heredos
@heredos@heredos.net
Location: 48.776396,2.335207
39 following, 25 followers
Ayé, j'ai envie de me mettre à écrire. Je sais comment commencer. Il me faut juste un outil physique. LibreOffice, je ne suis pas fan, ça sent l'usine à gaz. N'importe quel site dont je ne connais pas les admins, c'est juste non.
Vous connaissez des outils libres qui me permettent d'écrire un roman ?
Et si l'apprentissage du markdown est un problème, il y a des boutons.
Mais c'est la même syntaxe que sur fedi
Obviously.
What about boneless buffalo wings?
smh coherence
Ajustez votre vocabulaire 🤣
@heredos Enfin!
C'est Jean Granger qui va etre heureux!
the issue with generative AI to me is not the technology but purely the ethics behind the training data if you gave me a local diffusion model or LLM that I had to train myself, or is otherwise trained only on data by people that consent, and I can then stretch and bend that technology to achieve artistic results...I think that is totally fine kind of like those early projects with image diffusion models where people tried to break the software and find words that, when inputted as negatives would give weird results that say things about their training data publicly available models are, even if they say they are, The things that you
@alexia i kinda disagree, given the things that AI really accels at are also really horrible (facial recognition for mass surveillance, giving it a giant dataset of an entire social site and asking "where give me everyone who said something like .. " and so on. "automatically point the gun at the person and shoot them + now with image recognition!" and other such 'tasks' which they are currently making ai for the express purpose of "solving" ..
the fact that in general the AI "dream" is to create slave labour where they can't fight back; and all the people going off about "safety" is how they might try not let you do that, while going off about much "benefits to humanity" when, such "benefits" in practice are, badly written stories and art and .. the things it's "good at" are just 'can hurt people easier'
in general, i think these are much bigger issues and in general indicate a issue with the technology as a whole, that is much much larger than just unconsentual training data.
facial recognition for mass surveillanceWell this is not generally under "generative AI", though. That's a different category altogether and would go under computer vision, which isn't at all what I was talking about
Either way, I don't think there's a "good versus bad" type of way to look at this because I could equally say that various machine learning applications have been found in medicine to detect cancers and diseases early, which is objectively a good thing and would also fall under the AI umbrellaterm
there isn't a single way to look at this, but especially not a "it's all good" or "it's all bad" perspective here, I think.
@alexia and i would say pointing to the few 'good' things it does, while missing all the context behind those and the general field as a whole
is like defending the state's existence because sometimes it does welfare, or .. saying the military is fine because sometimes they make hospitals or the internet;
like i would say that is ignoring the primary focus on what "ai" and 'ai research' is doing and is trying to do.
@Li as I said, nuance.
@alexia .. there is no nuance here though;
ai is developed with the express purpose of hurting people, and its excused via false claims of a "utopia" built upon slave labour; the fact it _sometimes_ does something not so bad. doesn't change that.
ai cannot exist ethically within the current framework it is researched & developed in..
@Li I don't think it's fair to just disregard the time spent by people to fight cancer because they are, as we all, part of a system that imposes terrible things on everyone that is ran by terrible people
@alexia maybe if the same people using it for that, weren't also turning around and using it to be horrible to everyone, but they are all part of that same system and all contribute to it. thus they are all partly responsible for it;
sigh .. in the same way no one saying to abolish the military and that it is a bad institution.. is saying we should also dismantle down the internet just cause they made it..
i am saying ai isn't magically fine if done with 'ethically obtained training data', nor is it so if you cherry pick a couple not so bad things it's been used for.
my point is more along the lines of : "a nuclear bomb with open source tracking system and built with ethically sourced uranium is still a nuclear bomb."
for a GenAI example; and what it is actually being effectively used for, notably; creating intentionally misinformation, and propaganda; it's very good at that..
also just considering the power those who control the system prompt, training data, and consider the how much blind trust people place into AI and what it says, they take it as fact without checking any of it;
for example, when they make an "uncensored ai" where "uncensored" just means "we told it to say far right bullshit" .. and put it on one of the largest social media sites ever,
everyone blindly believing what it says because fancy ai said it;
hence; GenAI is incredibly useful for encouraging not to look at any details for things, and where you want to say something that sounds correct and convincing but where you don't care if its actually true or not, as such it excells at making propaganda and misinformation, and fails at anything where it being like correct actually matters at all.
@alexia also these tie in to eachtother alot, there are things like deepfakes which are a bit of computer vision and a bit of GenAI, and pretty much serve the only purpose of getting people to do things you for whatever reason couldn't just get them to do normally. which is generally not very good.
@alexia it also feels weird to me to use this exact system that imposes terrible things on everyone; to claim its 'unethical' in the first place;
like copyright and the entire concept of 'interlectual property' is, itself part of that horrible system that they impose on everyone, it's not legitimate, it shouldn't exist at all, you shouldn't actually have a complete monopoly on how something you made gets used,
putting forward as seemingly the only reason why it would be bad, when in reality, copyright is itself part of the problem, and is a system used to excuse doing immense harm to people; in the name of preventing 'infringements' ..
it is like; the least of the issues with GenAI even.
these last three are open source according to the osi definition, meaning you can get access to their training data and train them yourself if you have enough money and technical knowledge.
This is usually a big green flag that the model has not been trained on copyrighted or otherwise unconscenting data, because people would be able to tell much, much more easily, as you can just do a simple search in the publicly available datasets, as well as compare the final performance of a model you trained yourself that there is no unknown extra data involved.
As of yet i do not know any diffusion model trained this way (sadly), so i just have to trust that freepik did not lie about training F-lite solely on data they own
IP/Port: 163.44.120.53:5900
Hostname: v163-44-120-53.skta.static.cnode.jp
Client Name: computer:0
Location: Funabashi, Chiba, JP 🇯🇵
ASN: AS58791 GMO Internet Group, Inc.
VNC Password: N/A
ID: 39136746
Added to DB: 05/06/2025, 06:14:46 AM (UTC)
Last seen: 05/06/2025, 02:10:36 AM (UTC)
https://computernewb.com/vncresolver/browse#id/39136746
3 billion devices will soon be insecure with Windows 10 🔥🔥🔥
@microsoft
But not in Europe. 😝
Private users will continue to receive security updates. Consumer protectors have fought for Win10
@LeelaTorres@ieji.de overthrowing the European Union
'kids these days dont even know what a DS is'
why does everyone say stuff like this, but then when i actually talk to "kids these days" they seem to know all these things
In the end, AdBlock won and Youtube is usable again
What do you mean?
So I have slowly started working on FediQuest : An Emoji Battler poll driven Game!
Will you help fight the spam monsters and AI-slops who invaded the FediWorld with your arcane emoji mastery?
So how old is he now in 2025?
@ParadeGrotesque @Kingu no no no you got it wrong! 10 to 18 is 8 that makes the kid 18 in 2018, and 10 to 25 is 15, so you add the 15 to the 18 and thus in 2025 the kid is 33.
Duh.
We made a comic about street smarts (this is one of my favorites we've made).
TlDr: france has old estate agents going moe moe kyun on bread paper bags
(1/2)
Yes: | 0 |
No: | 0 |
What? No 3rd option?: | 0 |
Closed
I did a dirty install of PopOS and got it up and running again.
SO MUCH LEARNING
SO FUN
GRINDING MY TEETH
i learned a lot though, among other things, that i won't be touching it again even with a 10 meter stick
@heredos@heredos.net I wanted to try mint but something about installing nvidia drivers via terminal just didn’t cut it for me! I’m willing to try another distro but this one seems ok and it mostly just works?
Also, most modern distros ship with nvidia drivers now that nouveau kind of became the official one.
But yea, i recommend you stick to big-name base distros like debian, ubuntu, arch (jk), suse or fedora.
They have a lot of derivatives, especially "easy for begginers" ones, but these are often less tested, with less users and documentation in cases you need help and have some premade configurations with dubious choices.
@heredos@heredos.net what’s your recommendation then?
Those are just my personal picks though, and i'd still personally pick fedora as it comes with a lot more stuff to make a first experience more enjoyable and less barebones.
@heredos@heredos.net isn’t fedora owned by red hat?
remember hearing that there was no straight flag that wasn't made by a bigot .. so ... i'm kinda tempted to do one tbh..?
@heredos best debugging is when you put *(void*)0 = 0; so that it generates a crash dump and figure out where the fuck it's failing
data science class 2: web scrapping.(the double p isn't even a joke on my side)
The #ACNHDictionary of Obscure & Imaginary Words is now live at https://acnhdictionary.carrd.co! Check out the link to read the flipbook online or download your own PDF copy. If viewing the flipbook on a phone, horizontal layout is recommended. Shares also greatly appreciated to spread the word about this free piece of fan literature collectively produced by the ACNH community. 🩵
@lucscozygames This is such a cute project. I wish I had participated along the way, but so love how it turned out!
THREE (3) PUBLIC TRANSPORTS
Tonight's Low Quality Ad is for this Seal Pet Bed. Go ahead and try to name a better duo than dogs and seals. Wait, Batman and Robin. Salt and pepper. Bonnie and Clyde. Okay, there's tons of them. But I think your dog would still love this seal bed.
https://collabs.shop/rlqrhb
Please don't get this if you have a pet seal though, that would be traumatizing for the poor lil guy.
@heredos did you use a bridge