-157

Update, July 27, 2023: I posted a summary of Prashanth's keynote as an answer below.

We wanted to let everyone know that Stack Overflow will be at WeAreDevelopers World Congress in Berlin, Germany on July 27 and 28!

We’ll be participating in several different sessions where we’ll share the work we’re doing to help developers with the challenges they’re facing in 2023 and beyond. If you happen to be attending in-person or virtually, here’s when you can find us:

  • Thursday, July 27: Prashanth Chandrasekar, CEO of Stack Overflow, will be on the main keynote stage to share how Stack Overflow is evolving to bring community and AI together. He’ll talk about how combining the collective knowledge of the community with technologies that increase productivity, efficiency, and rapid learning, can help technologists get to solutions faster.

  • Thursday, July 27: Prashanth will be joined by Stack Overflow’s former CEO and cofounder Joel Spolsky for a fireside chat to share stories about the evolution of Stack Overflow over the last 15 years and what its future holds.

  • Friday, July 28: Join us for a workshop session where we’ll talk about how the developer community can play a crucial role in how AI accelerates and helps with the accuracy and quality coming out of future GenAI solutions.

If you’re interested in attending, you can register here and use code WWC23-FRIENDS10 to receive a 10% discount.

If you can’t make it, we’ve got you covered. We’ll post a recap on this Meta Stack Exchange post after the event and share more details on Stack Overflow Labs, our central hub to share experiments, demos, insights, and news - across all Stack Overflow products. The YouTube video of Prashanth's talk on the 27th is here.

7
  • 176
    "...how Stack Overflow is evolving to bring community and AI together" ... this should be interesting to hear, in light of the ongoing moderation strike. It seems to me AI is ripping the community apart more than anything else. "...how AI accelerates and helps with the accuracy and quality coming out of future GenAI solutions" ... not sure "accuracy and quality" has much of anything to do with GenAI solutions at present, so it'll be interesting to see how future solutions fare.
    – ggorlen
    Commented Jul 17, 2023 at 20:21
  • 113
    I wouldn't hold my breath with bringing community and AI together. It is pretty clear community does not want Gen AI here. Commented Jul 17, 2023 at 20:22
  • 56
    Will this be an event for users to earn the Not a Robot badge on Stack Overflow? Commented Jul 17, 2023 at 20:28
  • 1
    I tuned in with the assumption that the first thing I'll see in responses is tearing apart the staff. Assumption met. Commented Jul 17, 2023 at 21:01
  • 6
    You know, posting these tone-deaf announcements one after another and pretending nothing is happening seems pretty disrespectful to the community. I wonder if this crosses the point where this post qualifies for a ‘rude or abusive’ flag. Commented Jul 17, 2023 at 21:54
  • 5
    There as some posts related to "how Stack Overflow is evolving to bring community and AI together" like this answer and the follow up questions What does Prashanth mean by "Community" in his blogposts, and how do the mods influence his perception of it?, Is there any publicly available record since 2022 of Prashanth chatting with Meta Stack Exchange core community members?. I wonder if we could learn what Prashanth thinks about mod / Meta SE members nowadays.
    – Rubén
    Commented Jul 17, 2023 at 22:20
  • 135
    Just a note to @Rosie, the down votes here are for the message from Prashanth, not the messenger.
    – Chindraba
    Commented Jul 17, 2023 at 23:21

12 Answers 12

211

We’ll be participating in several different sessions where we’ll share the work we’re doing to help developers with the challenges they’re facing in 2023 and beyond.

The biggest challenge I (and many other developers) are facing this year is identifying unmoderated, unmarked AI sand and separating it from the pearls of hard work produced by actual experts. What session should I (we?) be listening to?

3
  • 23
    I'm surprised you didn't get any response from staff. Or not. And that's why I never bother to waste time giving feedback. Commented Jul 28, 2023 at 14:53
  • 2
    @ShadowWizardStrikesBack This isn't even just feedback. It's a genuine question about which talk will be addressing this issue, or if it will be addressed at all. They could've just replied with "none" or "the main one". Commented Jul 30, 2023 at 11:41
  • 2
    @Andreas they can do many things, they choose not to. They ignore us, and expect us to actually waste time in their events, in promoting their AI, and more. (And no, a dry post "we listen to your feedback" once every few months is pathetic, it's not showing anything.) Commented Jul 30, 2023 at 12:05
153

Stack Exchange and Stack Overflow are places for humans to find answers that are written and curated by experts in a variety of fields.

GenAI is incapable of the sort of understanding that makes someone an expert. Developers and other experts understand that if they have an expert-level question, they need an expert-level answer from someone who has dealt with the same sort of problems. GenAI can only simulate and pretend being an expert, selling snake oil promises of productivity while the output would be best at home with the output of a "cleanse". Occasionally correct extrapolation from a training set of actual expert answers is a curiosity, not a replacement for expert understanding.

I hope Prashanth Chandrasekar will promote Stack Exchange as the place for people to go for human expert answers to the questions they have, a beacon in a space increasingly saturated with embarrassing GenAI shortcuts that mislead and misdirect and waste productive time. When team members need expert human answers that can only come from their present and past coworkers, they can turn from the public Q&A sites to the familiar interface of Stack Overflow for Teams that their company uses to collect and maintain human expertise.

If you want to instead struggle with a solution that ChatGPT hallucinated that doesn't actually solve your problem, that's okay. If you want stuff like this, ask ChatGPT. If you want to annoy your coauthors by rewriting their work and then insist you're "helping" while they're stuck finding all the introduced errors, ChatGPT can do that. If you want to risk losing your professional license to practice law, GenAI can save you some time. If you want to see multiple human experts offer various solutions to a problem, come here instead.

Stack Overflow developers have been learning what others know, too: this technology isn't ready to replace humans. If we're going to use AI, let's limit ourselves to narrow, well-tested, purpose-built solutions. Let humans handle the rest.

Developers and other professionals around the world have trusted the Stack Exchange network for the questions that their jobs depend on. Let's not forget who built the content that they appreciate. While everyone else chases the shiny new toy, let's recommit to what works. It's not innovative to chase the crowd.

16
  • 1
    And of course AI as it stands can only answer with knowledge it gained up to training. Anything since, it won't have knowledge of. Thus, e.g. Stack Overflow becomes useless, given ChatGPT was trained on its current knowledge database. New questions, it won't be able to answer (just more debugging, "how to" duplicates etc, not about android 7000)
    – Adriaan
    Commented Jul 18, 2023 at 14:34
  • 3
    @Adriaan, But it doesn't 'gain knowledge' does it? Only living things do that. And we barely, and rather incompletely, understand how that works. These are very particular algorithms are suited for very particular tasks of sorting information and attempting to present it. None of them 'know' a thing about the output they are spitting out.
    – ouflak
    Commented Jul 18, 2023 at 14:57
  • 5
    something that doesn't exist can't gain knowledge.
    – Kevin B
    Commented Jul 18, 2023 at 15:02
  • Yeah, but it doesn't bring in money for shareholders. StackOverflow is a company, and that means one thing matters: money for shareholders. Clickbait brings in more money than nuanced content. See also Reddit, which is throwing away all its core contributors and replacing them with a feed of AI-generated memes. Commented Jul 24, 2023 at 12:38
  • @user253751 They aren't selling views on the public platform, though, they're selling private Teams instances as SaaS. An AI-repost site isn't going to sell corporate subscriptions. A site that is industry-famous for content that helps people do their jobs is. Commented Jul 24, 2023 at 13:27
  • 1
    @BryanKrauseisonstrike I thought Teams was winding down, so they could focus on generative AI. Turns out Q&A is less useful when you artificially limit who can A. Commented Jul 24, 2023 at 13:56
  • 3
    @user253751 I don't know why you would think that. Teams is their primary business. I think Q&A is far more useful when you limit it to humans, and I don't think that's an artificial limit. For Teams, there is of course further limit to who can answer by making an instance private, but this is also the whole point of the product, because it allows for internal questions that include privileged information, company-specific workflows, etc. Commented Jul 24, 2023 at 14:01
  • 1
    @ouflak There is no reason to think AI knowledge is categorically different from my knowledge or yours. Commented Jul 27, 2023 at 16:04
  • 7
    @user253751 There are lots of reasons to think AI knowledge is categorically different from my knowledge or yours. First of all, "AI" is not one thing, but a collection of things, and the vast majority are not at all similar to how biological brains store information. For LLMs in particular, they are instances of a specific statistical model that is not all that similar to human knowledge, and there is no underlying understanding of concepts. Maybe someday an AI will exist that has knowledge not categorically different from humans, but we are not there. Commented Jul 27, 2023 at 16:17
  • 1
    @BryanKrauseisonstrike "biological brains" is also not one thing, but a collection of things. If we upload your brain to a brain simulator it will still be your knowledge that is in the simulator, so the fact that it's digital is of no difference. Birds can know things and bird brains are not like human brains. In the future we will probably discover that human brains are just statistical models. Commented Jul 27, 2023 at 16:23
  • 7
    @user253751 Like I said, an AI may exist in the future that is more comparable to the ways we learn and store knowledge. Current GenAI models are not, so don't pretend like they are. Commented Jul 27, 2023 at 16:30
  • 1
    The MAJORITY of SE content is not experts asking expert-level questions and seeking expert-level answers, but beginners asking beginner-level questions and needing beginner-level answers. Stack Overflow's shareholders apparently know this, which is why they are shutting down the site and replacing it with a ChatGPT clone. Commented Jul 28, 2023 at 20:01
  • 1
    @user253751 I think the most typical user of SO is a professional programmer who has not managed to memorize every single bit of syntax from every single programming language out there, because that's not realistic. As you filter into the network sites, I do think some of them (including ones I answer frequently on) are mostly "novices/students asking experts", but still, students are in need of expert answers, too. They are the most vulnerable, in fact, to being mislead by plausible-sounding answers that aren't actually correct, which current GenAI commonly generates. Commented Jul 29, 2023 at 3:49
  • 2
    @user253751 Must be different now 45 minutes later because it looks just like I expect it to. Also remember that most people coming here aren't posting a question, they're finding one that's already posted and answered. Commented Jul 29, 2023 at 13:36
  • 2
    @BryanKrauseisonstrike social media CEOs are all optimizing for "engagement", which apparently means people coming and typing things on their site, not finding search results Commented Jul 29, 2023 at 13:55
122
  • Thursday, July 27: Prashanth Chandrasekar, CEO of Stack Overflow, will be on the main keynote stage to share how Stack Overflow is evolving to bring community and AI together.
  • Friday, July 28: Join us for a workshop session where we’ll talk about how the developer community can play a crucial role in how AI accelerates and helps with the accuracy and quality coming out of future GenAI solutions.

This only illustrates how out of touch SE management is with the community.

AI is currently the most controversial issue on (tech) SE sites. Its use in questions and answer is banned, as far as I know, SE-wide*, and to top it off, there's an unprecedented moderation strike going on about that very subject, at this very moment.

This continuous focus on how GenAI is "the next big thing" needs to end. I think I speak for a large part of the SE community when I say we're getting rather tired of being forced onto the GenAI hype train.

* SE-wide with a few exceptions.

0
101

Join us for a workshop session where we’ll talk about how the developer community can play a crucial role in how AI accelerates and helps with the accuracy and quality coming out of future GenAI solutions.

This conference is an ocean away, so there's no way I'll be there, but I'm interested in how the company can be an authority on "how the developer community can play a crucial role in how AI accelerates and helps with the accuracy and quality coming out of future GenAI solutions" after repeated demonstrations of not being able to assess appropriate applications for GenAI (or other types of algorithmic solutions) and not listening to what the broader developer (and non-developer) communities are saying about this topic.

4
  • 3
    There's a 99 euros online pass, I might just get one to listen to whatever Prashanth has to say about the "buy in" of the community in this GenAI hype train. I hope they'll have a Q&A after both talks, though I assume it will be heavily moderated.
    – Laf
    Commented Jul 18, 2023 at 15:21
  • @Laf You don't need to: it will be recorded.
    – wizzwizz4
    Commented Jul 19, 2023 at 14:24
  • 2
    @wizzwizz4 My post is about the workshop. It does not appear that will be recorded or otherwise available. Commented Jul 19, 2023 at 22:32
  • 2
    Oh. In which case, @Laf: I know you wouldn't record a video, but I'd appreciate some journalism, perhaps with some quotes here and there.
    – wizzwizz4
    Commented Jul 19, 2023 at 22:39
78

... Stack Overflow is evolving to bring community and AI together.

I... I don't know what to say.

He’ll talk about how combining the collective knowledge of the community with technologies that increase productivity, efficiency, and rapid learning, can help technologists get to solutions faster.

By 'technologies', do you mean something like the rather unsuccessful formatting assistant? You can talk all you want, but you need to show some concrete examples of AI in action on Stack Overflow, and I don't think that any of your products so far have been helpful or beneficial to the community. You wouldn't lie about our feedback, would you?

Friday, July 28: Join us for a workshop session where we’ll talk about how the developer community can play a crucial role in how AI accelerates and helps with the accuracy and quality coming out of future GenAI solutions.

Sure, I mean AI is built on and trained by developers. But the way you are going about integrating AI into the Stack Exchange network is not going to generate accuracy or quality.

By combining the collective knowledge and judgement of a community of millions of developers with technologies that increase productivity, efficiency and rapid learning, everyone can get to solutions faster.

(From program website)

You have listened to neither our knowledge nor judgement.

I am having a bad feeling about this; if someone actually checks the community's responses to the AI updates after the conference, they would be in for quite a surprise.

4
  • 5
    Perhaps it's worth noting the drop in quality when AI is trained on AI output (e.g. this paper). Commented Jul 18, 2023 at 1:37
  • 3
    @RebeccaJ.Stones they've also written a blog post about LLM training.
    – bobeyt6
    Commented Jul 18, 2023 at 1:38
  • 2
    @RebeccaJ.Stones so they underpaid humans to source training data and the underpaid humans just copy-pasted from AI. That's hilarious. Commented Jul 25, 2023 at 5:50
  • 1
    Here's what to say: nothing. Stack Overflow was bought by AI fanatics (for about 2 billion dollars) and we live within an economic system where buying a company gives you the right to make it do what you want, ignoring all feedback to the contrary. Feedback will not achieve anything unless it comes in the form of $$$. Maybe this out-of-touch-CEO-owner bubble will be the next to pop. Commented Jul 27, 2023 at 16:06
45

At the moment, I'm vaguely underwhelmed. I'm usually hopeful for brighter futures, but it seems that the Gen AI push has mostly brought suffering (including to former staff), and looking at some of the deleted posts, anger.

It’s a world gone a little mad, with the seductive whispers of the sibyls of Gen AI taken to be the truth. I'm personally less than impressed by what I've seen, and frankly there's a little terror over the next harebrained scheme that'll be collectively pushed upon us. It seems so many smaller organisations are rushing headlong into Generative AI projects that are half baked - MDN recently published a post mortem on a new AI product that went poorly

There probably are certain problems that could be solved but from what I've seen here and on the blogs, I'm unsure that anything that could be announced could have value.

I'm always happy to be wrong. But in general I don't see anything past a machine that speaks in coherent nonsense, folks who're moving from the last grift (cryptocurrency/web 3.0) to a new one, and a good chunk of the tech community switching from one set of cultist robes to another in worship of their new gods.

Outside the literal problem of 'natural language input generating 'plausible' responses' - occasionally in a convincing but terrible way, nearly resulting in regicide, occasional potential copyright infringement, or just claiming everything their search engine bot scrapes - I've... not seen much value in it. Even Copilot essentially is just 'we're taking code that might work', and using it to do smarter intellisense.

As for it'll help developers - well, there's CEOs talking about eliminating almost every traditional coding job in India.

It's a problem looking for a solution that causes more problems.

It's not going to help sort out the current moderator strike, and its aftermath, nor the workload of the community team, nor fix the backlog of features that could help communities thrive.

So - I'm not hopeful.

We'll see though. Once again, sometimes I'm happy to be wrong.

4
  • 1
    Along the same lines, another problem with GenAI is the unpredictable "drifting". For a recent example, Over just a few months, ChatGPT went from correctly answering a simple math problem 98% of the time to just 2%, study finds - the original article is paywalled, but is also available here.
    – dxiv
    Commented Jul 20, 2023 at 18:29
  • @dxiv I would suspect GPT was guessing in both cases and the guess just happened to shift over time, so not really an indication of anything. I don't see how a GPT model could determine whether an arbitrary number is prime other than just memorizing it and memorizing common patterns (ends in an odd digit other than 5) Commented Jul 25, 2023 at 5:51
  • @user253751 Who knows? There was this back in March: ChatGPT Gets Its “Wolfram Superpowers”! But that's precisely the point - we don't know, and we shouldn't blindly trust what we don't know. It will be the case someday that an AI may be able to give meaningful answers in general, and math answers in particular, while also being able to explain the step-by-step train of thought. But we are not there, yet, not by a very long shot.
    – dxiv
    Commented Jul 25, 2023 at 6:20
  • 4
    @dxiv an addition: GPT is can often recite algorithms from memory and then execute the steps it just recited, which is pretty cool, but there's no way for it to execute an algorithm without reciting it. If they manage to get that Wolfram connection to work well, it will greatly expand GPT's capabilities, although it will still be a bullshit generator at its core. Commented Jul 25, 2023 at 6:22
26

Will a recording or livestream of Prashanth Chandrasekar's keynote speech be available to watch? From the website this is the talk abstract:

Stack Overflow: Community and AI
Prashanth Chandrasekar, CEO of Stack Overflow, will share how Stack Overflow is evolving to bring community and AI together. By combining the collective knowledge and judgement of a community of millions of developers with technologies that increase productivity, efficiency and rapid learning, everyone can get to solutions faster.

The challenges developers faced back in 2008 are not the ones they are facing in 2023 and beyond. Stack Overflow believes the developer community can play a crucial role in how AI accelerates, increasing the accuracy and quality in GenAI solutions.

Yes, it's on YouTube.

Oh, and if possible, this one too:

Stack Overflow: Past, Present & Future
Stack Overflow’s current CEO Prashanth Chandrasekar and its former CEO and cofounder Joel Spolsky share stories about the evolution of Stack Overflow over the last 15 years and what its future holds.

4
  • 4
    That conference's presentations are usually posted on YouTube. Sample. Spolsky's starts at 05 h 30 min 57 secs. Commented Jul 18, 2023 at 8:13
  • 9
    The keynote speech will be recorded.
    – Rosie StaffMod
    Commented Jul 18, 2023 at 20:21
  • 1
    @This_is_NOT_a_forum here is a timestamped link: youtube.com/live/NFqI2FLMwOE?feature=share&t=19765
    – M--
    Commented Jul 19, 2023 at 16:56
  • 1
    Fun fact: Joel is on the fediverse, and he sometimes complains about the state of Stack Overflow after he sold it. Commented Jul 27, 2023 at 16:08
18

Thursday, July 27: Prashanth will be joined by Stack Overflow’s former CEO and cofounder Joel Spolsky for a fireside chat to share stories about the evolution of Stack Overflow over the last 15 years and what its future holds.

This might be interesting.

That last (i.e. most recent) thing I heard Joel say, 4 years ago ...

... was more-or-less that the public sites would continue pro-bono, and that the for-profit company would make its revenue from selling Teams.

3
  • +1 to this. Is StackOverflow not profitable enough, to the point that they feel the need to chase trends in order to draw more users? Commented Jul 25, 2023 at 9:17
  • A former CEO cannot bind future CEOs like that. Commented Jul 27, 2023 at 16:08
  • Circumstances change but I assume that was the view at the time, the basis on which the company was evaluated.
    – ChrisW
    Commented Jul 27, 2023 at 17:47
13

I am interested to know if there will be Q&A after any of these talks. I can think of number of challenging and crucial questions that can be asked on any of these sessions, some of which are asked here.

4
  • 2
    Maybe, but probably only available to attendees.
    – bobeyt6
    Commented Jul 19, 2023 at 19:18
  • 2
    @bobeyt6isstricken I am talking to folks on discord to see if anyone is willing to attend, and possibly we could be covering some of the expenses.
    – M--
    Commented Jul 19, 2023 at 19:19
  • It was available in this instance (2019), e.g., from 5 h 10 min 31 secs. Commented Jul 21, 2023 at 13:54
  • 2
    @This_is_NOT_a_forum if questions are "moderated" like this, I'd say it doesn't worth a penny to go, let alone £500+.
    – M--
    Commented Jul 21, 2023 at 14:03
11

I might be a bit late to make any changes, but I'd like to offer some of my thoughts on the events you have lined up:

Thursday, July 27: Prashanth Chandrasekar, CEO of Stack Overflow, will be on the main keynote stage to share how Stack Overflow is evolving to bring community and AI together. He’ll talk about how combining the collective knowledge of the community with technologies that increase productivity, efficiency, and rapid learning, can help technologists get to solutions faster.

Other answers have already covered the immense hypocrisy here. How can you even talk about AI bringing people together when your community is actively protesting against you for it? If you don't do something about it AI is going to rip your community apart, not build it.

After seeing the overview, I generally like the idea of improving the broken search and making a dedicated GenAI community. However, I think there are some fatal flaws here that only time will validate. First, no one is ever going to seriously use your GenAI-powered search. If I want answers, then I'm going to use Google, not one of the thousand ChatGPT rip-offs. Google has already solved your search problem; there's no need to reinvent the wheel.

On the other hand, I'm all for this new GenAI community. Honestly, it's a lovely idea and I have no criticisms of it. For once, we will finally have a dumping place for all the stupid, unanswerable, and off-topic questions on Stack Overflow! What could be better? (if you haven't noticed, this is sarcasm).

And finally, don't get me started on the new NLP collective. It's already been established that collectives are just not that great of an idea. But hey, if Stack Overflow gets their AI and I get one unified list of tags to ignore, then sure.

Thursday, July 27: Prashanth will be joined by Stack Overflow’s former CEO and cofounder Joel Spolsky for a fireside chat to share stories about the evolution of Stack Overflow over the last 15 years and what its future holds.

Considering the original fireside chats were given during one of the worst times (economically at least) for the public, I'd say this conversation is probably worthy of the title. However, I suspect that will be where the similarities end. In a midst of a moderator strike due to idiotic, unexplained policies we would rather have a real fireside chat explaining how you're going to work with the community. I beg you at the very least to acknowledge the strike, but my hopes are low.

I don't know about the rest of the community here, but I don't want to hear about the evolution of Stack Overflow (if I did, I'd just use Google). I want to hear about the future of Stack Overflow and how you're going to fix it.

Friday, July 28: Join us for a workshop session where we’ll talk about how the developer community can play a crucial role in how AI accelerates and helps with the accuracy and quality coming out of future GenAI solutions.

Crucial role in accelerating AI? Oh, you mean letting Stack Overflow sell our answers to GenAI data mining operations against the license we agreed to and without credit or compensation.

3
  • 5
    Re "Google has already solved your search problem": No, search is broken. For example, it will not return the canonical questions (with high-quality answers) from 2008 and 2009 (for whatever reason). Instead, it returns low-quality low-scored duplicate questions with subpar answers. As there is practically no effort to locate duplicates, search engine visitors are left stranded on those question islands (and the "Related" column is usually completely useless). Commented Jul 28, 2023 at 10:09
  • 1
    cont' - And yes, search engine results vary, and it is always possible to cherry-pick examples where it works. Commented Jul 28, 2023 at 10:14
  • 1
    Maybe "fireside" here stands for "the side that does the firing"?
    – tripleee
    Commented Aug 6, 2023 at 9:24
7

SE.inc's plans are self-contradicting

As Prashanth mentioned, our goal is to get technologists trusted solutions faster, and we’re leveraging AI and GenAI to facilitate this. We shared insight into:

  • How we want to make it possible for users to be able to receive instant, trustworthy and accurate solutions to problems using search and question-asking powered by Generative AI. Learn more about these updates here.
  • A dedicated GenAI Stack Exchange will serve as a place for a community that is centered around knowledge sharing for writing prompts for GenAI tools.

--- Rosie's summary of the Event and CEO Prashanth's talking points

With the second bullet point, the CEO is admitting that GenAI can be difficult to use, and people may need help building up the expertise to use it effectively - otherwise, having a Q&A site about writing prompts is nonsensical.

With the first bullet point, the CEO says that "instant, trustworthy, and accurate solutions" will be there for everyone.

These two concepts are mutually exclusive - either anyone needing help can just ask an AI and get a good, substantive answer, or writing prompts to get reliable information from an AI requires expertise the likes of which can be built by asking questions and getting answers to them.

How, exactly, does SE.inc reconcile these contradictory concepts?

-1

We Are Developers Event Summary

Today, Prashanth Chandrasekar, CEO of Stack Overflow, showcased several early previews of features of our upcoming AI offerings from the stage of WeAreDevelopers.

As Prashanth mentioned, our goal is to get technologists trusted solutions faster, and we’re leveraging AI and GenAI to facilitate this. We shared insight into:

  • How we want to make it possible for users to be able to receive instant, trustworthy and accurate solutions to problems using search and question-asking powered by Generative AI. Learn more about these updates here.

  • A dedicated GenAI Stack Exchange will serve as a place for a community that is centered around knowledge sharing for writing prompts for GenAI tools.

  • Stack Overflow’s Natural Language Processing (NLP) Collective will include a new feature called Discussions that will provide a focused space to debate technical approaches, explore implementation strategies, and share different perspectives so that members can make more informed technical decisions.

You can see a video from the event, read more about all our announcements for both Stack Overflow and Stack Overflow for Teams, and see early preview demo videos over at Stack Overflow Labs, our central hub to share experiments, demos, insights, and news - across all Stack Overflow products.

For each feature we announced today, you can register your interest to keep up with what we’re building or to let us know that you are interested in being part of an alpha or beta test.

17
  • 28
    The vision of "get technologists trusted solutions faster" is so disappointing. Y'all have a network of almost 200 international communities centered around sharing knowledge, and you want to turn it into just another search engine that pukes out robotic responses to whatever question gets dumped into it. Where do y'all think the knowledge that the robot is responding with is going to come from exactly?
    – ColleenV
    Commented Jul 27, 2023 at 12:42
  • 7
    @ColleenV The updates to the search engine that the team is looking into is about surfacing answers that have been written by community members. The summary comes from the specific community members per the sources listed below the ai-summarized answer. Sources are included to provide credit and attribution to the original post authors.
    – Rosie StaffMod
    Commented Jul 27, 2023 at 12:59
  • 4
    Thanks for clarifying, but my concern is a bit more abstract than that. Is that the vision for the future of Stack Exchange? To provide fast trusted solutions to technologists? Or did I misread and that's just for the AI offerings?
    – ColleenV
    Commented Jul 27, 2023 at 13:07
  • 1
    I have to say that my initial reaction to this is joy. Assuming the quality of AI content is good and that it works fine, this actually works well for both the SE communities and the page hits (profits) the companies are after. Ideally, the communities will be left alone to create meaningful content. Those who want to engage will engage. The rest will have a quick answer to their problem. Commented Jul 27, 2023 at 13:14
  • So these new updates to the search bar are like a Bing AI search restricted to SO?
    – bobeyt6
    Commented Jul 27, 2023 at 13:16
  • 6
    @ColleenV Stack Exchange doesn't exist without the community. What was presented today was a preview of what the team is working on. Features aren't set in stone and for the public platform features specifically, we need community input and feedback. It's why the CTA to my answer is to go to Stack Overflow Labs to indicate interest in alphas and betas. We'll also be coming to Meta to ask for feedback. Philippe held a couple of focus groups with moderators prior to today to get feedback that's already been shared with developers. We intend to do more focus groups in the future.
    – Rosie StaffMod
    Commented Jul 27, 2023 at 13:21
  • @Rosie When i registered interest a few weeks ago, I started getting SOfT spam based on 2021 data. Is there another way to register interest that isn't just another marketing email? I've already unsubbed, but certainly still want to know about alphas and betas related to the public QA product.
    – Kevin B
    Commented Jul 27, 2023 at 14:35
  • 1
    @KevinB thank you for flagging. It looks like there was an error and a fix was just pushed. Extra safety guards have been put in place. If you continue to receive marketing emails let us know.
    – Rosie StaffMod
    Commented Jul 27, 2023 at 15:37
  • 1
    @ColleenV The company vision remains unchanged. What Prashanth was speaking about in regards to "get technologists trusted solutions faster" is a goal around the initiatives shared today.
    – Rosie StaffMod
    Commented Jul 27, 2023 at 15:57
  • 3
    I currently use Google Search to get my trusted solutions. Is the SO-AI faster than this and how have you benchmarked this?
    – MT1
    Commented Jul 27, 2023 at 16:02
  • @ColleenV In most instances, corporate-speak means nothing except for the most obvious parts (the Stack Overflow brand is being reused for ChatGPT) and the rest should be disregarded altogether. Commented Jul 27, 2023 at 17:02
  • 3
    I'm quite disappointed, all this talk about AI and not a single mention of the strike that is happening because of it. But at least I'm not surprised.
    – bobeyt6
    Commented Jul 27, 2023 at 22:30
  • 1
    Why "Labs" instead of just continuing to use MSO and MSE? Because on Labs you can put out marketing fluff without receiving critical feedback that looks bad in the eyes of investors/buyers? Commented Jul 29, 2023 at 22:28
  • 1
    @AndreasdetestsAIhype More charitably: because the (relatively) new division of the company wanted their own shiny website to show off their shiny things, to make it clear that their work was distinct from Q&A / the Stack Exchange Network proper.
    – wizzwizz4
    Commented Jul 30, 2023 at 21:33
  • 2
    By the way, were the other events recorded?
    – bobeyt6
    Commented Jul 30, 2023 at 22:26

You must log in to answer this question.

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