Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Building another Autonomous Computer with 2x RTX PRO 6000 Blackwell. This is the ideal personal AI datacenter: 192 GB of VRAM on your desk. We're seeing ~300 tok/s with DeepSeek V4 Flash running locally. Own your compute. Own your intelligence.

Dee

5,697 subscribers

31,413 views • 7 days ago •via X (Twitter)

Science & Technology Education

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

384GB of VRAM. On your desk. Not in a datacenter. Autonomous, the company behind SmartDesk, just open sourced complete build guides for a personal AI computer that runs open models on hardware no one can switch off. It's called Autonomous Computer. Three configurations, one philosophy: a model you rent can be cut off overnight. A model running in your own house can't. → Home build: 2x RTX 5090 workstation → Team build: 4x RTX PRO 6000 Blackwell, 384GB VRAM → On-prem business build: 8x RTX 4090/5090, up to 256GB VRAM Every config ships with the full bill of materials, 3D printable and CNC housing files, wiring diagrams, BIOS tuning, and assembly photos. Software side covers OS setup, NVIDIA drivers, and serving open models through Ollama, vLLM, and llama.cpp. Fork it. Change it. Build it. Sell it. MIT License. 100% Opensource. Repo:

384GB of VRAM. On your desk. Not in a datacenter. Autonomous, the company behind SmartDesk, just open sourced complete build guides for a personal AI computer that runs open models on hardware no one can switch off. It's called Autonomous Computer. Three configurations, one philosophy: a model you rent can be cut off overnight. A model running in your own house can't. → Home build: 2x RTX 5090 workstation → Team build: 4x RTX PRO 6000 Blackwell, 384GB VRAM → On-prem business build: 8x RTX 4090/5090, up to 256GB VRAM Every config ships with the full bill of materials, 3D printable and CNC housing files, wiring diagrams, BIOS tuning, and assembly photos. Software side covers OS setup, NVIDIA drivers, and serving open models through Ollama, vLLM, and llama.cpp. Fork it. Change it. Build it. Sell it. MIT License. 100% Opensource. Repo:

Harman

47,954 views • 1 month ago

Your gaming PC can now serve frontier models at interactive speed using official checkpoints without extreme quantization! Qwen3.6 35B → 8GB RTX 4060 laptop @ 39 tok/s DeepSeek-V4-Flash 284B → RTX 5090 desktop @ 22-25 tok/s GLM-5.2 753B → RTX PRO 6000 workstation @ 15 tok/s Run your claude code or codex now with frontier model for $0 Meet FreeToken 🧵

Your gaming PC can now serve frontier models at interactive speed using official checkpoints without extreme quantization! Qwen3.6 35B → 8GB RTX 4060 laptop @ 39 tok/s DeepSeek-V4-Flash 284B → RTX 5090 desktop @ 22-25 tok/s GLM-5.2 753B → RTX PRO 6000 workstation @ 15 tok/s Run your claude code or codex now with frontier model for $0 Meet FreeToken 🧵

Shuo Yang

1,054,057 views • 3 days ago

Laguna S 2.1 running native NVFP4 Picked Blackwell hardware for this side-by-side: 🟢 DGX Spark | 19.44 tok/s | 162ms TTFT 🔵 RTX PRO 6000 | 108.54 tok/s | 266ms TTFT 🟣 4× RTX 5090 (TP4+EP4) | 145.81 tok/s | 521ms TTFT 117.6B MoE (8.5B active), quantized to FP4, on your own box.

Laguna S 2.1 running native NVFP4 Picked Blackwell hardware for this side-by-side: 🟢 DGX Spark | 19.44 tok/s | 162ms TTFT 🔵 RTX PRO 6000 | 108.54 tok/s | 266ms TTFT 🟣 4× RTX 5090 (TP4+EP4) | 145.81 tok/s | 521ms TTFT 117.6B MoE (8.5B active), quantized to FP4, on your own box.

stevibe

23,102 views • 1 month ago

Timelapse #177 (14 hrs) - Picked up an RTX Pro 6000 Blackwell and wired in another 1000W PSU - running two of them now - DeepSeek v4 Flash running locally on both GPUs - Distilling DeepSeek v4 Flash down into Nanbeige (still cooking) - Kernel bench runs for DeepSeek v4 Flash and Qwen 3.8 Max - Got it into chat but couldn't reliably get it into the OMP harness yet - Private contract work - days like this are stuff I actually want to build. and when it's the high-intensity grind, i still have to show up and get it done

Timelapse #177 (14 hrs) - Picked up an RTX Pro 6000 Blackwell and wired in another 1000W PSU - running two of them now - DeepSeek v4 Flash running locally on both GPUs - Distilling DeepSeek v4 Flash down into Nanbeige (still cooking) - Kernel bench runs for DeepSeek v4 Flash and Qwen 3.8 Max - Got it into chat but couldn't reliably get it into the OMP harness yet - Private contract work - days like this are stuff I actually want to build. and when it's the high-intensity grind, i still have to show up and get it done

Elliot Arledge

17,401 views • 20 days ago

Fluxgym: Train FLUX LoRA LOCALLY with LOW VRAM FLUX LoRAs are super exciting, but there has not been an easy way to train them LOCALLY using LOW VRAM machines. Fluxgym is a Gradio app for training your own Flux LoRAs on your own 12G, 16G, 20G+ VRAM computer for free.

Fluxgym: Train FLUX LoRA LOCALLY with LOW VRAM FLUX LoRAs are super exciting, but there has not been an easy way to train them LOCALLY using LOW VRAM machines. Fluxgym is a Gradio app for training your own Flux LoRAs on your own 12G, 16G, 20G+ VRAM computer for free.

cocktail peanut

237,764 views • 1 year ago

DeepSeek released R1, the best open-source reasoning model that is on par with OpenAI's o1! I have built a RAG app to chat with your PDF files using the DeepSeek R1 model, running locally on your computer. 100% Open Source

DeepSeek released R1, the best open-source reasoning model that is on par with OpenAI's o1! I have built a RAG app to chat with your PDF files using the DeepSeek R1 model, running locally on your computer. 100% Open Source

Sumanth

678,727 views • 1 year ago

Introducing PulseAI, where your data is your knowledge. PulseAI Platform is an agent-to-silicon AI infrastructure platform built to keep your data secure and sovereign, enabling the next AI enterprise. Offered in form factors for every stage of enterprise AI, PulseAI enables enterprise from initial pilot to hyperscale. Combining Gruve's AI-native services with NVIDIA-powered GPU architectures, it scales from 16 RTX 6000 PRO to 2,048 B300 GPUs, with security, storage, and compute integrated. The next enterprise is built on its own data and intelligence. We are here to enable it. Book your consultation: Read more on PulseAI from our Chief AI & Technology Officer Matt Robinson:

Introducing PulseAI, where your data is your knowledge. PulseAI Platform is an agent-to-silicon AI infrastructure platform built to keep your data secure and sovereign, enabling the next AI enterprise. Offered in form factors for every stage of enterprise AI, PulseAI enables enterprise from initial pilot to hyperscale. Combining Gruve's AI-native services with NVIDIA-powered GPU architectures, it scales from 16 RTX 6000 PRO to 2,048 B300 GPUs, with security, storage, and compute integrated. The next enterprise is built on its own data and intelligence. We are here to enable it. Book your consultation: Read more on PulseAI from our Chief AI & Technology Officer Matt Robinson:

Gruve

66,180 views • 2 months ago

New LFM2.5-2.6B hits DeepSeek-V4 level on tool calling and runs 3.7x faster! We ran Liquid AI 's new LFM2.5-2.6B against DeepSeek-V4-Flash on one box with 4x RTX 5090. Both got the same three jobs, and each one only completes if the model fires every tool call Topics: -weather and local time in six cities -one budget into six currencies -four hotels checked and booked for one date Outputs: LFM2.5-2.6B: 35/35 tool-calls, 366 tok/s, 19s DeepSeek-V4-Flash: 35/35 tool-calls, 77 tok/s, 70s Both models made all 35 calls, and both fired twelve of them in one turn. LFM was trained for agents, and tool calling is where that shows. Strong result for a model you can run right on your phone

New LFM2.5-2.6B hits DeepSeek-V4 level on tool calling and runs 3.7x faster! We ran Liquid AI 's new LFM2.5-2.6B against DeepSeek-V4-Flash on one box with 4x RTX 5090. Both got the same three jobs, and each one only completes if the model fires every tool call Topics: -weather and local time in six cities -one budget into six currencies -four hotels checked and booked for one date Outputs: LFM2.5-2.6B: 35/35 tool-calls, 366 tok/s, 19s DeepSeek-V4-Flash: 35/35 tool-calls, 77 tok/s, 70s Both models made all 35 calls, and both fired twelve of them in one turn. LFM was trained for agents, and tool calling is where that shows. Strong result for a model you can run right on your phone

atomic.chat

112,171 views • 19 days ago

EVERYONE IS BUILDING CUSTOM GPTS BUT ALMOST NOBODY IS TRAINING THEIR OWN MODEL FROM SCRATCH $100 and one weekend is all it takes with andrej karpathy’s nanochat you pretrain a gpt-2 grade base model on 400 billion tokens of web text, then fine-tune it on your own obsidian notes, your ai chat history, your writing it answers in your voice, recalls your own ideas, runs locally with no api and no subscriptions complete builder’s guide with every command in the article ↓

EVERYONE IS BUILDING CUSTOM GPTS BUT ALMOST NOBODY IS TRAINING THEIR OWN MODEL FROM SCRATCH $100 and one weekend is all it takes with andrej karpathy’s nanochat you pretrain a gpt-2 grade base model on 400 billion tokens of web text, then fine-tune it on your own obsidian notes, your ai chat history, your writing it answers in your voice, recalls your own ideas, runs locally with no api and no subscriptions complete builder’s guide with every command in the article ↓

leopardracer

15,859 views • 1 month ago

Anyone with 8GB or 12GB VRAM setups needs to understand that "-ncmoe" is the key flag to boost performance on llama.cpp Here are my results for Qwen3.6 35B A3B, with 64k q8_0 context on a 8GB RTX 3070Ti: ⚪️ no flag → 8.7 tok/s RAM: 13.6GB & VRAM: 7.8GB 🔴 -ncmoe 35 → 27.5 tok/s RAM: 12.1GB & VRAM: 4.3GB 🟢 -ncmoe 30 → 32.5 tok/s RAM: 12GB & VRAM: 5.6GB 🔵 -ncmoe 25 → 40.9 tok/s RAM: 12GB & VRAM: 6.9GB Please note the ram and vram usage you see are total usage of a windows pc, with the model running. My friend's setup: 8GB VRAM and 16GB RAM. You can boost performance by switching to Linux, just something to keep in mind. Basically, this flag keeps the MoE experts in the first X layers on your CPU + RAM, instead of eating all your VRAM straight away. This is a smart hybrid offload way that lets you run bigger models without OOM while keeping the rest on your GPU for speed. As we can see on the data, there's a sweet spot. When we lower it from 35 to 25, speed bumps +50% because there are more layers on your GPU (look at the VRAM usage). The key here is to play around with the number and fit as much as possible on your VRAM, goal is to have 1GB/800MB headroom to avoid stress. ↓ server flags below

Anyone with 8GB or 12GB VRAM setups needs to understand that "-ncmoe" is the key flag to boost performance on llama.cpp Here are my results for Qwen3.6 35B A3B, with 64k q8_0 context on a 8GB RTX 3070Ti: ⚪️ no flag → 8.7 tok/s RAM: 13.6GB & VRAM: 7.8GB 🔴 -ncmoe 35 → 27.5 tok/s RAM: 12.1GB & VRAM: 4.3GB 🟢 -ncmoe 30 → 32.5 tok/s RAM: 12GB & VRAM: 5.6GB 🔵 -ncmoe 25 → 40.9 tok/s RAM: 12GB & VRAM: 6.9GB Please note the ram and vram usage you see are total usage of a windows pc, with the model running. My friend's setup: 8GB VRAM and 16GB RAM. You can boost performance by switching to Linux, just something to keep in mind. Basically, this flag keeps the MoE experts in the first X layers on your CPU + RAM, instead of eating all your VRAM straight away. This is a smart hybrid offload way that lets you run bigger models without OOM while keeping the rest on your GPU for speed. As we can see on the data, there's a sweet spot. When we lower it from 35 to 25, speed bumps +50% because there are more layers on your GPU (look at the VRAM usage). The key here is to play around with the number and fit as much as possible on your VRAM, goal is to have 1GB/800MB headroom to avoid stress. ↓ server flags below

left curve dev

166,478 views • 3 months ago

We’re building the future of the Personal Computer for macOS—AI that’s truely your digital assistant. Access to your filesystem, always on, deep integration with your native OS and apps. So proud of the team! Excited to bring this vision to the masses 🙌

We’re building the future of the Personal Computer for macOS—AI that’s truely your digital assistant. Access to your filesystem, always on, deep integration with your native OS and apps. So proud of the team! Excited to bring this vision to the masses 🙌

Phil LaPier

42,562 views • 5 months ago

Deepseek V4 Flash 0731 (Q2) - 12 tokens/sec - Single RTX 4090 - 650+ tokens/sec prefill - 250k context - no kv cache quantization! DeepSeek just dropped the official V4 Flash 0731 two days ago with a massive agent capabilities upgrade. The official benchmarks are literally crushing their own V4-Pro-Preview on agentic tasks like Terminal Bench 2.1 and DeepSWE. Unsloth AI said they couldn't wait to bring it to local devices, and they delivered. If you thought my 118B Poolside Laguna S 2.1 MoE run last week on a single GPU was wild, hold onto your hardware. I just successfully ran Unsloth’s brand new 91GB DeepSeek-V4-Flash-0731 (UD-IQ2_M) GGUF entirely locally. And I pushed it to a mind-bending 250,000 context window. The VRAM ceiling is an illusion if you know how to optimize llama.cpp. Here are the benchmarks and the cheat codes to run a local frontier class model yourself. For the hardware and setup, I used a single NVIDIA RTX 4090 (24GB VRAM) hooked up via a PCIe 4 bus, running Ubuntu 22.04 LTS and CUDA 13.0. You don't need a massive enterprise server for this, if you have more than 80 GB of standard DDR4 RAM and a 24GB card like an RTX 3090 or 4090, you can run this exact stack yourself. All benchmarks were run using a massive 28k token prompt to truly stress test the prefill limits. no kv cache quantization THE BENCHMARKS (Scaling Context): # 80k Context (Baseline: -b 2048 -ub 2048): Prefill: 465.43 t/s | Decode: 13.00 t/s | VRAM: 22.87 GB # 80k Context (Optimized: -b 4096 -ub 4096): Prefill: 643.15 t/s | Decode: 12.20 t/s | VRAM: 23.00 GB (Notice how doubling the batch flags spiked my prefill throughput by nearly 200 t/s with almost zero VRAM penalty) # 180k Context (-b 4096 -ub 4096): Prefill: 629.18 t/s | Decode: 11.92 t/s | VRAM: 23.40 GB # 250k Context MAXIMUM (-b 4096 -ub 4096): Prefill: 619.02 t/s | Decode: 11.54 t/s | VRAM: 23.40 GB # THE SECRET SAUCE (Why this works): Unsloth’s UD-IQ2_M quant is ~91GB across 3 files. Since I only have 24GB of VRAM, the PCIe 4 bus and system RAM have to do the heavy lifting. The magic bullet is the --no-mmap flag. By completely bypassing OS disk paging, I forced llama.cpp to load the massive model weights directly into the system RAM upfront. Combined with Flash Attention (-fa on) and exactly 12 CPU threads (--threads 12), I maintained an incredibly stable 11.5+ tokens/sec decode speed even at a quarter million token context. # THE EXACT COMMAND: ./build/bin/llama-server -m /workspace/models/DeepSeek-V4-Flash-0731-UD-IQ2_M-00001-of-00003.gguf -c 250000 -fa on --port 8080 --threads 12 -b 4096 -ub 4096 --no-mmap -v Local conversational and agentic coding AI is fully here. You don’t need an API or an H100 cluster. Qwen 3.8 27b drops next week making the 24GB VRAM tier even more worthwhile. What does your current local AI rig look like, and what's the craziest model you've managed to squeeze into it? Official huggingface GGUF links from Unsloth and performance graphs are dropped in the replies below!

Alok

44,971 views • 22 days ago

Behold... GPT-OSS (20B) running 100% locally in your browser on WebGPU. This shouldn't be possible — but with Transformers.js v4 and ONNX Runtime Web, it is! A new class of AI apps is emerging. Zero-install, infinite distribution. Simply visit a website and run models locally.

Behold... GPT-OSS (20B) running 100% locally in your browser on WebGPU. This shouldn't be possible — but with Transformers.js v4 and ONNX Runtime Web, it is! A new class of AI apps is emerging. Zero-install, infinite distribution. Simply visit a website and run models locally.

Xenova

311,512 views • 6 months ago

Today we're launching Genome Intelligence — explore your genome with the latest AI models, without handing them your genome. With Genome Intelligence: - Explore blood results, diagnoses, medical records, and more alongside your genetics. - Get the latest genetic research as simple prompts you can ask of your own genome. - Get your genome automatically re-annotated each month to keep it up to date with the latest research. Genome Computer is a Public Benefit Corporation legally bound to never sell or license your individual genetic data. It's your data and always will be. Get started at The Genome Computer Company.

Today we're launching Genome Intelligence — explore your genome with the latest AI models, without handing them your genome. With Genome Intelligence: - Explore blood results, diagnoses, medical records, and more alongside your genetics. - Get the latest genetic research as simple prompts you can ask of your own genome. - Get your genome automatically re-annotated each month to keep it up to date with the latest research. Genome Computer is a Public Benefit Corporation legally bound to never sell or license your individual genetic data. It's your data and always will be. Get started at The Genome Computer Company.

David Ball 🌻

183,294 views • 22 days ago

I didn't expect DeepSeek v4 PRO (not Flash) to run well on the Mac Studio M3 Ultra with 512GB of RAM. This is 2 bit quantized with the same DwarfStar recipe used for Flash. 433GB GGUF file. 130 t/s prefill, 13 t/s generation. Prefill in the video is low because small prompt.

I didn't expect DeepSeek v4 PRO (not Flash) to run well on the Mac Studio M3 Ultra with 512GB of RAM. This is 2 bit quantized with the same DwarfStar recipe used for Flash. 433GB GGUF file. 130 t/s prefill, 13 t/s generation. Prefill in the video is low because small prompt.

antirez

171,478 views • 3 months ago

Your AI, Now Powered by Your Own Computer. 🚀 What if your AI could actually do the work on your laptop? Claw Intelligence makes it possible. 1️⃣ Link: Register your device with a simple terminal command. 2️⃣ Command: Tell the AI to handle file transfers or environment setups. 3️⃣ Execute: Watch as the AI prepares your local workspace for you. Check out the video to see how we sync with your computer to streamline a BTC monitor task! 📈 Tyr Now:

Your AI, Now Powered by Your Own Computer. 🚀 What if your AI could actually do the work on your laptop? Claw Intelligence makes it possible. 1️⃣ Link: Register your device with a simple terminal command. 2️⃣ Command: Tell the AI to handle file transfers or environment setups. 3️⃣ Execute: Watch as the AI prepares your local workspace for you. Check out the video to see how we sync with your computer to streamline a BTC monitor task! 📈 Tyr Now:

Claw Intelligence

12,355 views • 4 months ago

Finally got my character running in-game using UE5's GASP project! The feeling of seeing your own design move is indescribable.Who understands this feeling? 🎉

Finally got my character running in-game using UE5's GASP project! The feeling of seeing your own design move is indescribable.Who understands this feeling? 🎉

贤瑜Lucas

81,776 views • 2 months ago

NVIDIA JUST MADE A WORLD-CLASS AI CODER COMPLETELY FREE WHERE YOU SPEAK TO YOUR SCREEN AND THE APP BUILDS ITSELF it runs locally on your own computer using Claude and Ollama with no subscriptions, no cloud, and no technical setup headache.

NVIDIA JUST MADE A WORLD-CLASS AI CODER COMPLETELY FREE WHERE YOU SPEAK TO YOUR SCREEN AND THE APP BUILDS ITSELF it runs locally on your own computer using Claude and Ollama with no subscriptions, no cloud, and no technical setup headache.

0xMarioNawfal

57,505 views • 2 months ago

A $1,800 MINI PC ON YOUR DESK JUST MADE $5,040 A YEAR IN AI SUBSCRIPTIONS POINTLESS Claude Code Max, ChatGPT Pro, Cursor. That's $420 a month just to keep the workflow alive. A GMKtec mini PC with 128GB of unified memory runs the same stack locally for a one time cost. Ollama, Linux, local endpoints, Claude Code pointed at your own machine. Every prompt stays on your desk. Nothing touches the cloud.

A $1,800 MINI PC ON YOUR DESK JUST MADE $5,040 A YEAR IN AI SUBSCRIPTIONS POINTLESS Claude Code Max, ChatGPT Pro, Cursor. That's $420 a month just to keep the workflow alive. A GMKtec mini PC with 128GB of unified memory runs the same stack locally for a one time cost. Ollama, Linux, local endpoints, Claude Code pointed at your own machine. Every prompt stays on your desk. Nothing touches the cloud.

0xMarioNawfal

50,730 views • 2 months ago

Your data is your edge, but only if your AI is built on it. Rent a generic model and so can your competitor. The companies with an edge are deploying custom models that they own and improve over time. Our co-founder Rhythm Garg recently stopped by South Park Commons to share how companies are owning their intelligence with Applied Compute.

Your data is your edge, but only if your AI is built on it. Rent a generic model and so can your competitor. The companies with an edge are deploying custom models that they own and improve over time. Our co-founder Rhythm Garg recently stopped by South Park Commons to share how companies are owning their intelligence with Applied Compute.

Applied Compute

118,922 views • 2 months ago