Mouse&Keyboard

CPU usage heatmap comparing "Heterogeneous short running thread scheduling policy" on a 7800X3D in Cyberpunk 2077. "All Processors" vs "Prefer Performant Processors"

@ServeTheHome No price anywhere.

The NVIDIA DGX Station systems have undergone a major change since their announcement, but they are starting to ship. We have more details from GTC this week servethehome.com/nvidia-dgx-sta…

I’m giving away a copy of Crimson Desert for PC (Steam) To enter, simply follow these steps: 1. Make sure you’re following me @NikTek 2. Reply to this tweet and you’re in! The winner will be randomly selected in 24 hours. Good luck to everyone! Stay tuned for the winner announcement and future giveaways 🤎

@totoche Moi je l'utilise et ça marche bien, bon je fais que du social media, mais franchement ca me sauve des dizaines d'heures.

Tous ceux qui me bassinaient avec OpenClaw il y a 1 mois. "T'es en retard", "tu dors", "c'est le futur". Un mois plus tard, silence radio. Plus personne n'en parle mdr

@LibertyPillMeme If this is DLSS 5, I buy it.

@sudoingX Meta: "ok let's fire him"

this guy has 29 models on huggingface at page 2 ranking. no lab behind him. no sponsorship. $2,000 from his own pocket on GPU rentals. he compressed GLM-4.7 to run on a MacBook and quantized Nemotron Super the week it dropped. all public. all free. nvidia is a trillion dollar company with hundreds of teams but they are not the ones quantizing models middle of the night and pushing them out before sunrise. if nvidia stopped tomorrow their employees stop working. people like @0xSero would not. that is the difference between a paycheck and a mission. @NVIDIAAI you talk about making AI accessible. the people actually doing it are right here. 29 models deep burning their own compute with no ask except more hardware to keep going. you do not need to build another program. just look at who is already building for you. one GPU to this man would produce more public value than a hundred internal sprints. i am not asking for charity. i am asking you to invest in someone who already proved it.

Meta annonce 16 000 licenciements. En interne, on parle plutôt de 22 000. Depuis 8 mois, les équipes documentent tout : décisions, debug, archi… Officiellement pour “mieux partager la connaissance”. En réalité, ces données ont servi à entraîner des agents IA internes. Aujourd’hui, ces systèmes gèrent déjà ~70% des modifs de code. La consigne côté management est simple : garder les “AI multipliers”… et virer les autres, considérés comme des “legacy dependencies”. Le nouveau modèle ? 1 senior par équipe + IA + offshore sous Cursor à 28k$/an vs 240k$ dans la Bay Area. Le calcul est vite fait. Les équipes les plus “AI-first” livrent 3x plus vite avec 60% de personnes en moins visiblement. Le plus ironique dans l’histoire ? Ceux qui sont virés en premier sont souvent ceux qui ont construit ces outils. Deux ans à automatiser leur propre job. Le futur du travail est brutal... On reviendra sur ces licenciements semaine prochaine dans Silicon Carne.

If you think AI film can’t be art then explain this.

"All Processors" seems to hit a tiny bit more L3 Cache, and tiny bit less DRAM. Them it is a tiny bit better in frametimes. And thanks to @CapFrameX for the bench software of course! CPU Heatmap: github.com/HumbleUser33/C…

CPU usage heatmap comparing "Heterogeneous short running thread scheduling policy" on a 7800X3D in Cyberpunk 2077. "All Processors" vs "Prefer Performant Processors"

🚨BREAKING: SUPER MICRO CO-FOUNDER ARRESTED FOR SMUGGLING $2.5B IN NVIDIA GPUs TO CHINA >SMCI co-founder Yih-Shyan "Wally" Liaw arrested today >personally holds $464 MILLION in SMCI stock >charged with smuggling BILLIONS in Nvidia servers to china >used a southeast asian shell company to funnel $2.5B in servers to chinese buyers >$510 million worth shipped in just THREE WEEKS in spring 2025 >built thousands of fake dummy servers to fool U.S compliance auditors >caught on surveillance camera using a HAIR DRYER to swap serial number stickers >coordinated the whole thing over encrypted group chats >SMCI down 12% after hours >faces up to 30 years in federal prison ITS SO OVER…

@0xSero Be carefull 23% hygrometry causes static electricity hazard. You should be more at 60%

Putting out a wish to the universe. I need more compute, if I can get more I will make sure every machine from a small phone to a bootstrapped RTX 3090 node can run frontier intelligence fast with minimal intelligence loss. I have hit page 2 of huggingface, released 3 model family compressions and got GLM-4.7 on a MacBook huggingface.co/0xsero My beast just isn’t enough and I already spent 2k usd on renting GPUs on top of credits provided by Prime intellect and Hotaisle. ——— If you believe in what I do help me get this to Nvidia, maybe they will bless me with the pewter to keep making local AI more accessible 🙏

The performance difference enabling Ray Reconstruction is brutal on Crimson Desert, goes from ~80s fps with 4K DLAA to ~55 with DLSS on Quality They must be enabling higher quality RT settings when RR is on, there is no way otherwise

"Most people bashing DLSS 5 are on the peak of ignorance" Veteran game artist shows exactly how much of a difference lighting can make. 🔗 wccf.tech/1k1l4

PewDiePie just exposed what's happening to our brains. The 111M subscriber YouTuber says we've "industrialized nothingness" and reveals the cheat code to escape. This isn't just another tech rant. It's a wake-up call. All you need to know 🧵⬇️

@NikTek Well he is right despite the haters army that did not understand that it was not just a 2D post processing.

Jensen Huang has responded to the comments calling DLSS 5 AI slop: "First of all, they're completely wrong. The reason for that is because, as I have explained very carefully, DLSS 5 fuses controllability of the geometry and textures and everything about the game with generative AI. It’s not post-processing, it’s not post-processing at the frame level, it’s generative control at the geometry level." He has also added: "This is very different than generative AI; it’s content-control generative AI. That’s why we call it neural rendering" What are your thoughts on this?

@KimNoel399 Quand t'es journaliste tech FR vaut mieux faire du négatif pour racoler toute la communauté tech / gamer ultra toxique de France.

La différence de point de vue FR et US du DLSS 5 est dingue. D'un côté DLSS 5 est catastrophique, de l'autre DLSS 5 est une technologie incroyable.

Intel Panther Lake Unified Memory architecture Intel Core Ultra X7 358H has a memory bandwidth of 200+ GB/s Both CPU and iGPU can access 96GB LPDDR5X 9600 MT/s Source: youtu.be/X9eYQTkzxqU?si…

Well, that DLSS 5 could be very interesting... #DLSS5

@mkurman88 This is because your GPU mem controler is also saturated (not only GPU compute load). That's why I tell you to check for mem chip temp (memory junction).

@PurePlayerPC Haha, you should hear the sound it made; it normally operates at a maximum of 77 during heavy load.

I sent 5 concurrent requests to Qwen3.5 27B and almost burned my card. from 55 to 89 °C in seconds, and RTX 3090 begged for help