12:41 AM
And that's always going to be an advantage the 'human' web, places like this will always have. As much as people hype generative AI, it lacks nuance, experience and the ability to 'fit' things together as a whole. — Journeyman Geek 8 secs ago
1:03 AM
This post doesn't pass the bar for the "Needs Details or Clarity" close reason. — Robotnik 18 secs ago
1:27 AM
Do I get an upvote for my answers every time that an AI uses it? This is what we encourage from the (human) community, and I'd accept that if the AIs played along too. Otherwise, this becomes like Reddit, and all the content contributed by users is just feeding a nasty business model. — mike 23 secs ago
3 hours later…
4:10 AM
@super-starball-ultra, people don't always explain things exactly as they are, they explain how they think they are. Maybe they should test this stuff won't break since it's close to what they claim to be changing. (1) the images in the examples are on imgur now. (2) S.O. is changing some imgur urls to non-imgur url. (3) if the new server doesn't provide the same or compatible CORS permissions than any URLs changed in these samples will break them. — gman 21 secs ago
@Will no, LLMs, by definition, do not "recite". They could luck into a exact replica text, but there's absolutely no form of "intent" to duplicate exact text. But again, you're nitpicking terminology. — Cerbrus 47 secs ago
2 hours later…
5:56 AM
I still want an option to remove the ad for Indeed on the sidebar. It’s appearing on every community, it just limited, to Stack Overflow. — Ramhound 31 secs ago
How do I remove the link to this MVP to Indeed. I will not click on the link on purpose as I feel Indeed is an evil predatory company and in my experience has never actually prompted a real job. Please add an option to our settings to remove the link from the sidebar. — Ramhound 18 secs ago
6:16 AM
The only thing sad is your would delete your answers instead of just simply not contributing anymore. — Ramhound 18 secs ago
Removing help contributions once they have been submitted is a form of vandalism, if you don’t agree with the license to the contribution you are submitting, simply don’t submit it in the first place. — Ramhound 25 secs ago
In my opinion there was no way to message this announcement where individuals who have contributed many years to a community, would have taken it positively, since the announcement is primarily negative in if you have contributed to community significantly. — Ramhound 34 secs ago
"I know of a user whose account has been suspended for 7 days because they started deleting their stuff." for reference, the suspension has no relation to the OpenAI deal. It's standard practice and has been enacted every time an account has tried to mass remove their content. There is monitoring in place for such occasions for many years now. Only the deletion in this case is related to the OpenAI deal. The presentation here may make somebody believe that the suspension is somehow "abnormal" rather than a standard. — VLAZ 54 secs ago
6:44 AM
when you participate on this site and contribute content, you are agreeing to the terms of service, which includes giving SO Inc. the license to commercially exploit your content. you may not like it (it doesn't fill me with a warm and fuzzy feeling), but it's a fact. — super-starball-ultra 50 secs ago
Note that users trying to remove and/or deface their contributions and being suspended over it has been happening for years. There is monitoring in place for just such occasions. It all follows standard policy, as well. Neither what you tried to do, nor the result is anything new for the platform. — VLAZ 27 secs ago
@zcoop98 All the more reason to delete them now before their next announcement along the lines of "Also, we fed all your work into the AI thirty-five minutes ago" — Shadur-don't-feed-the-AI 1 min ago
@ramhound I submitted those contributions well before this "AI" LLM bubble was ever a thing. If they decide to retroactively alter the terms of the agreement, they don't get to be upset when I decide to retroactively withdraw my agreement to it. — Shadur-don't-feed-the-AI 9 secs ago
This vaccilating non-answer would be deleted by the mods instantly if it were from anyone but staff. I hope you're ashamed of yourself, but I'm guessing that the prospect of the money you're hoping to get from this AI deal is smothering your conscience. — Shadur-don't-feed-the-AI 18 secs ago
@Shadur-don't-feed-the-AI - Your agreement cannot be revoked, you might take a moment, to read the agreement you are trying to unilaterally revoke. Doesn’t change the fact, if you vandalize contributions, the community will just reverse your vandalism. — Ramhound 11 secs ago
@ramhound You do realize that "I am altering the deal. Pray I don't alter it any further" was a line uttered by someone who was unequivocally the bad guy when he said it, right? Do yourself a favour and stop banging the "this is going to be done, just lie back and enjoy it" rhetoric. — Shadur-don't-feed-the-AI 48 secs ago
@Shadur-don't-feed-the-AI - I most certainly am not happy with the deal, my 1500 answers will forever exist, if future answers will be submitted has yet to be determined. However, if users vandalize their contributions, I won’t hesitate to flag that behavior. — Ramhound 55 secs ago
7:30 AM
See here how to have your answers deleted for you: meta.stackexchange.com/a/399735/297249 — henning no longer feeds AI 50 secs ago
@AMtwo license or not, the GDPR gives you a right to have your answers deleted. See here how to have your answers deleted for you: meta.stackexchange.com/a/399735/297249 — henning no longer feeds AI 54 secs ago
license or not, the GDPR gives you a right to have your answers deleted. See here how to have your answers deleted for you: meta.stackexchange.com/a/399735/297249 — henning no longer feeds AI 1 min ago
That will only delete your PII, not the content you posted because you agreed to the ToS and you gave SE an irrevocable license to that content. You'll be forgotten, your content will not. — rene 22 secs ago
What do you mean with "reporst the work" on Codidact? You mean like copying our answers/questions there? — SilentCloud 55 secs ago
File size, but the bot owner already handled it by now, so guess all is good. Thanks! — Shadow Wizard Love Zelda 9 secs ago
8:24 AM
You could, of course, just also take note of the fact that pretty much the entire community is extremely displeased with the openAI announcement and make it clear to staff that they really need to rethink their stance posthaste before more people leave. — Shadur-don't-feed-the-AI 55 secs ago
8:57 AM
Hi 6nagi9, can you please provide the question you want to ask? See the [site-recommendation] tag description: "Give as much detail as possible about the question you'd like to ask". — Joachim 37 secs ago
That's a bit short-sighted once you write a question or an answer on an SE site, it's stored, if you delete or vandalise the post it still exists on SE infrastructure. You can't control that. — user692942 30 secs ago
@Catija currently this text comes from Transifex. It always easier to update translations this way than ask staff (mostly not familiar with Russian language) to do something with customization. — αλεχολυτ 29 secs ago
You know the datadumps exist, right? Much easier to use that instead of scraping. — rene 46 secs ago
@Cerbrus you're the one nitpicking terminology, trying to police people using a perfectly ordinary meaning for the concept of a "copy". When people talk about a mechanism like an LLM producing copies, there is no reason to assume they mean that the mechanism has some sort of "intent" or any other kind of understanding of the concept of copying. They just recognise the presence of design characteristics that cause copies systematically to be made, regardless of whether the system designers intended that behaviour in particular. — Will 14 secs ago
@bta That's interesting, I didn't know about the arbitration thing. It seems a very flimsy basis to commit such a massive copyright violation (considering all the exceptions), but I suppose it's hard to put anything past them at this point. — Peter 33 secs ago
9:47 AM
the potential legal issue I pointed out is not a license violation but a gdpr violation. Rather than losing the right to host content, stackoverflow only faces fines of up to 10% of annual revenue. — user1937198 53 secs ago
Questions and answers posted to the SE network are not personal information. Please try to stick to the actual law when you want to propose using it. More information en.wikipedia.org/wiki/Right_to_be_forgotten "The right to be forgotten (RTBF[1]) is the right to have private information about a person be removed from Internet searches and other directories in some circumstances" — VLAZ 10 secs ago
I would be extremely surprised if there were no SE data in the GPT training datasets, eg en.wikipedia.org/wiki/Common_Crawl — PM 2Ring 32 secs ago
10:25 AM
In any case, I would advise folks savvy on writing SEDE queries to prepare some for detecting patterns of questions deleted by protest. — E_net4 33 secs ago
I expect a lot of people want the answer to this question more than most other questions, but it depends how they implement it and what constraints are imposed on them by their contract-partners (and if they pay attention to the community's wishes). They've not told us about that yet. — W.O. 44 secs ago
11:11 AM
Interesting, I haven't used Gemini before and it didn't create something like that for me. It doesn't look entirely like attribution to me, more like a combination of search and genAI results. But in general the question how to attribute seems almost impossible to me, many genAI answers will not be clearly based on a single source and I don't see how attribution could work there. — Mad Scientist 6 secs ago
Hi michael chikoba, welcome to the Stack Exchange Network Meta site! I'm not sure which search brought you here but the problem you describe will not be answered on this specific site. To get an expert's answer for the topic of your question you'll have to find and then re-post on the proper site. Check How do I ask a good question and What is on topic on the target site to make sure your post is in good shape. Your question is definitely off-topic on Meta and is better deleted here. — rene 45 secs ago
@MadScientist In Gemini, you need to click a button for it to highlight content. It will either cite its sources or say no supporting sources can be found. I'd agree that it's not attribution, but I also don't think how CC defines attribution is appropriate, either, since it gives so many outs. But based on how LLMs work, I don't think you can get better than this. The best you can do is trace statements to similar training material. It would be better to link to the Top N supporting sources instead of just one, but it's still a guess since all of the training material impacts the next output. — Thomas Owens 42 secs ago
11:33 AM
@Lundin Automatically post AI-generated answers, and exclude them from the data collected by OpenAI, maybe? — Andreas moved to Codidact 16 secs ago
@Joachim I have already asked the question in GenAI. Do you still need me to update my query above with the question I want to ask? — 6nagi9 17 secs ago
11:50 AM
unless the relevant parts have been cut-off from your picture, in my opinion that is NOT an example of "Gemini providing attribution". Gemini did two things: first, it used Google Search to provide a link to a only tangentially related SO question (notice that it clearly stated "Google search found", NOT "the following Gemini answer was based on"). Second, it used Gemini LLM to create a statistical word chain based on the question. And if you notice, the content it generated is not found in the SO question in the first place. — SPArcheon 41 secs ago
@SPArcheon You are confusing "citation" and "attribution". Citation is an ethical concern to trace a potential source of an idea. Attribution is a legal concern regarding identifying content used, adapted, remixed, transformed, etc. The examples clearly show Gemini citing sources. I don't see any content sufficiently identical to warrant attribution, so citation is the proper thing to do here. However, even if the content was used in a manner that required attribution, what Google is doing does meet the requirements of the CC Attribution clauses. — Thomas Owens 14 secs ago
@SPArcheon Which point did I miss? I've also elaborated in my answer on citation vs attribution. — Thomas Owens 59 secs ago
that even if you want to differentiate between citation and attribution (I don't think it really matters on this topic, but you are correct in stating them being sightly different concepts) what Gemini did there by linking an unrelated question that doesn't anywhere express the concept found in the LLM generated answer is neither Citation nor Attribution. — SPArcheon 17 secs ago
“The email support of Stack Exchange was very friendly.” So, what is the reply given from the email. — user1176409 30 secs ago
Welcome Attila. As this relates to a single site on the network, it should be posted on that site's meta rather than here on main meta. See math.meta.stackexchange.com. — W.O. 53 secs ago
12:29 PM
@SPArcheon You're referring to the technical correctness of the citation. I looked at both examples again, and I don't see issues. The Gemini example that cites the SO question does have an answer that claims that one return comes from structured programming. The SGE example points to a Software Engineering question that has several answers that refer to readability, understandability, and control flow. I don't agree that the answer doesn't include the concepts expressed in the linked sources. — Thomas Owens just now
That AI is making up a somewhat plausible citation is actually worse than none at all. I am highly skeptical that the search results it provided have anything to do with how the algorithm arrived at its answer. Some of them look like AI generated SEO content, but since you only provided a screenshot I can barely read, it's difficult to tell. — ColleenV 21 secs ago
I'll agree to disagree. If I answer with some quotes from Gandalf, imho linking a page that links to another page about Gandalf is not "citing" your sources. In the same way if I am talking about the Agile Manifesto I will cite the official page, not an SO question that has a link to the Manifesto page. But you are free to think the opposite. — SPArcheon 9 secs ago
A related answer from a while back: meta.stackexchange.com/a/388569/273494 genAI is perfectly capable of providing plausible looking citations. That doesn't mean they are relevant. If you look at the citations in Jon's answer, they have multiple problems. The underlying problem is that we have no way of knowing if the citation/attribution is correct because we have no visibility into how it was determined to be relevant. — ColleenV 21 secs ago
@SPArcheon I was always taught to cite what you read. If you read the Agile Manifesto, you cite the Agile Manifesto. If you read my answer on SE or PM that summarizes and interprets the Agile Manifesto and didn't read the original content, then you cite my answer. And if you read both, then you cite both. Of course, this gets fuzzy when you talk about a system that is capable of "reading" and "remembering" everything, what do you need to cite? I've read lots of things that I'm sure influenced my thinking, but I can't cite every single book or post I've ever read on a given topic. — Thomas Owens 42 secs ago
@ColleenV I did check the citations. The linked pages had content that supported the generated content. I didn't analyze to see if the citations themselves were AI generated or not. But in this small sample, there is no evidence of making up citations. This doesn't say anything about other content containing similar information that could also be cited or anything about the correctness of the sources cited. However, the citation is there and the cited sources support the generation. You can always try Gemini or SGE for yourself and see. — Thomas Owens 48 secs ago
I can search Google and give you "citations" that are relevant. That doesn't make them accurate and doesn't help find where the idea came from. It's just a chain of hearsay about where the idea came from that misleads people into believing they know a fact by they way it's presented. — ColleenV 6 secs ago
@ColleenV I don't understand that argument. These citations are relevant and correct. Does that mean that they are the original source? No, but that's not the point of a citation. Citations are meant to be a chain. I could write something citing you and ideally you've cited what you read and they cite what they read and it leads back to an original source. Citing the sources used is socially responsible and the ethical thing to do. — Thomas Owens 42 secs ago
You assume because they're relevant snd confirm what you believe to be true about where that idea came from that they're correct. Why would you even need a citation in that case? Where did the person that wrote the SE post get the idea? These aren't citations. They're search results. — ColleenV 7 secs ago
@ColleenV I believe you have a fundamental misunderstanding of citations. Citations are not about correctness. They are only about tracing the history of an idea so that one can understand where it comes from to evaluate its correctness on their own. It would be up to the person who wrote the SE post to cite their sources (if any - something could be based solely on experience and not based on someone else's ideas), and we have a policy on citations. You can use the citations (or lack thereof) to decide if you trust a source. — Thomas Owens 28 secs ago
This is an example of what citation looks like quoteinvestigator.com/2024/05/08/new-idea — ColleenV 41 secs ago
How do search results match that definition of citation? None of those things cited are the origin of the concept. — ColleenV 13 secs ago
@ColleenV That is not what a citation needs to look like. Is that a good example? Sure. But there are plenty of formal and informal ways to cite someone else's work. — Thomas Owens 25 secs ago
@ColleenV It's not about the origin. It's about what was consumed to make the content. That may or may not be the origin of the concept or idea. Again, this is a fundamental misunderstanding of what a citation is and what purpose it serves. You may have concerns about how the citation is presented, and those may be valid. However, a citation is there and that citation also meets the requirements for attribution under CC BY-SA 4.0. — Thomas Owens 25 secs ago
How do you know it's not lying? You believe because an SE post was returned that it's properly attributed. What if it's not and that post had nothing to do with the information it generated? There's no way to prove that citation is accurate. That it's relevant is not proof. — ColleenV 46 secs ago
1:19 PM
A proper citation of data in an AI model would include the some sort of date or model version, not just a link to a page that was purportedly scraped to make the model. — ColleenV 47 secs ago
@ColleenV Again, accuracy of the cited source is irrelevant. Citations are only about pointing to the source(s) used to trace ideas. The training of the model and creation of the output are also two different concepts entirely. I do agree that there is insufficient information about the training of the model, but the cited content doesn't actually need to be part of the training in order to summarize. Gemini, for example, can summarize a file in Google Drive, including one that I just wrote and filled with novel content. Citing and/or attributing training material is a separate discussion. — Thomas Owens 52 secs ago
1:38 PM
@rene Nope, I forgot about those. Someone needs to let all the users defacing their content know about the datadumps as well, I guess... — F1Krazy 44 secs ago
1:50 PM
Unless you opted out - I think there's also a binding arbitration agreement. While its entirely possible to bring a company to its knees with binding arbitration, a class action suite would be tricky — Journeyman Geek 48 secs ago
2:14 PM
@rene See stackoverflow.com/help/licensing CC BY-SA requires attribution. LLMs and attribution are fundamentally incompatible, and they clearly have no idea what they are talking about when they pretend that they can give attribution. — Gantendo 40 secs ago
@rene See stackoverflow.com/help/licensing CC BY-SA requires attribution. LLMs and attribution are fundamentally incompatible, and they clearly have no idea what they are talking about when they pretend that they can give attribution. OpenAI would rather ask for forgiveness than permission, and SE cares little about licenses. — Gantendo 18 secs ago
@αλεχολυτ I mean... I understand that... but (assuming things haven't changed much since I was laid off), Adam is at least on the team doing this, so could easily confirm it. I respect trying to do things on your own - I just don't know how it'd work in this case, if it's even possible. As far as I'm aware, the Traducir strings are just that - strings - so any attempt to add formatting will just be rendered in plain text. The only option is to actually have the modal edited. — Catija 25 secs ago
My father wrote his PhD years ago. For every note he took from the literature he wrote up a card that told where it came from. That is a "citation". Taking some text and re-running your stochastic search on your corpus with a prompt to "produce something that looks like a citation" is not a citation. This answer is fundamentally wrong. — JonathanZ 49 secs ago
@AdrianHHH The "other" form is just a consent form. It's needed for ethical research and is simply there to help you protect your rights. It's pretty easy to read. The NDA on the other hand is complex. The team agrees that it could use some simplification. However, not all research includes an NDA. Typically if one of these are included you're being asked to participate in a session with new, shiny mockups—how exciting! — Piper 48 secs ago
@JonathanZ The rigor behind citations in a PhD thesis is very different than the rigor behind citations in an SE post (on most sites) or a blog post or a summary of search results. When publishing a thesis or in an academic journal, you would likely be expected to validate your citations and structure them according to strict formatting rules. You cannot compare how you cite sources in a PhD thesis or other academic journal to citing sources used to summarize search results. — Thomas Owens 1 min ago
Hi Jyotsna Bhakale, welcome to the Stack Exchange Network Meta site! I'm not sure which search brought you here but the problem you describe will not be answered on this specific site. To get an expert's answer for the topic of your question you'll have to find and then re-post on the proper site. Check How do I ask a good question and What is on topic on the target site to make sure your post is in good shape. Your question is definitely off-topic on Meta and is better deleted here. — Journeyman Geek 23 secs ago
"a good many folks here seem to be users with old accounts" - Yes, but in some unscientific observation, it also seems that most of those users haven't posted anything new in years (even after checking for deleted content). — NotTheDr01ds just now
Google's search results tags every result with the link it was found at. (Or at least it did for the first twenty years.) That again is a citation. You cannot compare rerunning your stochastic next-word selector to produce something that looks like a citation with an algorithm that actually tracks where it got something from. There is no rule that would prevent someone from creating a LLM that tracked this, but current ones most definitely do not, it's hard to see how to tweak them to do so, and we should not describe what they produce as "citations" until they do. — JonathanZ 1 min ago
First time I think I've ever downvoted Shog9 ;-). I understand (and yet ... don't) the anger about OpenAI. I've been fighting bad AI answers here probably more than anyone else, and I really don't think I'm exaggerating when I say that. But I actually see the OpenAI/SE deal as a win-win, as much as it can be. The content was already out there, scraped, used for training, etc. SE could go the "we'll sue" route like so many others (and I'm sure there was the threat of that) ... — NotTheDr01ds 23 secs ago
I find a lot of irony in how I'd have only found out about this by someone scraping a page on stackoverflow and removing all the important details — Sayse 29 secs ago
... but the deal that they brokered with both Google and OpenAI appears to require that responses from those LLMs will be attributed, and link back to the content here. That's an improvement over the existing situation. And I hope we all assume that SE is trying to make money on our answers - Hasn't that been the whole business model from the start? (not a rhetorical question - You worked there, so I'm hoping you have some insight) — NotTheDr01ds 19 secs ago
To the best of my understanding, the value SE users get out of this is that the company is claiming they will spend the income to invest in the public platform, which is something that is sorely needed. — Catija 28 secs ago
Hi Sadok Abouda, welcome to the Stack Exchange Network Meta site! I'm not sure which search brought you here but the problem you describe will not be answered on this specific site. To get an expert's answer for the topic of your question you'll have to find and then re-post on the proper site. Check How do I ask a good question and What is on topic on the target site to make sure your post is in good shape. Your question is definitely off-topic on Meta and is better deleted here. — rene 35 secs ago
I'm hopeful too, @NotTheDr01ds - but hope doesn't moderate vandalism, hope doesn't write code, hope doesn't enforce the terms of business arrangements, and hope certainly doesn't stand in for clear and honest communication about those terms. We're gonna need to see a lot more before we know what the real fallout is gonna be. As a former employee, I know what I'd have been doing right now in this situation, and I expect nothing less from current staff. — Shog9 12 secs ago
@NotTheDr01ds Why should unpaid volunteers keep mitigating the bad decision making of people who profit from their work already? Let Mr. Chandrasekar et al. reap what they've sown. — Dan Mašek 59 secs ago
@JonathanZ Except Google's SGE actually has functionality that tracks what content was used to produce the summary. Gemini and SGE are not just "stochastic next-word selectors". That is only a fraction of what they do. When you do a search, SGE receives the top search results and uses the LLM to summarize content and then structure it as a nice, human-readable summary. Again, you could argue about how the citations are presented, but you cannot say that what it doing is not a citation. — Thomas Owens 1 min ago
First an ad-ridden search engine and now ClopenAI? Bye StackOverflow... — pSrIoGcNeAsLs - bye stackGPT 33 secs ago
@Gantendo the content is dual licensed. SE also has a license to our content and that license doesn't require attribution. I'm not arguing whether I agree with all this shenanigans and whether I trust one tech company over another, I'm just trying to get it straight whether this violates anything. I don't believe it does so this argument is not going to win it. We need to come up with a better one, potential one that will survive in court. — rene 51 secs ago
@DanMašek cause to a certain extent the company hasn't been great stewards for years - I don't think the goal is to enable the company to do whatever, but rather ensure the damage to our communities is minimised in spite of those decisions, and ensuring protests have a positive impact that's more than burning down your account — Journeyman Geek 38 secs ago
@henningnolongerfeedsAI Uh, no. That's not how it works. Plus, any mod can and will undelete your deleted answers, because deleting your answers is "griefing". You can simply delete your account (and your PII), and then your answers and comments will remain, anonymized. — mbomb007 25 secs ago
I doubt the company will spend the money/effort on network sites besides SO and perhaps a couple of the other bigguns, on the rest of the network it's going to be up to moderators and users, realistically. (I trust CMs will look into things if it comes to their attention, I just don't believe that they'll be assigned sufficient priority towards hunting for it) — Bryan Krause 13 secs ago
Yep, they're integrating their search results. And that part does have citations. Of course those bits have been available for some time by just going a Google search. Maybe there's some use for AI to do selection on Google search results - I wonder how it picks? But regardless, LLM produced text currently cannot tell us what external source lead it to produce the sentence it produced, and anything that might make people believe that "Llama produce citations" is deceptive. — JonathanZ 51 secs ago
Hi Ola Śniegucka, welcome to the Stack Exchange Network Meta site! I'm not sure which search brought you here but the problem you describe will not be answered on this specific site. To get an expert's answer for the topic of your question you'll have to find and then re-post on the proper site. Check How do I ask a good question and What is on topic on the target site to make sure your post is in good shape. Your question is definitely off-topic on Meta and is better deleted here. — rene 41 secs ago
3:44 PM
@JonathanZ Again, you are incorrect. Gemini links generated statements - often one or a few sentences - to specific sources that support the ideas summarized in the sentence(s). SGE does something similar. How it picks is irrelevant. Why did I read and cite Bob's book on the topic instead of reading and citing Alice's book on the same subject? It doesn't matter. And I'm not claiming the LLMs produce citations - they don't and can't without additional stuff around them. Gemini and SGE can and do produce valid citations for content because they have more stuff than just the underlying LLM. — Thomas Owens 44 secs ago
@Shadur But by posting at all you've already lost that battle; deleting your content will only harm the public site, make it worse and create more work for the folks who stick around. I don't see any reason to believe that deleting your content will somehow keep it from being used in this agreement in any way; you've already irrevocably licensed your content by posting on SE, you can't retract that part. You can only control what you post here in the future, according to your own convictions about whatever this announcement will actually mean going forward. — zcoop98 46 secs ago
3:59 PM
@rene but OpenAI used data from SE before there was any agreement between SE and OpenAI. — Gantendo 14 secs ago
I, too, believe this is where the OpenAI partnership will lead; even just in a practical sense, I think there has to be a reason, some practical benefit for OpenAI to partner with Stack, and personally I believe that benefit is going to be more than bare-minimum access to Stack's content (which is widely believed to already have been scraped). I'd wager, given that it's being advertised on the OverflowAPI page, that the partnership will include some sort of access to purpose-built API/ infra to support whatever the end product will look like. — zcoop98 14 secs ago
Well, the cost needn't be huge @Bryan - a few minutes running a query a periodically would suffice. Vs. normal users hammering SEDE once a week and then filtering out false-positives by hand because SEDE doesn't have revision history for deleted posts. — Shog9 32 secs ago
The pushback on this is also interesting to me; I've been using the Search Experiments AI tool for a bit now, and I have yet to find an attributed response that doesn't reasonably summarize the cited source... personally, I just haven't seen any reason to doubt the "citations" (whether in the proper sense or not is besides the point) provided, because they've always matched up with the source when I click through. Now, sometimes that source is just bad or contains wrong info, but that's not an AI summary problem. It doesn't seem to be generating that data, it seems to be summarizing. — zcoop98 35 secs ago
@zcoop98 Exactly! We can debate what format the citations should be in and whether their presentation is appropriate. We can also debate Google's responsibility to ensure correct and validated content is returned via search results. But it is undisputable that citations exist, that citations are linked to generated content, and (so far, in my experience) that the summarized content and provided citations match up regardless of the correctness of the content being summarized. — Thomas Owens 56 secs ago
How is this an answer? It seems to be just a pessimistic view of the future, but it's not specific to the Stack Exchange / OpenAI partnership. — S.L. Barth is on codidact.com 42 secs ago
Have you considered copying your posts to the Physics Codidact site ? Codidact is a non-profit made by former Stack users, who were sick and tired of the way SEI treats their users... and built their own Q&A instead. Going there is a nice way of poking your tongue at SEI. (Or to make a more rude gesture, if you like :-) ) — S.L. Barth is on codidact.com 46 secs ago
@S.L.Barthisoncodidact.com Do you know if codidact disallows the GPT web crawler? — doublefelix 39 secs ago
But that OpenAI used your content is a problem between you and OpenAI, not something SE can or need to fix. What SE can do is use their license of your data to get reimbursed for use of the body of knowledge by OpenAI going forward so both SE and its communities gets somewhat compensated: SE in money, the community by having more and better features on the public platform as a result of that. — rene 50 secs ago
@Shog9 I didn't say the cost needed to be huge, only that I don't expect it to be spent. — Bryan Krause 41 secs ago
@doublefelix I think they do, looking for a reference. They do have a policy against posting LLM-generated content. — S.L. Barth is on codidact.com 26 secs ago
From someone who tried deleting his answers, and thought it through: I'm aware that those posts have already been scraped, but by updating them on the live site, the next scrape is likely to overwrite the previous scrape in the training data. So by editing, rather than deleting your answers, you are more likely to have them effectively removed from OpenAI (and other) training data. — doublefelix 28 secs ago
I think this is better served as a separate bug report as this question is only about the url rewrite in posts, comments etc. Can you please also post this on Stack Apps, tag it with bug and meta. I'll slap a status-review on it afterwards. This needs different eyes within the SE dev team. — rene 46 secs ago
...looks like it's not blocked at this moment :-( But I think they'd be open to discussing it on their Meta. — S.L. Barth is on codidact.com 56 secs ago
This partnership reads as if you were incapable of making OverflowAI work, so you're selling out to someone actually able to create LLMs. — T. Sar 37 secs ago
5:04 PM
"Rosie says that..." - management says many things. They are not always correct. — Shadow Wizard Love Zelda 24 secs ago
It's way too big. The bug is the error message that should say that. — Shadow Wizard Love Zelda 46 secs ago
Interesting it's actually only 120k so not too big, I'll try to check other possible limitations then. — Shadow Wizard Love Zelda 30 secs ago
Can I suggest adding "Is StackExchange planning to license subscriber content to OpenAI under anything other than a CC-BY-SA?" There is some ambiguous wording in the ToS that suggests this might be possible. It would explain what OpenAI is getting out of the deal (since they have all subscriber content already under CC-BY-SA). — Peter 58 secs ago
Well my best guess is the image dimensions which are huge (2373x1889) but only a developer can confirm. If you can resize that to e.g. 800px width (and respective height) this would probably work, and you can update the bug report with that info. — Shadow Wizard Love Zelda 16 secs ago
@ShadowWizardLoveZelda thanks I'll let dev debug as resizing may also change other gif info. But I guess you're correct. — Franck Dernoncourt just now
5:45 PM
(Disagreed with the edit removing the screenshot: need to show error message) — Franck Dernoncourt 21 secs ago
@anonymous about the edit proposal: thanks but the screenshot more info shows it's a Q (not A or chat), shows error message location+color, shows gif image preview, etc. — Franck Dernoncourt 45 secs ago
@rene - Been over this every other time this comes up. Proper attribution is required in order for the license to be honored, which GenAI doesn't do. It is a clear violation. There are numerous lawsuits which are about to become legal landmark cases, and this legal basis will be used here against Stack Exchange should it come to that. — Travis J 51 secs ago
Despite the "CC BY-SA" technicalities, which are definitely worth better visibility to the community as the current bounty suggests, this question raises an even more important topic: The company has said, "attribution is non-negotiable." I'm actually not against a partnership with OpenAI but I think it is important for the company to stay true to this promise to continue maintaining some level of trust with us, the community. — Josiah Yoder 26 secs ago
I understand that legally, Stack Overflow can do whatever they want with the data. But practically, to maintain community trust, the company DOES need to maintain an attitude of keeping promises such as "Attribution is non-negotiable". — Josiah Yoder 15 secs ago
6:15 PM
The gif and the screenshot are both easily readable on my phone (S23, i.e. a regular phone). — Franck Dernoncourt 51 secs ago
@FranckDernoncourt That's great for your phone and your eyes. On my phone and with my eyes, I absolutely cannot make out the (pink?) text on the dark background. In any case, I still think you get my point. — testing-malady 45 secs ago
I was not very fond of the idea that AI will train on our content for quite some time now (it has been over a year since the initial discussion about selling content for here on Meta) and company could definitely do a better job when announcing things and communication with users here. However, our content has been used for training AI along before all that so all anyone can do is not post new content if the don't want to be source for AI training. Vandalizing the content that is already part of AI only hurts people that don't want to use AI as a source of information. It is my duty to clean. — Resistance Is Futile 26 secs ago
Anyway, hopefully SE devs won't exclude 4k monitor users, or phone users whose screen are >1080p (i.e.. most recent phones). It's all about font sizes. — Franck Dernoncourt 28 secs ago
Nice catch! We're tracking this internally already, I've got a fix ready and it's just going through my team's review. — Kyle Pollard ♦ 14 secs ago
@FranckDernoncourt This is an actual screenshot of your question from my phone (iPhone 13 Pro Max - sure it's not the latest). See the difference between the error message on the screenshot and the error message rendered as text? I think it's far worse to exclude anyone who can't read an animated gif 2,300+ pixels wide than to prevent people from forcing them to. — testing-malady 15 secs ago
@JourneymanGeek I'm well aware of that (corp not being good stewards). And I certainly don't endorse any vandalism, I frown on that (it's against the spirit of the site, against the license agreement, rather childish, etc.). But let's be honest, the relationship between the company and the community has reached exploitative levels. It's unhealthy, and there's a point where you need to say "that's enough". — Dan Mašek just now
@JourneymanGeek I'm well aware of that (corp not being good stewards). And I certainly don't endorse any vandalism, I frown on that (it's against the spirit of the site, against the license agreement, rather childish, etc.). But let's be honest, the relationship between the company and the community has reached exploitative levels long ago. It's unhealthy, and there's a point where you need to say "that's enough". — Dan Mašek 1 min ago
I agree it's preferable add text when there is an images/gif. But I don't put the screenshot, I'm risking getting comments that the screenshot may have answered. — Franck Dernoncourt 24 secs ago
@FranckDernoncourt That's why I suggested that if you really, really, really think people need a screenshot to read an error message, you can link to it. — testing-malady 26 secs ago
6:58 PM
@zcoop98 Maybe "the folks who stick around" should look into getting an AI to clean things up for them, since they're okay with feeding all their work into an LLM to begin with. — Shadur-don't-feed-the-AI 20 secs ago
@6nagi9 It's up to you, of course, but I think that lack of detail might be the reason for the downvotes, as it's a perfectly fitting question for Meta. — Joachim 5 secs ago
@ResistanceIsFutile I'm not saying "don't clean". I'm saying, this is the company's mess to clean up. We've got folks in the comments talking about using SEDE to identify this stuff, using 10k tools, etc. - all of those suck for mass deletion. There are mod tools for handling mass deletion, but if those aren't sufficient then the solution isn't "burn insane amounts of time working around deficiencies in the system". This is literally a "few minutes" type of job for someone with employee access, so either they're gonna clean up their own mess, or you should seriously question things. — Shog9 8 secs ago
7:29 PM
@Shadur Framing fellow users who also care about the site and attribution of their content into opponents because they aren't willing to break rules is not a great approach. — zcoop98 7 secs ago
Every once in a while I get reminded that letting you go was really bad day for the whole network. This is one of those days. I am afraid that if we leave this to become company's problem it will never get resolved. — Resistance Is Futile 55 secs ago
If the problem really gets out of hand, then we will have no choice... but if it gets out of hand, then we will also have bigger problems. — Resistance Is Futile 10 secs ago
7:42 PM
I'm not worried, @Resistance - this is hardly the first time these sites have faced mass vandalism, and years ago my team devoted a lot of time and effort into developing and tuning systems to minimize the damage. Assuming those systems are still in place, still tuned, this is at most an annoyance. Folks are upset, and they have some reason to be - while I don't think this is an effective form of protest, I get it. And frankly, if it helps some folks' peace of mind to do a bit of vandalism and move on, so be it - as long as we don't let it waste too much time or effort. — Shog9 9 secs ago
8:20 PM
@S.L.Barthisoncodidact.com - This agreement is neither a major product change nor a policy change and so is not required to be announced first to mods. It would have been good if they had, though. — Mithical 32 secs ago
@testing-malady We can't link to it if we can't upload it. Posts should be self-contained, but images uploaded to the Stack Exchange image host are usually considered part of the post, even if they're just linked. — wizzwizz4 39 secs ago
There's a lot of misinformation out there recently about this partnership and the company's stance and directives to mods with regards to AI Generated content and moderator actions as a result of users vandalizing their content, is anything being done to address this? — Kevin B 24 secs ago
8:47 PM
Hi Divij Goyal, welcome to Meta! I'm not sure which search brought you here but the problem you describe will not be answered on this specific site. To get an answer from users that have the expertise about the topic of your question you'll have to find and then re-post on the proper site. Check How do I ask a good question and What is on topic on the target site to make sure your post is in good shape. Your question is definitely off-topic on Meta and is better deleted here. — Sam Onela 56 secs ago
9:09 PM
I previously somewhat defended StackOverflow after the Google announcement. This time I can't justify that anymore. We need proof and concrete information how SO fights for the values they supposedly hold so high. Consider me on strike, for whatever that's worth. — leftaroundabout 55 secs ago
@dan1st: You're assuming that StackOverflow doesn't violate the license that allows them to make and perform copies of the posts. (Of course, user attempts to delete the content have no effect on that, either StackOverflow complies with the license and has the right to distribute copies, or they act outside the license and every page served is a copyright violation) — Ben Voigt 42 secs ago
@Ramhound: What if I agree to the license (CC BY-SA) for the content, but StackOverflow doesn't (by their actions)? — Ben Voigt 31 secs ago
@bta: Many of us opted out of mandatory arbitration. The fact that the clause is there probably prevents a class action, unfortunately, but there still is room for a massive lawsuit with hundreds of plaintiffs. — Ben Voigt 28 secs ago
9:27 PM
That'd be great, thanks! If the upper bound is confirmed, I'll follow-up with a feature request. — Franck Dernoncourt 27 secs ago
I am not worried either. It is not a nice situation, for sure, but we still have way bigger problems with users that are actively posting AI generated answers. — Resistance Is Futile 6 secs ago
"Questions are closed under unclear pretenses." - This is false. - The reason questions are closed are listed in the close reason. — Ramhound 18 secs ago
9:58 PM
Legally, Stack Overflow has to abide by CC BY-SA. The answer quoted that. The other paragraph doesn't create a new license, it's just restating rights granted under CC BY-SA. Note that "irrevocable" does not mean "unconditional". — Ben Voigt 51 secs ago
And even if you argue that it does create a second license, relicensing is not one of the rights granted, so that supposed second license cannot be used for any partnership, OpenAI or otherwise. — Ben Voigt 57 secs ago
10:38 PM
@Will And none of that matters, as LLMS can't perform proper attribution by themselves. — Cerbrus 58 secs ago
« first day (351 days earlier) ← previous day next day → last day (11 days later) »
Transcript for
May8
May9
May10
♦ Meta Stack Exchange Comment Archive
A comprehensive archive of all Meta Stack Exchange Comments. C...
