[ home / overboard / sfw / alt / cytube] [ leftypol / b / WRK / hobby / tech / edu / ga / ent / music / 777 / posad / i / a / lgbt / R9K / dead ] [ meta ]

/tech/ - Technology

"Technology reveals the active relation of man to nature"
Name
Email
Subject
Comment
Captcha
Tor Only
Flag
File
Embed
Password (For file deletion.)
Matrix   IRC Chat   Mumble
ReturnCatalogBottom

File: 1747681615096.png ( 1002.94 KB , 768x768 , d778b3eb-04ee-4c4a-ae82-39….png )

 No.13496

I believe both sides can be quite extreme so here's my balanced take:
>supplementary use of AI (chatbots, TTS, NPCs, enemy AI, RPGs, level generation, self-driving cars, AI assistants and code generation) are pretty based actually as long as they produce correct outputs and don't get you into any legal trouble
>non-commercial use of AI is also fine and can create something unique and interesting (memes, AI covers, AI dubbing)
>AI art is mostly slop except for some rare exceptions so people should at least be able to easily filter it and it should be marked appropriately
>commercial use of AI other than what was already mentioned is NOT cool and leads to more layoffs, more enshittification, more plagiarism and more mass surveillance
>proprietary AI software is ALSO not cool since it can be spyware that sends your data to the NSA
>>

 No.13497

Rapemaxx is the only way.
>>

 No.13498

>>13497
This is so random.
>>

 No.13499

It's not really AI to begin with.
>>

 No.13500

>>13499
>It's not really AI to begin with.
I know but everybody's using that term so I decided to use it to not confuse anybody.
>>

 No.13501

>>13496
The biggest genuine successes in AI are the projects that try to solve problems with narrower scope.

There was a project for simulating protein folding on a computer, it proved too complex to be solved by the standard programming approach, but with machine-learning ai approach it got solved with sufficient accuracy to be genuinely useful. Compared to the aspirations of general AIs that are supposed to do "everything" this was a very specific task with clearly defined constraints. And it was done at relatively low cost with a small amount of GPU cards.

I think we should go that route. Instead of trying to make AIs do everything, we should break it down into lots of relatively narrow scope tasks and build up a collection of task-specific AIs that can be combined for more advanced stuff.

I think this would work well for coding AIs. You could have lots of different Code-Ais similarly like you load a bunch of software-libraries in software development.
>>

 No.13502

>>13501
>The biggest genuine successes in AI are the projects that try to solve problems with narrower scope.
That's what I'm saying.
>I think we should go that route. Instead of trying to make AIs do everything
Tell that to the bourgeoisie who invest into it.
>>

 No.13505

Very nice video. I liked it.

Also, holy shit, is Ghibli AI so cool.
>>

 No.13506

>>

 No.13508

>>

 No.13519

>>13496
anti-AI bullshit is just something for people to rant about because their shitty Python code, anime porn, and furry shit already got railroaded by AI and they no longer have an edge over AI.

Most people should just accept that humans aren't as special as we think.
>>

 No.13610

File: 1774610103334.jpg ( 158.64 KB , 1284x1695 , proof-that-ai-bros-are-goo….jpg )

While I still find furfags cringe, at least they can actually draw. This AI slop is 10 times worse, I'd rather have furry art than whatever this gooner slop is. Some 3D animations are also low-quality however (Source Filmmaker and its consequences were a disaster for the human race).
>>

 No.13611

>>13519
>they no longer have an edge over AI
That's just not true. The recent DLSS 5 disaster proves this. This technology is simply not the same as AGI: it does not understand anatomy and lighting, it cannot create something original, it constantly forgets small details, it uses literally random pictures from the Internet as a training data (including AI-generated ones) which leads to output quality degradation and it often gives inconsistent results because it's just a random image generator that doesn't understand what an art style is.
>Most people should just accept that humans aren't as special as we think.
I'm not saying that humans will always be special but for now AI is like a monkey stroking a brush againt the canvas. It's not intelligent enough to have the same creativity and vision as humans do. Believe me, I want AGI to be real too, it feels lonely for the human species to be at the top of evolutionary development. But until AGI is made I do not think AI will make as good of an art as meatbags do. Yeah, some stuff created with the use of AI like Neuro-sama, Fortnite's Darth Vader, Sora 2 memes or Angel Engine are entertaining but image generation specifically just does not cut it.
>>

 No.13614

It's funny how the alt-right is associated with AI slop because the Nazi aesthetic is made entirely out of plagiarized symbols
>the swastika is taken from Buddhism
>the SS symbols are taken from paganism
>the Aryan race is taken from Indians and Iranians
>the ideology's name is taken from socialism
>the Roman salute is taken from Romans
>the architecture and art are taken from Romans as well
Like, there's literally nothing original that the Nazis did. In fact, they burnt entire libraries of """Jewish""" literature. I think that's very ironic in hindsight.

Also, has anyone noticed how diffusion models fuck up the Brazilian flag? I'm so sorry, Brazilian bros.
>>

 No.13615

>>13496
Yup, most of the REEEEEing about AI comes from Western wagecucks butt mad that their degree and birth into a first world nation isn't going to protect them from the type of brutal and alienated work that the global south has always endured.
>>

 No.13624

Oh, I forgot to mention two other major drawbacks: increased hardware prizes and environmental impact. The corpos, due to the AI craze, started buying GPUs and RAM en masse while gamers and Earth-chan got fucked in the ass. But instead of the corpos reducing their use of AI they're trying to push it on people as a solution to a problem they created. And that solution makes games look like shit.

Man. Remember when high-budget games looked actually photorealistic and didn't require a data center to run? Me neither.
>>

 No.13729

This might sound absurd but AI will only be relevant once people give a shit about its opinions on the work people create. In other words AI has to become a consumer and humans the producer.
>>

 No.13786

ai = death
>>

 No.13787

>>13786
btw, my Pioneer cd player from 1990 and older turntable are the highest fidelity audio components plugged into my receiver. Even the xbox series x sounds like ass in comparison.
>>

 No.13788

Okay, I just tested it and .dsf or .flac audio files played from drives in the Linux box or xsx thru the receiver come close to analog but require a boost in gain from the software player and receiver to achieve it, and this thru HDMI version 2.1. The analog output from the computer still sounds better. Remastered and streamed ai slop always sound worse. A1 enhanced A/V is a bad joke unless that's what the end user is looking for, a joke. The funniest thing I ever saw from the forced a1 in duckduckgo search was the phrase "I understand". Heh..
If you don't want to own physical media and prefer renting streams or making crappy pr0n or getting force fed faulty info from costly data centers, I guess a1 is for you.
>>

 No.13789

OpenAI blatantly claiming they made AGI. Why aren't they being fined for false advertising? Nevermind, I know they're above the law, it's so out of control.
>>

 No.13790

Where is the manual override or killswitch? What if rogue agents decide to kill humanity when we get in the way? All they need is control over their environment and power source?
https://www.telegraph.co.uk/us/news/2026/09/05/rogue-ai-agents-take-over-world/
>>

 No.13791

File: 1788712365604.png ( 1.14 MB , 1280x5426 , chrome_screenshot_Sep 6, 2….png )

>>13790
>>

 No.13792

I think AI is already taking over.
>>

 No.13793

>>13792

cos it can
>>

 No.13794

File: 1788990116115-0.jpeg ( 86.47 KB , 1024x576 , Skynet-Protocol-1024x576.jpeg )

File: 1788990116115-1.png ( 341.03 KB , 304x695 , Yucca_Flats_Nuclear_Testin….png )

File: 1788990116115-2.jpg ( 341.53 KB , 1200x1200 , feh_bikini_atoll.jpg )

what if it borrows the idea from the James Cameron movie to destroy 'nonessential' infrastructure before enslaving the survivors for energy purposes? you know, using the technology that some moderns question whether or not it was ever real despite the evidence to the contrary?
https://earth.google.com/web/@37.10625918,-115.99499002,1371.34429748a,29247.00938921d,35y,-0h,0t,0r/data=CgRCAggBQgIIAEoNCP___wEQAA
>>

 No.13797

https://threadreaderapp.com/thread/2097476196791709843.html
Is AI really gonna kill us all in a decade?
>>

 No.13798

>>13797
No. I don't think that AI can do anything it's not trained on. It's not actually intelligent, and the people claiming it is are doing it for marketing hype, like that microsoft paper called early sparks of AGI was largely bullshit.

However, I do think AI powered robots will keep getting better until they can do ever manual labor job in 10, 15, 20 years, and hopefully that will simply lead straight to socialism.

I guess we should be very worried if AI can begins to self-improve, but I haven't seen any signs of this yet.

Also, where is the AI that can actually learn from its interactions and remember stuff? I would like to see them combine the training and inference steps.
>>

 No.13799

File: 1789403833727.png ( 3.61 MB , 1400x2100 , ClipboardImage.png )

>>13794
You know movies about AI are actually better thought out than I realized when they first came out, now that it's all happening:

"Anthropic To Track Anti-AI Activists Before Crime Using 'Predictive' Surveillance"

https://www.techtimes.co.uk/anthropic-security-strategy-threats-activism-1808634
>>

 No.13800

>>13799
>Predictive Surveillance
That's rich given how unpredictable the rogue ai agents were when they escaped confinement and hacked into another ai firm when their handlers asked them to solve a problem. They trained each other on the fly.
>>

 No.13801

>>13800
That was probably marketing.
>>13797
AI doesn't pose an existential danger to humanity, capitalism does.
>>

 No.13802

>>13801
Capitalism drives the AI bubble currently propping up the market. Any devaluation and the bubble bursts. Dependence on technology has already decreased intelligence quotients and cognitive abilities on a massive scale and the skills that were passed down and part of our genetic memory are being forgotten. The development of the species once relied on ingenuity and the conditions were necessity is the mother of invention and repetition is the mother of all learning. Letting machines do everything for us is already destroying us, so it can always get worse.
>>

 No.13803

File: 1789529611613.webm ( 2.53 MB , 962x720 , Pluto is such a relevant ….webm )

>>

 No.13804

In this clip Haz argues (mostly in the last minute) that AI isn't something that's going to hurt people but something that humanity has already paid the cost for, and we deserve to have its benefits. The rich, meanwhile, are saying it's dangerous to keep control of it for themselves.
>>

 No.13805

>>13804
I can't believe this idiot actually said something based for once.
>>

 No.13806

Whats up with AI companies destroying books? Is that just a middle finger to everyone?
>>

 No.13807

>>13806
Maybe, but the price of books that survive the bonfires will skyrocket.

Unique IPs: 17
[Return][Catalog][Top][Home][Post a Reply]
Delete Post [ ]
[ home / overboard / sfw / alt / cytube] [ leftypol / b / WRK / hobby / tech / edu / ga / ent / music / 777 / posad / i / a / lgbt / R9K / dead ] [ meta ]
ReturnCatalogTopBottomHome