FRANCAIS (traduit)
Le problème :
L’utilisateur a publié les commentaires suivants que je considère comme de purs mensonges :
- https://tarte.nuage-libre.fr/comment/2504870
- https://tarte.nuage-libre.fr/comment/2504863
- https://tarte.nuage-libre.fr/comment/2504780
Contexte :
J’ai publié des commentaires dans les communautés Gaming et PCGaming en essayant d’être objectif sur une nouvelle technologie, le DLSS 5. Cela n’aide personne quand la critique ne repose sur rien de concret et se contente de traiter cela comme de la « bouillie d’IA » (« AI slop ») comme on en voit depuis des lustres.
Cela a provoqué la colère de beaucoup de gens, car de nombreux utilisateurs veulent simplement pouvoir détester sans discernement. Le fait qu’à un moment donné, j’aie nié qu’il s’agissait d’une IA générative n’a rien arrangé, car c’était une erreur : le PDG lui-même l’a qualifiée ainsi. Cela a même conduit à un avertissement de la part d’un administrateur de mon instance qui a lui-même des avis très tranchés sur le sujet.
Dans chacun de mes commentaires, j’ai pourtant veillé à illustrer le problème dans le contexte de NVIDIA — une entreprise que je critique vivement pour la manière dont elle a érodé l’industrie — et à souligner les limites de cette technologie, simplement pour essayer de recentrer la discussion sur ce qu’est réellement cette technologie. Les gens ne veulent tout simplement pas l’entendre.
Actions :
J’ai signalé ces messages aux communautés respectives, mais face à la toxicité ambiante envers quiconque ne rejoint pas activement le mouvement de dénigrement systématique de l’IA sur cette instance, ou peut-être par manque de disponibilité des modérateurs, aucune mesure n’a été prise. C’est pourquoi je fais appel à vous.
Affirmations :
Je ne suis ni un robot ni un « assistant marketing », comme en témoignent clairement plusieurs de mes commentaires critiquant NVIDIA pour son monopole abusif et sa contribution à la bulle de l’IA via sa collaboration avec des cartels mondiaux. Il est absurde de prétendre que NVIDIA présenterait cela comme sa propre « stratégie marketing », et pourtant, nous en sommes là.
Il prétend que j’utilise de « nombreux pseudos » : j’en ai deux actifs, plus un troisième créé récemment suite aux problèmes soulevés par l’administrateur susmentionné, afin d’éviter d’impliquer l’instance dans ses préoccupations à l’avenir. C’est à la limite du mensonge : mes seuls autres pseudos sont un compte inactif et un autre provenant d’une instance qui n’existe plus, lemm.ee.
Il prétend que je diffusais de fausses informations, une accusation qui n’est réellement valable que pour la période où je ne reconnaissais pas l’usage de l’IA générative ; un point que j’ai rectifié bien avant qu’il ne poste ses commentaires en modifiant mes messages précédents suite à l’avertissement de l’administrateur.
Il parle aussi de « positivité toxique », ce qui, franchement, semble n’être qu’un moyen de masquer le fait que je ne partage pas l’avis général consistant à classer cela comme de la simple « bouillie d’IA », ou que j’apprécie certains aspects très précis et contextualisés de cette technologie — alors que mes commentaires restent majoritairement critiques.
S’il veut avoir ses propres opinions sur moi, aussi négatives soient-elles, libre à lui, mais ses affirmations sont purement et simplement fausses. C’est de la diffamation pure et simple.
Malheureusement, je dois également préciser que la partie de ce message rédigée en français a été générée avec l’aide de Google Gemini.
ENGLISH (Native)
The problem :
The user has made the following comments that I consider outright lies:
- https://tarte.nuage-libre.fr/comment/2504870
- https://tarte.nuage-libre.fr/comment/2504863
- https://tarte.nuage-libre.fr/comment/2504780
Context :
I’ve been making comments in Gaming and PCGaming communities that try to be objective about a new technology that has come out, DLSS 5. It doesn’t help anyone when the criticism is not founded in reality but just devolves into acting like it’s the image generation AI slop we’ve seen for ages now.
This has made a lot of people mad, because a lot of users would just like to hate undiscerningly. It hasn’t been helped that at one point I was dismissing it of being generative AI, because that was wrong: the CEO actually referred to it as such. This even led to a warning from an admin of my instance who themselves have strong opinions about this issue.
In any comment I’ve made, I’ve also made sure to try to illustrate the issue in the context of NVIDIA itself, which I’m pretty critical of given how much they’ve eroded the industry, and to criticize the limitations of this technology as well, just to try to shift discussion into what the technology actually is. People just don’t want to hear it.
Actions :
I’ve reported them to the respective communities, but because of how much toxicity there is against anyone just not actively jumping on the AI slop bandwagon in this instance, or just because they may be unavailable, they’ve taken no action. As such, I appeal to you.
Assertions :
I am not a bot or “marketing aide”, as should be obvious from several of my comments criticizing NVIDIA of being an abusive monopoly feeding the AI bubble through cooperation with global cartels. It should be absurd to claim that NVIDIA would claim this to be “NV’s marketing”, yet here we are.
He claims the use of “many aliases” - I have two active, another one created recently given the issues the aforementioned admin has brought up so as to avoid involving the instance in his concerns in the future. This is borderline false, the only other aliases I have is another inactive one, and one from an instance that no longer exists, lemm.ee.
He claims I was providing misinformation, a claim really only valid when I wasn’t acknowledging the technology employed generative AI, a claim I resolved well before he made his claims by editting my previous comments after the admin’s warning.
He’s also claiming “toxic positivity” which, frankly, just seems to be a way to whitewash that I am not going along with how everyone is just trying to classify this as the image generation AI slop we already have, or that I have a very limited and contextualized things I like on some very discrete aspects of the technology - my comments are still mostly criticism.
If he wants to have his own opinions about me, negative as they may be, he is welcome to, but the claims are flat out false. It’s just flat out libel.
Unfortunately, I also have to add a disclaimer that the part of this that is in French used Google Gemini.
@verdi@tarte.nuage-libre.fr, the fact that another user has a different opinion on a product than yours does not imply that the user is part of nvidia’s advertisement team.
If the user says something false which is not a opinion but a fact, you can give it a source or demonstrate it. If the user ignore the message and continue to says something objectively false (something that is not a opinion), then it is strange but it doesn’t mean that the user is a bot or working for nvidia. It could be that the user, because of its opinion, has difficulties to change his mind, or it could be that your explanation was not clear, or it could be that the user missed it and didn’t read it. Remember that humans can make errors, and will make errors in their life.
What define something generated by AI and something modified by a complex algorithm is not the same for everyone.
If you don’t want to see the messages of the user you can block him, but you can’t imply things (the use of likely doesn’t change anything) you can’t prove about a user, and you are not supposed to say multiple times the same thing about the user. This behavior can be toxic over time and be considered harassement.
You will always find people with different opinions. On piefed, if you don’t want to see certain opinions, you can use the block feature.
Point 1:
I said
hethey spammed positivity and misinformationedit: didn’t mean to misgender anyone is just French defaults to male neutral and that colours up my English.
It’s a fact admitted by himself.
Point 2:
Libel implies I spread misinformation about them. That is patently false, my statement was “LIKELY a bot or a marketing aide” (emphasis post hoc), that implies probability and not certainty. I offered evidence of both the spam and misinformation, the user tried to address (some would call it mask) the misinformation but the fact of the matter stands, they spammed misinformation under several pseudonyms and that misinformation was overwhelmingly (and unreasonably) positive about a technology that has real consequences for working people. The conclusions about his motivations were presented in no uncertain terms as speculation for any reader.
Point 3:
Accusing me of libel is the social media equivalent of a slap suit, my contribution was to the point and offered evidence of spam and my EVIDENT speculation as to why the spam was being incurred. Over 50 users agreed.
Point 4:
If misinformation and misinformed opinions are protected speech, so is speculation as to the motivations for the dissemination of false information under several pseudonyms.
You’re welcome to act in accordance to your moderation ethos, I will continue to try and avoid people being misinformed by users trying to make a quick buck.
Bonus point: GPTZERO flags more than 50% of that user’s comments, larger than 250 characters, as AI generated, derive from that knowledge what you will.
The problem is not that you are totally wrong (because you are not totally wrong when you say that the user spread some misinformation), the problem is how you say it.
Making an answer to attack its arguments is ok but attack against the user is not ok. Attack on users can lead to toxicity.
Using the word likely does not erase the accusation. Someone reading the message will remember the accusation and not try to verify it. This is probably why the user uses the term “libel”. The word “likely” doesn’t change what the reader understand.
Accusing someone to be bot, or to be payed, without any proof will not help to fight against misinformation.
If misinformation and misinformed opinions are protected speech
They are not. I don’t moderate content in remote community when content is posted by a remote account.
If the user makes repeated misinformation in the same community, the moderators of the community are responsible to act to stop the misinformation.
In one sentence : the form is problematic. You can say that specific comment are misinformation and say which parts are false with explanations.
Bonus point: GPTZERO flags more than 50% of that user’s comments, larger than 250 characters, as AI generated, derive from that knowledge what you will.
AI detectors are not realiable unfortunately.
Understood, I’ll dial it down in future contributions.
edit: part of his additions to the comments are now revealed to also be misinformation.
@administrateur@tarte.nuage-libre.fr
I’m actually having to reply to myself because for some reason it isn’t federating correctly, at least as of now: https://web.archive.org/web/20260318212358/https://lemmy.ca/post/61961274 (And after posting this, it only shows this one comment. It’s also been having issues federating the latest edits, but then again I’ve continued doing so until now to leave it ready to be archived: https://web.archive.org/web/20260318222736/https://lemmy.ca/post/61961274 )
The user continues to claim I continue to misinform, which I don’t actually think I’m doing but that’s up to him. They have not eliminated or edited any part of the claims that were deliberately false. He can claim I’m misinforming all he wants, it’s not deliberate and I do not believe I am.
This thread isn’t about what he disagrees with, and I doubt he is interested in any actual discussion if he isn’t bringing it up directly. He could, say, edit his comments to eliminate his false accusations with this evidence he thinks proves my claims wrong, but isn’t.
But, to defend myself, and I have admitted and corrected myself when I was wrong, what Jensen is saying now has not contradicted anything regarding previous claims from actual quotes on how the technology works.
Citing myself:
-
It takes color and motion vectors and passes it as input through AI - same as previous versions of AI.
-
It generates no new shapes or geometry.
-
It just applies changes to lighting and materials, basically creating a mask over the original.
Citing what they’ve told us:
DLSS 5 introduces a real-time neural rendering model that infuses pixels with photoreal lighting and materials. Bridging the divide between rendering and reality, DLSS 5 empowers game developers to deliver a new level of photoreal computer graphics previously only achieved in Hollywood visual effects.
DLSS 5 takes a game’s color and motion vectors for each frame as input, and uses an AI model to infuse the scene with photoreal lighting and materials that are anchored to source 3D content and consistent from frame to frame. DLSS 5 runs in real time at up to 4K resolution for smooth, interactive gameplay.
Citing what Jensen actually is telling us now:
“It’s not post-processing, it’s not post-processing at the frame level, it’s generative control at the geometry level,” he said.
This is not inconsistent with what we have been told, the filter takes a game’s color and motion vectors and changes the defined “photoreal” lighting and materials of the geometry.
It is inconsistent if you interpret “generative control” to mean changing the geometry, and we will know soon enough, but my own attempts to test these claims suggest otherwise:
https://i.imgur.com/HeQvqVO.gif
Regardless, if it does turn out to be misinformation, it is not deliberate, it is consistent with what has been told to us, and I’ve even bothered to double-check it myself. I find it rather hypocritical to complain about misinformation while continuing to to promote it, given I have not seen them take any action regarding their comments yet. In contrast, the moment I realized I was wrong about my claims about generative AI not being involved, I corrected myself.
I don’t think his comment about “more than 250+ words, so it must be a bot” comment is in good faith either, given the extents I’m going through to document, reply, and edit this out. I mean the guy just literally posted a 320+ word post, so either he’s not being self-aware or worse. I apologize for this word stew, but the guy is making this personal.
Nor do I think he actually cares about whether I am continuing to “spread misinformation” given that he avoided telling me. Being fully aware I have gone back to edit my comments when told, he avoids slapping what he thinks is an obvious “I told you so” to my face.
I hope you don’t mind if I don’t pass it through the auto-translator this time now that I realize you also speak English fairly well :)
Federation does not work probably because nobody from lemmy.ca subscribed to this community. Maybe it will work better with ping. @theobvioussolution@lemmy.ca
I don’t think his comment about “more than 250+ words, so it must be a bot” comment is in good faith either
The user didn’t said that, its comment says that he used AI detector on your comments larger than 250 characters and that 50% are flagged as AI generated. The user could think that it means that you use AI to correct your comments, or to organize them, or to write them completely.
I hope you don’t mind if I don’t pass it through the auto-translator this time now that I realize you also speak English fairly well :)
I wouldn’t say fairly well but yes I can speak English
The user continues to claim I continue to misinform, which I don’t actually think I’m doing but that’s up to him. They have not eliminated or edited any part of the claims that were deliberately false. He can claim I’m misinforming all he wants, it’s not deliberate and I do not believe I am.
If I understood correctly, the user said indirectly that the user will change the form in future contributions. So no future accusation to be a bot or payed by nvidia.
Are previous comment very important for you ? Posts after a few days will probably be lost in thousands of other posts with no interactions.
-

