aren't set in stone and for the public platform features specifically, we need community input and feedback.

(I'm fairly sure they said words to this effect on the podcast, too.)

Stack Exchange have explained some things being implemented are experimental, which means some of them are inevitably going to "fail" (i.e., we succeed in figuring out how not to do things; laying the groundwork and gaining experience). It's why I ask questions like this. It'd be nice to get features more welcome by the community, and save time by identifying what's going to go wrong before committing too many resources.

Barring some historical blemishes, Stack Exchange have been incredibly interactive with users when it comes to development. For comparison Reddit, a few weeks ago, suddenly decided to retire "coins", which some users paid real-world money for. They had a strike not long ago too, which the CEO completely ignored; he didn't even take it seriously. Stack Exchange is listening, although they have to include input and factors outside of meta.SE and meta.SO (marketing, business, etc.).

In the end, if Stack Exchange want us to actually use genAI tools productively, they're going to need to develop these tools using user feedback.

Improve this answer
1
  • 14
    SE has a giant pile of tech debt that has been swept under the rug and ignored for the sake of this. There's a lot of feedback that's ignored. I don't think at this point that SE as it is has the benefit of the doubt when it comes to the desires of the folks who're most active here. We're not yet at the back to point of community driven development and management
    – Journeyman Geek
    Jul 29 at 2:29
-9

Will SE seek and act on community input in developing OverflowAI?

Yes, they'll use the community-generated data to train their AI.

Improve this answer
6
  • 16
    I doubt that's what was intended by the question.
    – Mast
    Jul 27 at 16:47
  • 6
    @Mast it's probably someone's intent though.
    – Franck Dernoncourt
    Jul 27 at 17:23
  • 5
    This doesn't answer the question.
    – Cerbrus
    Jul 28 at 16:54
  • 3
    @Mast Who cares? It’s funny. :P
    – Andreas moved to Codidact
    Jul 28 at 17:27
  • 1
    @Cerbrus actually it does answer the question, quite literally.
    – user10186832
    Jul 30 at 18:45
  • No, it's just a sarcastic joke...
    – Cerbrus
    Jul 30 at 19:20

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .