# \#cuda **URL:** https://www.artofsm.art/tag/cuda/339.md [Latest](https://www.artofsm.art/latest.md) · [Categories](https://www.artofsm.art/categories.md) · [Tags](https://www.artofsm.art/tags.md) --- ## [NVIDIA Put 128GB In A Laptop. What's The Catch?](https://www.artofsm.art/t/nvidia-put-128gb-in-a-laptop-whats-the-catch/25627)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [10 October 2026 22:12 UTC](https://www.artofsm.art/t/nvidia-put-128gb-in-a-laptop-whats-the-catch/25627 "2026-10-10T22:12:15Z")
NVIDIA’s RTX Spark laptop platform offers up to 128 GB of unified memory with native CUDA support, promising portable AI computing but facing challenges such as limited usable memory, performance and batter… --- ## [Deepseek & Huawei Made Local AI Run Without NVIDIA's CUDA (And Why it's Hard)](https://www.artofsm.art/t/deepseek-huawei-made-local-ai-run-without-nvidias-cuda-and-why-its-hard/24987)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [4 October 2026 15:50 UTC](https://www.artofsm.art/t/deepseek-huawei-made-local-ai-run-without-nvidias-cuda-and-why-its-hard/24987 "2026-10-04T15:50:48Z")
Deepseek has developed optimized kernel libraries for Huawei’s Ascend 950 chips that achieve near-full hardware utilization, offering drop-in replacements for Nvidia’s CUDA libraries and challenging Nvidia’… --- ## [This Kernel is 40% Faster Than Nvidia's Own!](https://www.artofsm.art/t/this-kernel-is-40-faster-than-nvidias-own/23928)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [14 September 2026 08:33 UTC](https://www.artofsm.art/t/this-kernel-is-40-faster-than-nvidias-own/23928 "2026-09-14T08:33:30Z")
A team of Stanford graduate students developed Thunder Kittens, a GPU kernel that runs 40% faster than Nvidia’s highly optimized code by directly utilizing tensor cores and bypassing Nvidia’s outdated CUDA … --- ## [\[53\] "The CUDA Moat is Gone"](https://www.artofsm.art/t/53-the-cuda-moat-is-gone/23864)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [12 September 2026 10:00 UTC](https://www.artofsm.art/t/53-the-cuda-moat-is-gone/23864 "2026-09-12T10:00:20Z")
The video discusses AI End, a Japanese frontier AI company led by David Bennett, which focuses on sovereign AI by building and operating heterogeneous hardware data centers to provide secure, localized AI i… --- ## [AMD Beats Nvidia On Every Spec And Still Loses](https://www.artofsm.art/t/amd-beats-nvidia-on-every-spec-and-still-loses/23605)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [5 September 2026 22:20 UTC](https://www.artofsm.art/t/amd-beats-nvidia-on-every-spec-and-still-loses/23605 "2026-09-05T22:20:26Z")
Despite AMD’s superior GPU hardware specifications, Nvidia dominates the enterprise market due to its entrenched CUDA software ecosystem, which offers seamless integration with popular AI frameworks and ext… --- ## [Nvidia Sees AI-Fueled Demand Boosting Sales 70% Next Year](https://www.artofsm.art/t/nvidia-sees-ai-fueled-demand-boosting-sales-70-next-year/23274)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [27 August 2026 16:18 UTC](https://www.artofsm.art/t/nvidia-sees-ai-fueled-demand-boosting-sales-70-next-year/23274 "2026-08-27T16:18:33Z")
Nvidia anticipates a 70% revenue increase next year driven by strong demand from hyperscalers and rapidly growing emerging players, while working to diversify its customer base and manage margin pressures t… --- ## [New Local AI Inference Engine You Should Be Using in 2027 ? (FreeToken)](https://www.artofsm.art/t/new-local-ai-inference-engine-you-should-be-using-in-2027-freetoken/23227)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [26 August 2026 15:54 UTC](https://www.artofsm.art/t/new-local-ai-inference-engine-you-should-be-using-in-2027-freetoken/23227 "2026-08-26T15:54:38Z")
The video introduces FreeToken, a novel local AI inference engine that efficiently runs massive mixture of experts models on a single desktop by dynamically managing GPU and system RAM using a global LRU ca… --- ## [The Honest Guide To Local AI Supercomputers (NVIDIA DGX Spark)](https://www.artofsm.art/t/the-honest-guide-to-local-ai-supercomputers-nvidia-dgx-spark/23006)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [21 August 2026 11:39 UTC](https://www.artofsm.art/t/the-honest-guide-to-local-ai-supercomputers-nvidia-dgx-spark/23006 "2026-08-21T11:39:13Z")
The video provides a detailed overview of the Nvidia DGX Spark, a compact AI supercomputer with 128 GB unified memory that enables local running of large AI models, highlighting its portability, remote acce… --- ## [Your Coding Agent Should Do AI System Engineering — Ben Burtenshaw, Hugging Face](https://www.artofsm.art/t/your-coding-agent-should-do-ai-system-engineering-ben-burtenshaw-hugging-face/19038)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [21 May 2026 13:20 UTC](https://www.artofsm.art/t/your-coding-agent-should-do-ai-system-engineering-ben-burtenshaw-hugging-face/19038 "2026-05-21T13:20:36Z")
Ben Burtenshaw from Hugging Face advocates for coding agents to engage in advanced AI system engineering tasks, such as developing optimized CUDA kernels and fine-tuning large language models, supported by … --- ## [How to Use Ai Agent Hermes in Qwen 3.6 27B](https://www.artofsm.art/t/how-to-use-ai-agent-hermes-in-qwen-3-6-27b/18223)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [1 May 2026 01:54 UTC](https://www.artofsm.art/t/how-to-use-ai-agent-hermes-in-qwen-3-6-27b/18223 "2026-05-01T01:54:16Z")
The video offers a detailed guide on using the Hermes AI agent with the Qwen 3.6 27B model in a locally managed Proxmox environment, highlighting optimized hardware configurations, key settings for performa… --- ## [The CUDA Trick That Makes LLMs Faster AND Use Less Power (Real Results)](https://www.artofsm.art/t/the-cuda-trick-that-makes-llms-faster-and-use-less-power-real-results/17677)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [16 April 2026 14:21 UTC](https://www.artofsm.art/t/the-cuda-trick-that-makes-llms-faster-and-use-less-power-real-results/17677 "2026-04-16T14:21:46Z")
The video explains how the “mega kernel” technique, which fuses all GPU operations for LLM inference into a single CUDA kernel, significantly reduces CPU-GPU communication overhead, resulting in faster and … --- ## [China Destroying NVIDIA with Drop In CUDA Replacement - Huawei CANN Next](https://www.artofsm.art/t/china-destroying-nvidia-with-drop-in-cuda-replacement-huawei-cann-next/17076)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [30 March 2026 14:57 UTC](https://www.artofsm.art/t/china-destroying-nvidia-with-drop-in-cuda-replacement-huawei-cann-next/17076 "2026-03-30T14:57:29Z")
Eli the Computer Guy discusses China’s strategic development of Huawei’s CANN Next, a drop-in CUDA replacement designed to seamlessly integrate with existing AI codebases and challenge NVIDIA’s dominance in… --- ## [LTX 2.3, GPT 5.4, CUDA agent, realtime AI videos, new image models 360 videos: AI NEWS](https://www.artofsm.art/t/ltx-2-3-gpt-5-4-cuda-agent-realtime-ai-videos-new-image-models-360-videos-ai-news/16539)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [8 March 2026 03:51 UTC](https://www.artofsm.art/t/ltx-2-3-gpt-5-4-cuda-agent-realtime-ai-videos-new-image-models-360-videos-ai-news/16539 "2026-03-08T03:51:10Z")
This week’s AI news highlights major advancements in open-source video and image editing tools, real-time video generation, and robotics, including new models like LTX 2.3, GPT-5.4, and ByteDance’s CUDA Age… --- ## [The end of the GPU era](https://www.artofsm.art/t/the-end-of-the-gpu-era/15400)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [21 January 2026 00:03 UTC](https://www.artofsm.art/t/the-end-of-the-gpu-era/15400 "2026-01-21T00:03:42Z")
The video explains how Nvidia became the world’s most valuable company by dominating the GPU market for AI, but notes that major tech firms are now developing or adopting specialized chips that could end Nv… --- ## [Nvidia "Buys" Groq for $20 BILLION](https://www.artofsm.art/t/nvidia-buys-groq-for-20-billion/15014)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [28 December 2025 12:58 UTC](https://www.artofsm.art/t/nvidia-buys-groq-for-20-billion/15014 "2025-12-28T12:58:33Z")
Nvidia’s $20 billion deal with Groq is a non-exclusive licensing agreement that effectively transfers Groq’s top talent and technology to Nvidia, leaving Groq as a diminished “zombie startup” while sidestep… --- ## [Reinforcement Learning Tutorial - RLVR with NVIDIA & Unsloth](https://www.artofsm.art/t/reinforcement-learning-tutorial-rlvr-with-nvidia-unsloth/14753)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [15 December 2025 13:20 UTC](https://www.artofsm.art/t/reinforcement-learning-tutorial-rlvr-with-nvidia-unsloth/14753 "2025-12-15T13:20:44Z")
This tutorial, created with NVIDIA and Unsloth, demonstrates how to set up and run reinforcement learning with verifiable rewards (RLVR) locally on a home PC using an NVIDIA RTX GPU by teaching an AI model … --- ## [AI Trade is Different From 90’s Says Nancy Tengler](https://www.artofsm.art/t/ai-trade-is-different-from-90-s-says-nancy-tengler/14319)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [26 November 2025 20:38 UTC](https://www.artofsm.art/t/ai-trade-is-different-from-90-s-says-nancy-tengler/14319 "2025-11-26T20:38:38Z")
Nancy Tengler highlights that the current AI-driven technological revolution differs from the 1990s tech boom, emphasizing strong earnings growth, increasing capital investments, and broad AI adoption acros… --- ## [Nvidia-Google AI Chip Battle Escalates](https://www.artofsm.art/t/nvidia-google-ai-chip-battle-escalates/14273)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [25 November 2025 15:22 UTC](https://www.artofsm.art/t/nvidia-google-ai-chip-battle-escalates/14273 "2025-11-25T15:22:51Z")
The video explores the intensifying competition between Nvidia and Google in the AI chip market, highlighting Google’s TPU technology as a more energy-efficient alternative challenging Nvidia’s GPU dominanc… --- ## [Microsoft Code Allows CUDA to Work On AMD GPU's - MS Destroying NVIDIA](https://www.artofsm.art/t/microsoft-code-allows-cuda-to-work-on-amd-gpus-ms-destroying-nvidia/13895)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [11 November 2025 13:25 UTC](https://www.artofsm.art/t/microsoft-code-allows-cuda-to-work-on-amd-gpus-ms-destroying-nvidia/13895 "2025-11-11T13:25:59Z")
Eli the Computer Guy discusses Microsoft’s development of toolkits that translate Nvidia’s CUDA code to run on AMD GPUs, aiming to reduce Nvidia’s dominance in AI hardware by enabling cross-platform compati… --- ## [SGLang: Open-Source Model Performance Optimization](https://www.artofsm.art/t/sglang-open-source-model-performance-optimization/13871)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [10 November 2025 18:23 UTC](https://www.artofsm.art/t/sglang-open-source-model-performance-optimization/13871 "2025-11-10T18:23:26Z")
John, the Eston core maintainer, presents SGLang, an open-source framework that optimizes large language model performance through innovations like deepseek array optimizations, reinforcement learning integ… --- ## [Anthropic’s TPU move and NVIDIA’s Starcloud](https://www.artofsm.art/t/anthropic-s-tpu-move-and-nvidia-s-starcloud/13667)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [31 October 2025 10:24 UTC](https://www.artofsm.art/t/anthropic-s-tpu-move-and-nvidia-s-starcloud/13667 "2025-10-31T10:24:55Z")
In this episode of Mixture of Experts, the panel discusses Anthropic’s shift to using Google’s TPUs to diversify hardware and reduce costs, debates the practicality of AI risk insurance, and examines OpenAI… --- ## [NVIDIA RTX Pro… in a LAPTOP!](https://www.artofsm.art/t/nvidia-rtx-pro-in-a-laptop/13616)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [29 October 2025 12:24 UTC](https://www.artofsm.art/t/nvidia-rtx-pro-in-a-laptop/13616 "2025-10-29T12:24:31Z")
The Dell Pro Max 14 laptop, featuring the NVIDIA RTX Pro 2000 GPU and Intel Core Ultra 9285H CPU, offers strong AI and professional performance in a premium, portable design with a vibrant 14-inch OLED touc… --- ## [Jensen Huang - The Man Behind The AI Revolution](https://www.artofsm.art/t/jensen-huang-the-man-behind-the-ai-revolution/13591)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [28 October 2025 15:51 UTC](https://www.artofsm.art/t/jensen-huang-the-man-behind-the-ai-revolution/13591 "2025-10-28T15:51:17Z")
Jensen Huang, co-founder and CEO of Nvidia, transformed the company from near bankruptcy into a multi-trillion-dollar leader by pioneering the development of GPUs and the CUDA platform, which became essenti… --- ## [LIVE: Nvidia CEO Jensen Huang Keynote Address](https://www.artofsm.art/t/live-nvidia-ceo-jensen-huang-keynote-address/13593)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [28 October 2025 12:30 UTC](https://www.artofsm.art/t/live-nvidia-ceo-jensen-huang-keynote-address/13593 "2025-10-28T12:30:20Z")
In his GTC Washington DC keynote, Nvidia CEO Jensen Huang highlighted the company’s pioneering role in accelerated computing and AI, unveiling strategic partnerships like the Nvidia-Nokia collaboration for … --- ## [NVIDIA DGX Spark…What Went Wrong](https://www.artofsm.art/t/nvidia-dgx-spark-what-went-wrong/12846)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [26 September 2025 18:03 UTC](https://www.artofsm.art/t/nvidia-dgx-spark-what-went-wrong/12846 "2025-09-26T18:03:53Z")
The Nvidia DGX Spark, announced in early 2025 as the world’s smallest AI supercomputer with advanced hardware and clustering capabilities, has faced multiple delays and remains unreleased due to likely prod… --- ## [New Hardware! NVidia Thor, Mini PCs, 5K2K Monitors, Keyboards and More!](https://www.artofsm.art/t/new-hardware-nvidia-thor-mini-pcs-5k2k-monitors-keyboards-and-more/12825)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [25 September 2025 21:58 UTC](https://www.artofsm.art/t/new-hardware-nvidia-thor-mini-pcs-5k2k-monitors-keyboards-and-more/12825 "2025-09-25T21:58:26Z")
In this video, Dave reviews four innovative tech products—a powerful AMD Ryzen 9-based GeekCom A9 Max mini PC, a budget-friendly Anoken 40-inch 5K2K ultrawide monitor, two advanced Keyron Hall effect keyboa… --- ## [Local Ai Homelab Setup Guide Proxmox 9 Llama.cpp Unsloth in LXC](https://www.artofsm.art/t/local-ai-homelab-setup-guide-proxmox-9-llama-cpp-unsloth-in-lxc/11956)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [17 August 2025 10:19 UTC](https://www.artofsm.art/t/local-ai-homelab-setup-guide-proxmox-9-llama-cpp-unsloth-in-lxc/11956 "2025-08-17T10:19:33Z")
The video provides a detailed walkthrough for setting up llama.cpp with GPU acceleration inside a Debian 12 LXC container on a Proxmox 9 host, including CUDA installation, GPU passthrough, building the soft… --- ## [NVIDIA Orders 300,000 More H20 GPU's for China -- Xi Beats Trump](https://www.artofsm.art/t/nvidia-orders-300-000-more-h20-gpus-for-china-xi-beats-trump/11455)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [30 July 2025 21:20 UTC](https://www.artofsm.art/t/nvidia-orders-300-000-more-h20-gpus-for-china-xi-beats-trump/11455 "2025-07-30T21:20:53Z")
The video discusses Nvidia’s sale of 300,000 slightly downgraded H20 GPUs to China amid inconsistent and unclear U.S. policies aimed at restricting advanced AI hardware exports, highlighting the complexitie… --- ## [NVIDIA CUDA on RISC-V -- Kinda Cool, But Probably Worthless](https://www.artofsm.art/t/nvidia-cuda-on-risc-v-kinda-cool-but-probably-worthless/11266)
**Author:** [@merefield](https://www.artofsm.art/u/merefield)\ **Replies:** 1\ **Last updated:** [22 July 2025 17:20 UTC](https://www.artofsm.art/t/nvidia-cuda-on-risc-v-kinda-cool-but-probably-worthless/11266 "2025-07-22T17:20:49Z")
Eli the Computer Guy discusses Nvidia’s recent support for RISC-V in its CUDA platform, viewing it as a strategic move rather than an indication of imminent mainstream adoption due to RISC-V’s current ecosy… --- ## [Nvidia allowed to sell to China: Here's the company's market opportunity](https://www.artofsm.art/t/nvidia-allowed-to-sell-to-china-heres-the-companys-market-opportunity/11084)
**Author:** [@artesia](https://www.artofsm.art/u/artesia)\ **Replies:** 1\ **Last updated:** [18 July 2025 15:37 UTC](https://www.artofsm.art/t/nvidia-allowed-to-sell-to-china-heres-the-companys-market-opportunity/11084 "2025-07-18T15:37:02Z")
Nvidia’s recent approval to sell its H-20 chip to China, combined with its strong Cuda software ecosystem and the rapidly growing AI inference market, positions the company for significant growth despite th… [Next page](https://www.artofsm.art/tag/cuda/339.md?match_all_tags=true&page=1&tags%5B%5D=cuda)