OpenAI signs AI deal with Condé Nast
- Thread starter JournalBot
- Start date
It's a phone that is being sold as Google's first Pixel for Gemini. Ergo our conclusion: "If you're not into AI, the 9 is not a must-have upgrade from the Pixel 8 (or even 7)." It's germane IMO. I have no problem with that usage.
Upvote
20
(24
/
-4)
I know they have been because I've gotten chat agents to write poems about Ars comments with details that could not be random.
That was last year. And the data here is valuable, especially paired with metadata. I don't personally care if the model is trained on me but it'd be nice if people had a choice even if that was on a per-user basis.
Upvote
7
(9
/
-2)
train_wreck
Ars Scholae Palatinae
I understand, I was not trying to be insulting or anything. Thanks for all y’all do here. Also resistance is futile.
Upvote
20
(21
/
-1)
They do. My understanding is that data comes from Common Crawl which anybody can use. There isn't a need to scrape sites directly for training purposes.
Upvote
7
(8
/
-1)
You do realise you replied directly to a member who used what looks most likely to be a real name?
But anyway, back to the matter of everything's already been scraped so it's not worth editing - I subscribe to the view that morally my posts are mine, if cn/ars has some legal fine print that says their yours then I would put forward the idea that the anticipated uses of said posts did not include training of AI.
Mortally, we should have been given the choice to opt our posts out FIRST!
Ps another sub lost - sorry much as I want to support ars I can't square this circle - I will re-sub [and participate in forums] if my posts can be excluded
Upvote
23
(27
/
-4)
I would resubscribe if and only if a knob is given that allows all of my posts to be deleted from Ars' databases forever. I don't mind paying for quality journalism at all. I no longer want my posts on Ars to be available for any AI to scrape.
Blocking things in robots.txt is clearly not going to be enough, and I no longer trust that Ken actually understands why some of us might want our consent to matter, or care if he does. At this point, only making significant efforts to honour user consent will bring my home's three subs back to Ars. Trust, once lost, is very hard to regain. What the status quo was, or how things were before the announcement...none of that matters now.
Ongoing consent matters. Just because someone says "yes" at the beginning of a relationship doesn't mean you get to ignore them when they say "no" later on. I don't feel that's unreasonable, and it is a standard I believe we need to uphold in all aspects of our society.
Upvote
17
(27
/
-10)
I am a long-time lurker who has commented minimally to avoid scenarios exactly like this as I am a cynical non-american.
Funny you're not getting it - CN has thrown any "value" you had down the toilet and you're just taking it and even actively supporting it.
Do tell me, which would make a stronger statement to CN in terms of showing them how disliked this move is (provided you really want to fight it which seems doubtful):
- Option A: Suddenly and without any warning turn off an ability to edit comments older than 60 minutes - alienate your readers and any "sense of community" and any trust your readers had in you whilst your tongue is deep up the corporate bottom to keep the "value"
- Option B - Do nothing (maybe take a backup right now). Let the community run wild deleting posts. Escalate this, together with all the feedback to the corporate a** you are forced to lick right now, showing what their actions did to your value - while showing your readers that you care, with actions, and not just empty platitudes.
edit: this may or may not be somewhat of a throwaway account
Last edited:
Upvote
14
(34
/
-20)
Those of us who choose to post under our real names really have no expectations that our posts are anonymous to be fair.
Upvote
36
(38
/
-2)
I appreciate the personal response (and that generally you're been responding to a lot of people in this thread today).
To me, when you've split the review into "the piece about AI" and "the piece for people who don't want to hear about it", AI content in the latter is not germane.
However, I don't expect to agree with the editors 100% of the time, and it's more important to me that you're showing up here willing to set out your perspective. Thank you for talking openly about these policies.
Upvote
38
(39
/
-1)
You may need to edit the wording of this policy. I feel like most people would consider graphics (a/k/a "artwork") to be a "material", in which case the use of Gemini-generated artwork in this article violates the first bullet point.
Upvote
15
(19
/
-4)
Upvote
45
(50
/
-5)
You know what is a shield? Deleting what you've put online.
Most of what we put online was put there before generative AI trained on scraped content like /civis was a thing.
Regardless of your feelings about the given reason for wanting to delete old comments, it is a reason, your feelings about that reason don't change that.
Upvote
42
(47
/
-5)
@purecarrot - more seriously, the problem as I see it is two-sided:
Out of those two possibilities, neither makes it appealing to have AIs come to the fore given the current shape of our society. Not only we cannot guarantee that AIs could reliably be granted more influence on our future, but we humans also aren't prepared to deal with AIs partaking in it:
If AIs suddenly produce the majority of our cultural output, what will we watch/read/listen to tomorrow? In the first case, not much of value; in the second, our culture becomes that of machines and our minds will become used to thinking more like theirs too.
If AIs also start guiding our policies, in addition to our economy, will the human race still be in control of its destiny?
If AIs shape our knowledge through available information, will the human race still be able to change course if they don't realize what course it is on?
Again, all of the above are problems happening under simple AIs, more advanced but still similar to what we have now. I've not even looked at the alternative of sentient and self-improving AIs, which would be a very likely consequence of the second alternative above.
Before you discount me as a retrograde luddite again, I've long considered the future of humankind, and I don't think a human-AI war is ineluctable, nor do I see it as the only outcome of truly-intelligent AI. I'm convinced our species centuries from now would be unrecognizable to us today, from instant-access shared knowledge to enhanced genetics to cybernetics leading to very different body structures. I surmise we'll end up as a single-consciousness superorganism (or rather, a few such entities, given the limits of radiation over interplanetary/interstellar distances) for which the notion of individuality and intra-species violence will seem quaint. I believe human intelligence and 'artificial' intelligence will be melded and undistinguishable at that point.
But today's society is not ready for AIs, and today's AIs are not ready to seamlessly augment our species. Why rush unprepared and risk prematurely crashing the momentum in the best case, a dark age, or a catastrophe, ... or worse? I'd venture nobody knows the odds of the worst alternatives, but do we really need to find out?
Because a few corporations are struggling to come out on top of their own capitalistic game of survival, do we want to let them ruin it for all of us? Let's take it slow, with regulations and ethics committees and safeguards every step of the way. Let's make reaching the singularity as boring and eventless as we possibly can.
- if AIs are not actually intelligent, then their contributions are not going to help civilization move forward I'm afraid. There's already enough noise drowning actual information - coming from human contributions, we don't need it amplified further;
- but if AIs do come through in the short term as superior intellectual entities, I would selfishly appreciate that we first ensure that they won't get rid of human cockroaches as soon as they're given the chance, before we give them the keys to the kingdom (they can outpost us in the virtual world, and probably soon drones can outfight us in the real one). Even if we don't give them direct access to the means of controlling us, consider that pets like their human master more than they do other members of their own species, and I'm not convinced the same couldn't happen with humans.
Out of those two possibilities, neither makes it appealing to have AIs come to the fore given the current shape of our society. Not only we cannot guarantee that AIs could reliably be granted more influence on our future, but we humans also aren't prepared to deal with AIs partaking in it:
If AIs suddenly produce the majority of our cultural output, what will we watch/read/listen to tomorrow? In the first case, not much of value; in the second, our culture becomes that of machines and our minds will become used to thinking more like theirs too.
If AIs also start guiding our policies, in addition to our economy, will the human race still be in control of its destiny?
If AIs shape our knowledge through available information, will the human race still be able to change course if they don't realize what course it is on?
Again, all of the above are problems happening under simple AIs, more advanced but still similar to what we have now. I've not even looked at the alternative of sentient and self-improving AIs, which would be a very likely consequence of the second alternative above.
Before you discount me as a retrograde luddite again, I've long considered the future of humankind, and I don't think a human-AI war is ineluctable, nor do I see it as the only outcome of truly-intelligent AI. I'm convinced our species centuries from now would be unrecognizable to us today, from instant-access shared knowledge to enhanced genetics to cybernetics leading to very different body structures. I surmise we'll end up as a single-consciousness superorganism (or rather, a few such entities, given the limits of radiation over interplanetary/interstellar distances) for which the notion of individuality and intra-species violence will seem quaint. I believe human intelligence and 'artificial' intelligence will be melded and undistinguishable at that point.
But today's society is not ready for AIs, and today's AIs are not ready to seamlessly augment our species. Why rush unprepared and risk prematurely crashing the momentum in the best case, a dark age, or a catastrophe, ... or worse? I'd venture nobody knows the odds of the worst alternatives, but do we really need to find out?
Because a few corporations are struggling to come out on top of their own capitalistic game of survival, do we want to let them ruin it for all of us? Let's take it slow, with regulations and ethics committees and safeguards every step of the way. Let's make reaching the singularity as boring and eventless as we possibly can.
Upvote
9
(10
/
-1)
I understand wholeheartedly the intention behind this, in effect this action is alienting your core base of both subscribers and readers alike. Its difficult to interpret this decision in a positive light without prior notice. Only an explanation later provided in the comments of the article. A move that was deployed sometime near the release of this article, in anticipation that your community would in some way react like this. You'd have to pre-emptively consider the motivations your commentors would have to do exactly that, so you understand fully where that sentiment is coming from. However, to arrive at the concept of their actions of editing their own comments as being vandalism, is disheartening to say the least. You've already seen how this plays out with Reddit. I'll just quote something from exactly one of those articles:
The Internet can be a fickle creature, but if there is one lesson that seems to consistently ring true it’s this: don’t alienate your core users.
R: Whats Next?
Upvote
34
(38
/
-4)
I have over 33,000 public posts in our forum under my real first name, I have as much skin in the game as anyone really.
If you don't want to have your posts be public then I would continue not post.
Our rules have always, and forever, not allowed people to go back and edit their posts away. We left the window open so that things like perpetual threads could keep getting updates, people could mark items as sold in the Agora forum, etc. It was an honor system, and people abided by it.
When several people in a row broke that trust today I turned off that ability. I'll sort it out later.
We have worked really hard over the years to keep our forum's history. It would have been much cheaper and easier to dump all the old posts frankly. It matters to us. I'm not going to apologize for caring about the community.
And this very much is a community, our threads are shared fabrics of interactions. This isn't anyone's private feed. We're not Twitter or Facebook, and we've never pretended to be.
Upvote
13
(45
/
-32)
That's not "controlling people's speech". It's a reasonable step to prevent vandalism. If you want to correct what you said, you are free to do so by adding a new post. Get to it.
Upvote
1
(15
/
-14)
Lets be honest, the journalism industry is hurting for funds. While I can't really blame Conde Nast for taking the deal, it definitely leaves a dirty feeling. We know that only large companies with power will be able to extract licensing deals from OpenAI. Small outlets, which form an enormous long tail on the internet, will only have two options: block or get scraped for free.
Upvote
7
(8
/
-1)
To be fair, the full phrase is "generate material" and that is not what reportedly happened for that image.
Upvote
3
(3
/
0)
I don't have a dog in this fight. I think the frustration is that this kind feels how Reddit reacted, and people are still salty about it.
Upvote
28
(28
/
0)
The rules are the rules, and always have been. You agreed to them when you made your account, and every time you post.
I have a user group that restricts edits to 60 minutes that I manually add people to when they abuse the edit privileges. In this case I simply did it for the whole forum from my phone while I was out, because I'm not interesting in trying to track down a bunch of people mass editing things.
Upvote
-8
(31
/
-39)
My recollection of the language, posted again a few page back, is they can use it for anything.
Yeah, but with many posters there's enough of their posts to replicate their style. For example, let's say I scraped Ars and fine tuned on it, I'd just have to put your username in a similar prompt to whatever I tuned on to reply in your style to just about anything possibly with some personal details real or hallucinated.
I don't know how OpenAI would weight this content or even if they would bother, but it wouldn't be a terrible idea to use it as part of a dataset especially given upvoted answers tend to be factual and downvoted, the opposite. From a technical perspective you know I would like to but also that I have enough respect not to.
Upvote
-7
(0
/
-7)
Upvote
-16
(4
/
-20)
Well, unlike reddit, you've never been able to go back and delete your posts at Ars. This isn't new.
Upvote
11
(25
/
-14)
It's not, to be honest. Our biggest and best argument is that we're opposed to it, they don't need it, and it stifles participation. A bunch of people canceling subs and then still hanging out here ultimately has the opposite effect as intended.
I hope people can trust that we're making the best case we can.
Upvote
27
(33
/
-6)
Upvote
19
(33
/
-14)
66k posts? That's a lot of you. A valuable lot of you. At a penny a post that's $600. Are your posts worth a penny? That's up to Conde Nast and OpenAI. You won't see any of it.
Upvote
-3
(7
/
-10)
How the fuck can we prove that to you on demand?
When the OpenAI LLMs haven't been apparently updated with the Ars comments datasets yet?
Should I try it once all your 19,365 comments here actually make it into the latest ChatGPT? From the random tidbits that Aurich or you were at a certain conference at the time to you liking old game consoles or whatever, collating big data can absolutely give you a lot of PII about a person.
I am certainly NOT scrapping the whole 19,365 of your individual comment posts here into my home hypercluster just to prove a point, though
Yet the point still stands and is pretty valid.
I am pretty sure a big‑enough LLM could find out plenty of deep profiling PII info about you just by collating all the small tidbits in all your 19,365 posts. Generating a pretty accurate profile of you.
And sure, you can still say that it's the users who put it out there, but in plenty of jurisdictions like the EU, collating all of that stuff is by AI is still a potentially illegal privacy violation with some pretty fucking hefty fines attached.
Upvote
16
(29
/
-13)
train_wreck
Ars Scholae Palatinae
I understand the rationale for not allowing editing of posts, even though i don’t fully agree with it. I have a harder time understanding the rationale for disallowing deletion of posts.
Upvote
9
(15
/
-6)
THIS. This, right here.
The issue at hand is one of consent. Ars' response thus far has been "well, your posts are already out there, so it doesn't matter if we violate your consent", as well as "you said 'yes' at the beginning of the relationship, so we don't have to honour you saying 'no' now".
Either our consent matters, or it doesn't. So far, the message from Ars is that our consent does not matter, and that is Not Okay.
Upvote
29
(40
/
-11)
We could always do the closest thing to and edit them down to nothing. Now we can't. That is new.
Bing was correct when they called posters here pedantic.
Upvote
3
(10
/
-7)
You gave your consent when you agreed to the site's TOS.
The deletion of posts breaks the continuity of threads.
The thread that spawned the WHARRGARBL meme is a minefield of broken links and some after-the-fact post editing that renders it down to gibberish. To the point where the Know Your Meme page has details wrong because the necessary context no longer exists.
Upvote
12
(22
/
-10)
Exact same reason.
When you post to the community your post belongs to the community is how the Ars forums have always functioned. I mean, it's true legally on whatever level, but that's just the lawyer stuff I don't care about. What I care about is the fabric of things. We call each topic in a forum a thread, and they really do weave together to create a larger whole.
This isn't a private social media feed, where you're posting your own stuff, and can just nuke it whenever you're done with it. The Ars community fabric is made up of replies and interactions, and if people start ripping holes in that the whole thing frays and loses structure.
It's why every time we've moved forum platforms over the years we've tried to keep all the old posts and history intact.
It's why there is no delete function. Now if I could allow for a short deletion window for mistakes or double posts etc, I would. But the deletion and edit windows are tied together in the forum. Can't separate them.
So I guess if we decided to keep the 60 minute edit window we could also add a 60 minute delete window too.
Upvote
17
(29
/
-12)
Two points. 1) It can be abused by trolls who leave inflammatory comments and then wholesale alter or delete it to make the conversation look weird. 2) In the old forum software Ars ran on for about 20 years, deleting posts would fuckup the back-end and do weird things to pagination and link-backs.
Upvote
25
(26
/
-1)
You could if you broke the rules, and if I caught you doing it I would revert all your edits and remove your long term editing privileges.
Again, none of this is remotely new, these are the rules Ars has always had. That's why this user group exists:
Upvote
13
(27
/
-14)
- Status