到底 PewDiePie 要停止使用怎样的 AI？又要开始使用怎样的 AI？ | 机器之心发布 | | --- | | 机器之心编辑部 | 11 月 3 日，全球知名游戏博主 PewDiePie 发布视频，展示其自建本地 AI 系统的全过程。该视频目前浏览量已经超过 300 万，视频标题则赫然写着双关梗 "STOP: Using AI Right now"。 他投入 2 万美元，组装了包含 10 块英伟达 GPU（含 8 块改装版 RTX 4090 和 2 块 RTX 4000 Ada）的本地系统，支持 700 亿至 2450 亿参数的大模型运行，全程无 需依赖云计算。 通过 vLLM 框架，他试验了 Qwen-235B 等模型，构建了具备搜索、记忆、RAG（检索增强生成）和音频输出功能的定制化 AI 界面。 要知道 PewDiePie 是海外游戏博主顶流中的顶流，Youtube 粉丝超 1.1 亿，Ins 粉丝 2 千万。他下场部署本地 AI 引起全网巨大轰动。 一夜之间，本地部署 AI 大模型成为讨论热度极高的话题。 这要先从本地 AI 的特性说起。 在传统 AI 订阅模式中，大模型 + 云服务 = 账号 ...

标普500指数下跌0.92%，4月以来首次跌破50日均线这一关键技术支撑位。 市场恐慌指数VIX随之飙升12.97%，报22.39，显示投资者恐慌情绪大幅升温。 华尔街的恐慌情绪像病毒一样蔓延。 道指暴跌557点，标普500指数跌破关键支撑线，而"硅谷风投教父"彼得·蒂尔清仓英伟达的全部股份，正在引发一场关 于AI泡沫的深刻忧虑。 交易终端屏幕上的红色数字不断跳动。 华尔街刚刚经历了一个不眠之夜，道指暴跌557.24点，跌幅达1.18%，创下近一个月来最差表现。 彼得·蒂尔的清仓行为引发市场广泛关注。 他不仅清空了英伟达全部持股，还减持特斯拉20.76万股，减持比例高达76%。 蒂尔宏观基金三季度末持仓总市值仅为7440万美元，相比二季度的2.12亿美元大幅下降65%。 该基金同时新建仓苹果和微软，但总体呈现大幅收缩态势。 彼得·蒂尔今年早些时候曾警告英伟达估值过高，并将科技股估值飙升与1999-2000年互联网泡沫破灭进行了类比。 作为PayPal联合创始人和Facebook早期投资者，他在硅谷拥有巨大影响力，其投资动向备受关注。 在这场全面溃败中，"硅谷风投教父"彼得·蒂尔旗下基金清仓英伟达全部股份的消息 ...

Core Viewpoint - A modification involving the addition of a single resistor has allowed an RTX 4090 gaming laptop to outperform an RTX 5090 in certain benchmarks, highlighting the significant impact of power consumption on performance [1][10]. Group 1: Power Consumption and Performance - Power consumption is a critical factor that directly influences the performance of gaming laptops, with high-end models often boasting total power consumption exceeding 200W [3][5]. - The total power consumption typically refers to the combined power of the CPU and GPU, where higher power levels correlate with better performance due to enhanced cooling and power supply requirements [5][12]. - A user modified their ROG Zephyrus M16 by adding a resistor, effectively lowering the circuit resistance and allowing the RTX 4090 to draw nearly double its original power limit, resulting in performance that rivals the RTX 5090 [9][10]. Group 2: Benchmark Comparisons - After the modification, the performance of the RTX 4090 in the ROG M16 surpassed that of the RTX 5090 in most 3DMark tests, with the highest score in the Speedway benchmark showing a 9.6% lead [10][11]. - The overall performance improvement from the modification was over 20% in most benchmarks, with some tests showing increases of more than 35% [11][12]. Group 3: Manufacturer Limitations - NVIDIA is identified as the entity that sets power consumption limits for mobile GPUs, which restricts manufacturers from fully utilizing the hardware's potential [13][15]. - Despite the potential for higher performance through increased power limits, manufacturers often adhere to NVIDIA's restrictions to maintain product differentiation and avoid market conflicts [15][16]. - There are indications that NVIDIA may consider lifting power limits for future high-end models to cater to hardcore gaming enthusiasts seeking significant performance boosts [15][16].

Core Insights - 3D Gaussian Splatting (3DGS) is an emerging method for novel view synthesis that utilizes a set of images with poses to iteratively train a scene representation composed of numerous anisotropic 3D Gaussian bodies, capturing the appearance and geometry of the scene [2][4] - The CLM system proposed by the team allows 3DGS to render large scenes using a single consumer-grade GPU, such as the RTX 4090, by addressing GPU memory limitations [6][8] Group 1: 3DGS Overview - 3DGS has shown revolutionary application potential in fields such as 3D modeling, digital twins, visual effects (VFX), VR/AR, and robot vision reconstruction (SLAM) [5] - The quality of images rendered using 3DGS depends on the fidelity of the trained scene representation, with larger and more complex scenes requiring more Gaussian bodies, leading to increased memory usage [5] Group 2: CLM System Design - CLM is designed based on the insight that the computation of 3DGS is inherently sparse, allowing only a small subset of Gaussian bodies to be accessed during each training iteration [8][20] - The system employs a novel unloading strategy that minimizes performance overhead and scales to large scenes by dynamically loading only the necessary Gaussian bodies into GPU memory while offloading the rest to CPU memory [8][11] Group 3: Performance and Efficiency - The implementation of CLM can render a large scene requiring 102 million Gaussian bodies on a single RTX 4090 while achieving top-tier reconstruction quality [8] - Each view typically accesses only 0.39% of the Gaussian points, with a maximum of 1.06% for any single view, highlighting the sparse nature of the data [23] Group 4: Optimization Techniques - The team utilized several unique characteristics of 3DGS to significantly reduce communication overhead associated with unloading, including pre-computing the accessed Gaussian sets for each view and leveraging spatial locality to optimize data transfer between CPU and GPU [12][17] - The microbatch scheduling optimization allows for overlapping access patterns between consecutive batches, enhancing cache hit rates and reducing redundant data transfers [24][25] Group 5: Results and Impact - CLM enhances the training capacity of 3DGS models by up to 6.1 times compared to pure GPU training baselines, enabling the training of larger models that improve scene reconstruction accuracy while lowering communication and unloading overhead [27]

RT Justin Drake (@drakefjustin)Progress toward real-time proving for Ethereum L1 is nothing short of extraordinary.In May, SP1 Hypercube proved 94% of L1 blocks in under 12 seconds using 160 RTX 4090s. Five months later Pico Prism proves 99.9% of the same blocks in under 12 seconds, with just 64 RTX 5090s. Average proving latency is now 6.9 seconds.Performance has outpaced Moore's law ever since Zcash pioneered practical SNARKs a decade ago. Today's Pico Prism results are a striking reminder of that exponen ...

Core Thesis - Advanced Micro Devices, Inc. (AMD) is positioned as a strong investment opportunity due to its potential market share gains and the challenges faced by competitor Nvidia, with a target price range of $168–$187 over the next 12–18 months [2][5]. Financial Performance - AMD reported a 32% year-over-year revenue growth in Q2 2025, reaching $7.7 billion, driven by a 73% increase in gaming revenue to $1.1 billion and a 14% rise in data center revenue to $3.2 billion [3]. - Wall Street forecasts suggest a 15–20% compound annual growth rate (CAGR) for earnings per share (EPS) through 2027, despite near-term margin pressures from export controls [3]. Competitive Landscape - Nvidia's structural GPU reliability issues, such as problems with RTX 4090 connectors, create a competitive opportunity for AMD, which is seen as a stable alternative [4]. - AMD's RX 9070 XT shows strong performance and improved power efficiency, while its open-source ROCm platform enhances its data center positioning [4]. Market Opportunities - AMD could capture $3.6–$6 billion in incremental revenue from potential market share gains in the $120 billion discrete GPU segment, although Nvidia's ecosystem dominance poses challenges [5]. - The company's diversified revenue streams and competitive GPU offerings support the potential for multiple expansions, despite macroeconomic risks such as Federal Reserve rate hikes [5]. Historical Context - AMD's stock price has appreciated approximately 39% since May 2025, reflecting strong revenue growth driven by data center and Ryzen processor sales, as well as AI demand [6].

Multi-GPU Environment - Confirmed stable execution and compatibility of MPI runtime on a dual-GPU setup (RTX 4090 ×2, CUDA 12.8) [1] Hardware & Software - Validated MPI runtime on a dual-GPU setup with RTX 4090 ×2 and CUDA 12.8 [1]

Core Viewpoint - BluSky AI Inc. has signed a Letter of Intent (LOI) with Lilac to form a strategic partnership aimed at enhancing cloud compute provisioning and monetizing idle capacity in the AI ecosystem [1][2][3] Group 1: Partnership Details - The LOI allows BluSky AI to offer its GPU cloud computing resources, including unallocated inventory and customer capacity, for rent through Lilac's marketplace [2] - This collaboration is expected to increase the utilization of BluSky AI's compute assets while expanding Lilac's supplier base with high-performance GPU models such as NVIDIA B200, H200, H100, A100, L40, RTX 5090, and RTX 4090 [2][3] - The partnership includes a multi-pronged engagement strategy involving engineering integration, co-marketing efforts, and a customer acquisition framework, with a definitive agreement anticipated in the coming months [3] Group 2: Company Missions and Goals - BluSky AI aims to democratize access to AI compute and optimize resource efficiency by integrating idle capacity into Lilac's platform, enabling customers to generate new revenue streams [3][6] - Lilac's mission is to democratize access to critical AI infrastructure and create a more efficient cloud economy by connecting idle GPU capacity to AI developers and enterprises [5][7] Group 3: Marketing and Transparency Initiatives - BluSky AI will designate Lilac as a 'Preferred Marketplace Partner' and promote the platform within its ecosystem [6] - Both companies will collaborate on various marketing initiatives, including joint press releases, social media activations, and event partnerships [6] - BluSky AI will provide quarterly transparency reports on available GPU inventory to inform marketplace strategy and performance tracking [6]

Core Viewpoint - Bolt Graphics, a lesser-known chip startup, has claimed that its first GPU module, Zeus 4C, achieves performance levels 13 times greater than the RTX 5090 in path tracing scenarios [1][4]. Performance Comparison - In path tracing tasks at 4K resolution with 120 frames per second, Zeus 4C outperforms RTX 5090 significantly, but it is not designed for gaming applications [4][6]. - Zeus GPU models (2c26-064 and 2c26-128) have advantages in board power and cache compared to RTX 4090 and RTX 5090, but they lag in floating-point performance (FP64/FP32/FP16 vector tflops) [5][6]. - Zeus GPU utilizes LPDDR5X memory, which is designed for mobile devices, resulting in lower bandwidth compared to the GDDR7 memory used in RTX 5090, leading to potential performance issues in gaming scenarios [6][9]. Target Applications - Bolt Graphics focuses on high-precision graphics rendering rather than gaming or AI performance, targeting industries such as film visual effects, game rendering, and high-performance computing (HPC) [6][9]. - Path tracing is highlighted as a key technology for Bolt Graphics, providing realistic rendering effects widely used in various industries [7][9]. Architectural Design - The Zeus GPU series features a chiplet architecture, with models incorporating multiple compute and I/O cores to enhance performance [13]. - The design aims to address memory bandwidth limitations by offering multiple SODIMM slots for additional memory [13]. Market Position and Future Outlook - Bolt Graphics is seen as a niche player in the GPU market, potentially challenging established companies like NVIDIA and AMD in specific applications [20]. - The company has not yet disclosed benchmark testing details or how its performance compares to competitors, with developer kits expected in 2026 and full production in 2027 [20][21].

Core Viewpoint - Bolt Graphics, a lesser-known chip startup, has claimed that its first GPU module, Zeus 4C, outperforms NVIDIA's RTX 5090 by 13 times in path tracing scenarios [1][8]. Group 1: Performance Claims - Zeus 4C achieves 120 frames per second at 4K resolution in path tracing tasks, significantly surpassing the RTX 5090 [8]. - Despite its impressive path tracing performance, Zeus GPUs fall short in floating-point operations compared to RTX 4090 and RTX 5090 [9]. - The Zeus GPU series is designed for high-precision graphics rendering rather than gaming or AI performance, focusing on applications like movie visual effects and high-performance computing (HPC) [12][14]. Group 2: Technical Specifications - Zeus GPUs utilize LPDDR5X memory, which is optimized for low power consumption but has lower bandwidth compared to the GDDR7 memory used in RTX 5090, leading to potential performance issues in gaming scenarios [10]. - The architecture of Zeus GPUs includes multiple compute and I/O cores, similar to AMD's chiplet design, enhancing their computational capabilities [17][19]. Group 3: Market Position and Future Prospects - Bolt Graphics is positioned as a niche player targeting specific high-demand applications rather than competing directly with established players like NVIDIA and AMD [26][27]. - The company has not yet disclosed benchmark testing details or how its performance compares to competitors, and the developer kits for Zeus GPUs are expected to be released in 2026, with mass production in 2027 [29][30].