Ai Flux

1枚600WのGPUを3枚装備した

Testing, I swear x og card cache will be the death of me localmaxxing.com/en/marketplace…

@d33v33d0 What happened??!

@pupposandro You'd need to add 5KW of air conditioning to work in the same room with this beast lmao

Working hard to be able to buy one of these at some point in the future 7x RTX PRO 6000. Custom liquid cooled system, in a small under desk box.

Bros qwen 3.7 27b Coming soon Lets goooooo

Undervolt your 3090s right now!

THIS GUY SAW A $430 AI BILL AND BUILT HIS OWN AI LAB UNDER HIS DESK INSTEAD RTX 5090 + RTX 4090, 56GB VRAM, 128GB RAM, Proxmox and local Qwen / DeepSeek / Llama models running without API keys while everyone else is still paying every time they test a prompt. The best part of the setup: api_key: “not-needed”. His agents can scan GitHub, Reddit and RSS feeds, read notes, test ideas overnight and break without turning into another invoice. If something fails, he fixes the config, not the credit card limit. Most people rent AI by the token. He is turning a desk setup into a private machine that works even when the dashboard is closed.

⚡ Luce Megakernel just proved the NVIDIA efficiency gap is a software problem not a hardware one 🔬 a 2020 RTX 3090 at 220W now matches Apple M5 Max efficiency and delivers 1.8x the throughput 🔹 413 tok/s decode vs 267 tok/s on llama.cpp — same GPU, different software 🔹 1.87 tok/J — matching Apple M5 Max at less than a third of the system cost 🔹 All 24 layers of Qwen3.5-0.8B fused into a single CUDA kernel — zero CPU round trips 🔹 25x faster than PyTorch HuggingFace on the same hardware 🔹 Hybrid DeltaNet and Attention architecture — the first megakernel ever built for this pattern 🔥 Full breakdown and live benchmark below 👇 youtu.be/e6jY4goVIu0

We open-sourced the chassis design for the 8-GPU build. STEP files for every part. You can now build your own On-Premises Business AI Center. Why? - Protect your IP - Protect customer data - Save on inference costs - Train your own models Open source AI hardware is the way!

First contribution by Manuel Burgschachner, he has just setup his 8-node GB10 (Spark) Cluster before the full rack was mount and had a quick run of zai-org/GLM-5.1-FP8 (756 GB) using vLLM. View full benchmark at spark-arena.com/benchmark/sub1…

@isnit0 What's the GPU mix?

My brother-in-law runs his own data center/GPU farm at home. Currently makes him ~$3k/month profit. 30% solar powered. Aiming to get to 60% by end of this summer which will increase profits significantly (mostly down to adding batteries, not panels), then will buy more GPUs. He sells the compute via some 3rd party service that takes a cut. Apparently a lot of ML/rendering workloads. You can just do stuff!

@MegaSparked We'd love to join a class action if you decide to go that route - our channel was demonitized last week.

We are going to sue YouTube But before we do, our lawyer is currently in the process of sending a letter to YouTube. Over 2 weeks ago our channel @UnderSparked" target="_blank" rel="nofollow noopener">youtube.com/@UnderSparked was falsely demonetized under the "inauthentic content" policy. Demonetized doesn't mean no ads though, all it means is YouTube will still run ads on the videos. But they pocket all the revenue, essentially stripping all the revenue from us. We don't claim to be the pinnacle of quality content standards, but the following things do cause us to scratch our heads in confusion 1. All of the bigger channels in the same niche as us are monetized with no issues 2. We literally voice act, edit, etc, all our own videos while 100s of AI storytelling channels are still monetized 3. YouTube is fine with running ads on the videos, pocketing all the revenue. We assume if YouTube had such a big problem with the content, why still serve it to your advertisers? 4. 100s of reaction channels still stay monetized Some examples of bigger monetized channels that are in the same niche as us are: rSlash: @rSlash" target="_blank" rel="nofollow noopener">youtube.com/@rSlash DarkFluff: @DarkFluff" target="_blank" rel="nofollow noopener">youtube.com/@DarkFluff Oz's Vault: youtube.com/channel/UCoJ1A… We support these channels, the reason we are providing these channels as examples is to showcase how it feels like we're being targeted. This has caused extreme distress for my team and I. Many of our livelihoods and families rely on this channel. Our community relies on the UnderSparked channel. Many people support the UnderSparked channel because in a time of nonstop AI channels, we stepped up and used real voice actors to cater to that connection and community that viewers on YouTube crave. We truly believe this demonetization was done in error, all we want is a real human specialist to help us fix the issue. We're tired of constantly receiving only copy and pasted templated responses, constantly neglecting/dismissing our attempts at fixing the false demonetization. We're kindly requesting for a representative from @TeamYouTube to help us fix this demonetization issue with our channel. We don't want to send letters, or sue, or anything! We simply just want to go back to producing content for our community.

@mossab12 This happened to my channel last week - curious if you've had any luck getting connected with a @TeamYouTube partner?

YouTube DEMONETIZED my 8 year old channel and took $8,000 from me... THIS MADNESS HAS TO STOP They called my content "Inauthentic content." I am LITERALLY creating my own content. I spend hours researching from real sources: books, news reports, articles, and more. I write the scripts myself, edit the videos, and design every thumbnail by hand. YouTube saw all of that and said MASS PRODUCED. I do not want another bot response. I want a REAL person to sit down and actually watch what I make and my APPEAL VIDEO. Anyone who does will immediately see that this decision makes NO SENSE. For anyone facing the same “inauthentic content” madness, let’s unite and go VIRAL. YouTube has to listen to us. PLEASE SHARE THIS. @TeamYouTube @YouTubeCreators @YouTube

@RPHManime @TeamYouTube Hey, recently the same thing happened to my channel. Curious if you've had any luck resolving this with @TeamYouTube ?

Hey @TeamYouTube My channel was incorrectly flagged for “inauthentic content” on April 27. I tried appealing twice but every time i did my appeal got rejected by the automated system, without any human review. I don't make Ai generated videos, i don't even use templates to mass produce or reupload the same video with minimal changes, i make anime related content with my original stories, all of my videos have: - unique ideas - different & and unique scripts - edited by me. - Visuals and edits I make myself in Photoshop + Canva In the 1:39 video attached, i showed my content creation process. So please, I respectfully request a manual human review of my channel. I’ve spent the last 3 years building this channel and audience through original work, and I would truly appreciate another review. This is my channel: @rphmanime?si=Sd2rCBbd5JD58cX6" target="_blank" rel="nofollow noopener">youtube.com/@rphmanime?si=… Case id: 7-5118000041125 & 1-9196000040580 Please YouTube, I only ask for one human review of my content.

I have been tracking the trajectory of hardware my entire career. Sometimes there are advances that are extremely impressive! I just received my OpenMV AE3 this week and it is a stunning evolution of hardware, connectivity, and AI! Prior to buying this device, I had purchased a Milk V Duo S which blew my mind. for a total of $25 I was running an embedded Linux device with a camera and Ethernet where I could run YOLO models, perform object recognition, and report to the cloud via MQTT. This device cost $90, can also run YOLO models (and more), includes a MicroPython development environment and also includes WiFi and Bluetooth! Yes, triple the price but there is also an extremely active developer and support community. Amazing! —- OpenMV AE3 ow.ly/hEOR50YZxR5

I tested Mighty’s VIO against Meta’s LaMAria benchmark. It’s did well actually, better than other usual open-source ones (see leaderboard) specially considering Mighty is mono-imu! Keep in mind, this same VIO runs *unchanged* on the Mighty Camera apart from some minor config changes. Aria MPS is still goat though!

On-Premise Business AI Center After my posts on the 2-GPU and 4-GPU builds, people reached out asking how to build an 8-GPU box for their businesses. Why? - Protect their IP - Protect customer data - Save on inference costs - Train their own models Here's how to build one: 🧵

@ChloeLuong_ @autonomous_labs lamp?

Indie hackers, your next stop in Vietnam: Ho Chi Minh City (Saigon). Come visit @autonomous_labs. We're shipping AI hardware, everyday, 24/7, right next to the airport. Desks and chairs, plenty. Prototypes, lining up in our lab. And good coffee, always filled. Ping me!

左の写真はケースFANの配置は理解できるけど、 CPUの排熱をケースへ取り込んでる理由が不明( ๐_๐)