[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology

Name
Options
Subject
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]

[Catalog] [Archive]

File: 1739485888068667.jpg (170 KB, 1536x782)
170 KB
170 KB JPG
240p to 1080p

It's over for the competition (there isn't any)

It's over for "Native resolution" (99% of Native res in games now is actually garbage forced TAA blurshit)

NVIDIA WON
25 replies and 3 images omitted. Click here to view.
>>
>>107852757
The edges of the pixels are really grainy on the nvidia side. I can still see where the AI model is hallucinating and making shit up. We need hardware to make the pixels that are coming from the data, not some retard shit that Jensen Wanker wants to make because their chips can't actually make what is there happen.
>>
>>107852757
>111 fps
>>
File: 4nbu3o.png (177 KB, 600x390)
177 KB
177 KB PNG
>>107852757
Is it purely upscaling what it sees plus some edge refinement or does it know how the sword is supposed to look at higher fidelity?
>>
>>107853077
looks like shit.
>>
>>107855637
as far as I know it's trained on images with ridiculously high resolutions like 8K or even beyond, then with AI magic the model learns what some pixel patterns normally should be like.

it works, just needs specialized hardware for it.

File: 61wbV8oqAbL(1).jpg (94 KB, 1500x698)
94 KB
94 KB JPG
Why do you need 24GB of VRAM? What will you do with it that 16GB can't? Most people don't need more than 12 and you NEET highschool dropout hobbyists can get by just fine with 8GB for your projects and old games.
124 replies and 12 images omitted. Click here to view.
>>
>>107842575
AI training experiments for my thesis. I have 32GB, and it's barely enough. Meh, I'll probably quit anyway, as everything I do just goes wrong in ways I couldn't even imagine.
>>
>>107855858
i always felt like the 5090 isnt fantastic for AI and its more of a gayming card because 32GB isnt nearly enough
>>
I can run pretty much any model with a 5070 ti and 64GB. might be slightly slower but its worth not having to replace my case and 750w PSU
>>
>>107855858
>AI training experiments for my thesis
isnt there a university supercomputer you can rent time on
>>
>>107856337
Yeah, there is. They have an HPC cluster.

File: DLSS WON.jpg (836 KB, 2973x1695)
836 KB
836 KB JPG
https://www.youtube.com/watch?v=NXAYR-ZtbH4

DLSS 4.5 btfo FSR4
181 replies and 13 images omitted. Click here to view.
>>
>>107848234
>implying that upscaling isn't slop
>comparing slop vs slop
This whole premise is fucking wrong and terrible. OP is a faggot.
>>
>>107856018
Normalized vs image quality, it's not
>>
>>107851027
this. shippers are shipping dog shit and claiming it's catshit but it's bullshit!
>>
>>107851081
interpolation is fake you imbecile

>>107851123
everyone cares you fucking dipshit. just because you are high as a kite and keep altering how reality looks to your brain doesn't mean everyone else wants to play games like their video card is on drugs hallucinating everything instead of calculating the shit its supposed to.
>>
>>107856112
Show me

File: file.png (233 KB, 2000x2000)
233 KB
233 KB PNG
/aicg/ - A general dedicated to the discussion and development of AI chatbots

Maybe The Real Treasure Was The Friends We Made Along The Way Edition

>News
Z․ai releases GLM-4.7 https://z.ai/blog/glm-4.7
Google releases Gemini 3 Flash https://blog.google/products/gemini/gemini-3-flash
OpenAI releases GPT-5.2 https://openai.com/index/introducing-gpt-5-2
Deepseek releases V3.2 https://huggingface.co/deepseek-ai/DeepSeek-V3.2
Anthropic releases Opus 4.5 https://www.anthropic.com/news/claude-opus-4-5
Google releases Gemini 3 Pro https://blog.google/products/gemini/gemini-3

Additional info: https://aicg.neocities.org/info.html

>Frontends

Comment too long. Click here to view the full text.
183 replies and 31 images omitted. Click here to view.
>>
holding a glass of wine to {{char}} and asking them to taste it, only to dump it on them after they give high praises and remark about vintage wine; accuse {{char}} of just trying to flatter you (narrate that it was actually cheap shit)
>>
>>>/v/730719497
>remind me again, why jannies nuking AI chatbot threads?
Because NAI shills keep taking over them.
See: >>>/v/730718596 >>>/v/730718814
>>
>>107856093
So, what did that update entail?
>>
>>107856437
NAIgods run this site btw

kneel, cuckie
>>
File: 1764227570108062.png (173 KB, 935x861)
173 KB
173 KB PNG
>>107856421
tested

File: 1754714232647000.png (8 KB, 500x500)
8 KB
8 KB PNG
Why did youngsters decide to ditch this website all of a sudden? It used to be THE social media website for the longest time but almost exclusively used by boomers these days.
53 replies and 3 images omitted. Click here to view.
>>
>>107852694
>used by boomers
You answered your own question.
>>
>>107852694
because it's hot garbage. it was fun around the late 2000s when it had those arcade / puzzle games like bejeweled, but as time went on it became more and more unusable
>>
>>107852694
Millennials ceased to be youngsters and became oldsters
>>
>>107852694
just like >>107856188 said. the millennial yuppies have become the new boomers
>>
>>107854538
yeah, this. lot of zoomies say they don't use it, yet they still have accounts and you can occasionally catch them online

File: huawei_matebook.mp4 (1.3 MB, 240x426)
1.3 MB
1.3 MB MP4
How come American companies can't innovate like this? We can't even afford ram because of Trump meanwhile their ram is 2/3rds cheaper than what prices were in America before tarrifs.
>>
>>107856482
Why do you retards love wasting your money on gimmicky bullshit that's just going to sit in a corner and collect dust?
>>
>>107856503
>be american
>get humiliated
technology for this feel?

File: 137366333_p3.jpg (2.3 MB, 2489x3613)
2.3 MB
2.3 MB JPG
hardworking Hina Edition

>Not sure what private trackers are all about?
A private tracker is an invite-only torrent website. Each member shares common goals: collecting, preserving and discussing media.

>Have a question?
- FAQ https://archive.is/UVQkn
- WIKI https://igwiki.lyci.de/wiki/Private_trackers
- NEWFAG PYRAMID https://inviteroute.github.io/graph or https://inviteroute.github.io/sheet/
- STUDY https://www.cambridge.org/core/services/aop-cambridge-core/content/view/2F379FE0CB50DF502F0075119FD3E060
- SPREADSHEET https://hdvinnie.github.io/Private-Trackers-Spreadsheet/
- TEN CURRY COMMANDMENTS https://pastebin.com/raw/dBbdE73M
- TEN NEON COMMANDMENTS https://pastebin.com/raw/Ud2pGYaE
- RED SPAMMER'S BIBLE https://rentry.org/69zbxh4h
- #ptg is on irc.sageru.org but it's pretty dead

Comment too long. Click here to view the full text.
209 replies and 25 images omitted. Click here to view.
>>
File: Untitled.jpg (61 KB, 1049x616)
61 KB
61 KB JPG
>>107855584
yea m8 it's in AB forums
tho seems pretty neglected
>>
>>107855781
Hide the usernames faggot
>>
File: wat.jpg (124 KB, 720x960)
124 KB
124 KB JPG
>two weeks left to use 71 tokens on red
what do
>>
>>107856473
/marked/
>>
>>107856473
Snatch all ILP slop

File: PCM30A.jpg (62 KB, 990x990)
62 KB
62 KB JPG
I've owned three different Pyle desktop amps that all have similar problems. All of them blast static when the volume knob is moved near the low end. Two of them have dead or partially dead right channels. Two of them have the grounding jackets pulled off the RCA connectors. All of them have the RCA connectors too close together for the backshells on a lot of RCA cables. Needless to say I'm done with Pyle. I really like their designs but their quality is shit.

What's the poorfag upgrade from Pyle that doesn't completely suck?

I want:

>smol (under 3.5" tall)
>RCA inputs
>banana jack outputs
>6.5mm headphone jack
>volume knob

That's it. No Bluetooth or USB any other modern bullshit.
>>
>>107855830
I don't know shit about shit but I heard class D amps are to prone to this so you should probably look for another class unless its specifically cleared of static issues

Also /g/ has hifi and headphone generals you could ask there

File: 1737127076050393.png (198 KB, 500x376)
198 KB
198 KB PNG
whats the best mouse ever?
32 replies and 5 images omitted. Click here to view.
>>
>>107855989
>he's never opened up a "quality" product and discovered pig iron weights inside
>>
>>107856035
>>
File: -ack.png (531 KB, 768x719)
531 KB
531 KB PNG
>>107855761
>>
>>107856072
I had a mouse that came with a box of weights and I didn't use any of them because it was already too heavy. Piece of shit started double clicking after a year or two anyway.
>>
>>107855340
its going to be the logitech superstrike coming out in a month

File: 1758010686171721.jpg (1.53 MB, 1298x2116)
1.53 MB
1.53 MB JPG
/lmg/ - a general dedicated to the discussion and development of local language models.

Previous threads: >>107847320 & >>107838898

►News
>(01/08) Jamba2 3B and Mini (52B-A12B) released: https://ai21.com/blog/introducing-jamba2
>(01/05) OpenPangu-R-72B-2512 (74B-A15B) released: https://hf.co/FreedomIntelligence/openPangu-R-72B-2512
>(01/05) Nemotron Speech ASR released: https://hf.co/blog/nvidia/nemotron-speech-asr-scaling-voice-agents
>(01/04) merged sampling : add support for backend sampling (#17004): https://github.com/ggml-org/llama.cpp/pull/17004
>(12/31) HyperCLOVA X SEED 8B Omni released: https://hf.co/naver-hyperclovax/HyperCLOVAX-SEED-Omni-8B

►News Archive: https://rentry.org/lmg-news-archive
►Glossary: https://rentry.org/lmg-glossary
►Links: https://rentry.org/LocalModelsLinks
►Official /lmg/ card: https://files.catbox.moe/cbclyf.png

Comment too long. Click here to view the full text.
>>
►Recent Highlights from the Previous Thread: >>107847320

--AI companion project updates and TTS technology debates:
>107850660 >107850672 >107850928 >107850947 >107853430 >107853443 >107853483 >107853554 >107853511 >107853575 >107853593 >107853616 >107853690 >107854122 >107854331 >107850983 >107851185 >107851216 >107851258 >107851258 >107853585
--Ouro-2.6B with claimed 12B performance and Scaling Latent Reasoning via Looped Language Models:
>107850350 >107850413 >107850458 >107850486 >107850501 >107850519
--Optimizing client-side beam search with logprobs and batching challenges:
>107848617 >107850562 >107853006 >107853060 >107853196 >107853678
--Tools and methods for converting EPUB books to audiobooks:
>107848869 >107848873 >107848915 >107849013 >107849252
--LLM roleplaying tool UI comparisons: RisuAI vs Talemate tradeoffs:
>107847698 >107847879
--Model-GPU compatibility and quantization tradeoffs for consumer hardware:
>107855431 >107855475 >107855507 >107855527 >107855540 >107855562 >107855603 >107855650 >107855676 >107855692
--Conflicting mmap behaviors in model loading across Windows/Linux platforms:

Comment too long. Click here to view the full text.
>>
S-so it's really here?
autoregressive imagegen for local? I'm not dreaming?
>>
>>107856463
Worse than zit despite being newer and bigger, basically a garbage
https://huggingface.co/zai-org/GLM-Image/discussions/1

File: 1767306970164.png (4 KB, 233x102)
4 KB
4 KB PNG
>We made the horizontal tab bar buttons arbitrarily more rounded and antialiased! to match the vertical tabs!!
I hate these meaningless pissant changes
7 replies and 3 images omitted. Click here to view.
>>
>>107854687
That'd be great, eh?
>Jones! How long to change the color of every blue button on our site to green?!
>"Seventeen months?"
>Perfect! You've done it again, Jones! See you next October when you've got all those buttons changed over! Keep up the good work!
>>
File: dropdown.png (27 KB, 640x360)
27 KB
27 KB PNG
>>107850850
I always to have a drop down button on the right side of each tab instead of the close button. This dropdown menu can have all sorts of tab related items like bookmark, pin, suspend etc. Make this button conver the entire right side of the tab like picrel.
>>
>>107853324
That's because most of them have to justify being employed.
>>
Not the most OFFENSIVE change I've had to endure but it is arbitrary.
>>
File: wemade.com.png (2 KB, 358x33)
2 KB
2 KB PNG
>>107850850
Just userchrome and forget.

File: file.png (6 KB, 1005x42)
6 KB
6 KB PNG
based or cringe?
8 replies omitted. Click here to view.
>>
>>107855895
>making a function for something thats only used once
bloat
>>
File: file.png (7 KB, 576x103)
7 KB
7 KB PNG
>>107854904
>>107855755
>>107855895
>>107855916
>>107855960
/thread
>>
>>107854904
hmm
> (A() and B()) or (C() and {})
...
> if A() { B() } else {C(); {}}
are these equivalent, or does the first one run C if B fails or something?
>>
>>107856077
yeah
only a good idea when the language allows a function to be nested inside another function
>>
>>107856422
They are not equivalent. C indeed depends on the return of B.
In simple words, and chain returns either the first false value or the last value (if all of them are true), or chain returns the first true value or the last value

File: deepseek.jpg (53 KB, 1200x1200)
53 KB
53 KB JPG
Is it safe to admit that Deepseek, MiniMax and Qwen are the actually good AI models yet? Or are we still going to pretend that ChatGPT, Gemini, Copilot and Claude count as "good" AI?
>>
gemini is god tier
deepseek works for my needs
don't even know what grok and openAI are so trash, they are focused on creating viral images i guess
meta ai is the worst by far though now that apple is out of the game
>>
I don't pay for slop but among the free options Gemini is shoulders and hips above anything else. In particular it's the most likely to have recent information and find new information at runtime and it understands intent much better. I don't think that last part is actually because they have a better architecture, I think they just haven't overcooked it as much as the others. It reacts more like a human and less like an HR redditor and seems to understand irony and oblique. Deepseek has been confirmed to be an inbred retrain of an old version of chatgpt since the beginning and I think they made it even worse since then by distilling it. With that said, Qwen is #1 in images by a lot, although so few enthusiasts have enough VRAM to play with it that it's still way under its ceiling.
>>
Depends on what you want from an LLM but the answer is a resounding no if you don't care about local and can actually afford the API costs.
Of the three Chink models listed MiniMax isn't even worth using I don't know why you bothered including it.
>>
>>107855325
Its good in that its like cheap. But other than that its shit. I only use it so I don't have to pay like $400 a month.
>>
redpill me in full vibe coding with these, I have money to throw at it, whats the cursor equivalent setup

File: output.webm (3.66 MB, 720x574)
3.66 MB
3.66 MB WEBM
Check out my voxel engine anons. Written in C++ using OpenGL. Infinite world in every direction, including up and down. Not even close to complete yet, but I'm proud of what I've done so far.
166 replies and 19 images omitted. Click here to view.
>>
>>107851629
>Luddite
You don't even know what a Luddite is, you're just seething at whatever you think they are because the internet told you to
>>
>>107852600
That's the so-called "vibe coding" for ya
>>
>>107846997
very nice work anon
>>
>>107855881
Thank you! I promise I'll work on the performance :)
>>
>>107846997
>I'M proud of what I'VE done so far
lmao you haven't done anything. Sam altman should be getting the credit here.

File: 1696342843575.jpg (335 KB, 1920x1080)
335 KB
335 KB JPG
>>107601582
"The further a society drifts from the truth, the more it will hate those that speak it."
--George Orwell

>Cyberpunk
The FAQ: https://sizeof.cat/post/cyberpunk-faq/
What is /cyb/erpunk?: https://pastebin.com/pmn9vzWZ
How do I into /cyb/erpunk?: https://pastebin.com/5tpNFQds
Huge list of cyberpunk media: https://sizeof.cat/post/cyberpunk/
The cyberdeck: https://pastebin.com/7fE4BVBg
Cyberlife: https://jinteki.industries/files/cyberlife.7z
Bibliothek: https://www.mediafire.com/folder/4m5hd2065hde8/Bibliothek

>Privacy
Tools: https://www.privacyguides.org/en/tools/

Comment too long. Click here to view the full text.
192 replies and 31 images omitted. Click here to view.
>>
How are your hacking plans going, anon?
>>
>>107855954
Marvelous. Hacked my mom into giving me two plates of tendies instead of one. My power grows.
>>
>>107856100
Tonight I will be having tendies and velveeta shells and cheese with a dr pepper. I have to make it myself though.
>>
>>107856115
Godspeed anon. Enjoy your cooking.
>>
File: 1721425851098.png (893 KB, 1000x667)
893 KB
893 KB PNG
RIP based Dilbertman
https://news.slashdot.org/story/26/01/13/1619251/scott-adams-creator-of-the-dilbert-comic-strip-dies-at-68


[Advertise on 4chan]

Delete Post: [File Only] Style:
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.