OpenAI is embracing open source for the first time in five years, with two new models you can download right now making their debut.Get the full rundown here: https://t.co/o1GGstdU8H https://t.co/Ra40vqgAK4 ...

RT Muskonomy (@muskonomy)🚨BREAKING: Grok defeats Gemini 2.5 Pro in Kaggle’s AI Chess Exhibition ♟️The score’s now 3-2. Grok advances to the finals.Tomorrow: Grok faces off against OpenAI’s o3 in a battle for the chess crown. https://t.co/B8UQdYApCa ...

最近， 99% 的钱最后流向了 1% 的顶尖 AI 人才。 硅谷的 AI 人才战争已升级为 "精准斩首" 行动。 Meta刚从苹果挖走第三个AI核心人才，连Siri都快不会说"你好"了。 谷歌、亚马逊、微软、Meta忙着跟AI初创公司签协议，把顶尖人才收入囊中。 一年一亿美元，下班之前回答我，来不来？ …… 一个神秘的电话、一封突如其来的邮件，巨头开始天价抢人，硅谷的老板们人心惶惶，AI大牛们蠢蠢 欲动，这里面，华人工程师更成了重点目标。本期《科技很有料》AI料王半月谈，@DeepTech深科 技，@数读社，来扒一扒撑起了整个AI江湖的这些人都是谁？ @DeepTech深科技：Meta给 AI 人才提供的年薪底线是200万美元，曾给澳大利亚科学家Tulloch开出了 一份价值高达 15 亿美元的 6 年期薪酬方案，即使在以高薪著称的硅谷，这样的薪酬也极为罕见。 @数读社：扎克伯格给顶级AI人才余家辉开出了四年、总额高达3亿美元的薪酬方案，第一年的薪酬就 超过1亿美元。苹果基础模型团队负责人庞若鸣被挖角到Meta，他的薪酬更加离谱，水平达到了2亿美 元。 @ DeepTech深科技：Tulloch 2010年在 ...

Core Viewpoint - The collaboration between Amazon Web Services (AWS) and OpenAI marks a significant shift in the AI cloud service landscape, breaking Microsoft's monopoly on reselling OpenAI's software and services, and enhancing AWS's competitive edge in the large model cloud service market [3][15]. Summary by Sections Collaboration Announcement - AWS announced support for OpenAI's newly open-sourced models, gpt-oss (120b and 20b), and Anthropic's Claude Opus 4.1, through its platforms Amazon Bedrock and Amazon SageMaker AI [1][4][16]. Strategic Importance - This partnership allows AWS to fill a critical gap in its model library, enhancing its "Choice Matters" strategy, which emphasizes the importance of diverse model options for various industry needs [7][10][15]. Model Ecosystem Development - AWS's platforms now host over 400 mainstream commercial and open-source large models, facilitating a diverse AI ecosystem that accelerates technology adoption and innovation in the AI industry [10][18]. Performance and Cost Efficiency - The performance of gpt-oss-120b is reported to be three times more cost-effective than Google's Gemini, five times that of DeepSeek-R1, and twice that of OpenAI's o4, providing budget-friendly access to top-tier AI capabilities for small and medium enterprises [14][15]. Enhanced Model Deployment - AWS's Amazon SageMaker JumpStart allows for rapid deployment of advanced foundational models, including OpenAI's offerings, enabling efficient customization and optimization for AI applications [14][24]. Future Prospects - The collaboration is expected to create a win-win situation, expanding OpenAI's market reach while solidifying AWS's position as a leading platform for deploying and running various AI models [15][19]. AI Ecosystem Transformation - AWS is evolving from a cloud service provider to an AI capability aggregation platform, enhancing its role in the AI ecosystem and providing better service to customers and developers [19][29]. Model Selection Flexibility - The "Choice Matters" strategy addresses the diverse needs of different tasks, allowing developers to select models based on specific requirements, thus maximizing efficiency and effectiveness in AI applications [21][24]. Conclusion - The integration of multiple models into a single platform is anticipated to lead to a significant surge in AI application development, enabling innovative solutions through the combination of various models [30][31].

Core Viewpoint - Super X AI Technology Limited has launched its All-in-One Multi-Model Servers (MMS), which are designed to provide enterprises with a comprehensive AI infrastructure that supports the collaboration of multiple models, enhancing efficiency and reducing deployment complexity [2][4]. Product Features - The All-in-One MMS comes pre-configured with OpenAI's high-performance large language models, GPT-OSS-120B and GPT-OSS-20B, which offer superior performance in various benchmarks compared to leading closed-source models [3][4]. - This solution is characterized by its out-of-the-box readiness, allowing enterprises to deploy advanced generative AI applications immediately without extensive integration efforts [4][5]. - The MMS supports a diverse range of models, including inference, general-purpose, speech synthesis, and text-to-image models, facilitating deep integration with end application scenarios [5][6]. Market Positioning - SuperX aims to provide a competitive edge to enterprise clients in the generative AI space by offering a secure, reliable, and fully optimized platform that transforms traditional hardware into a complete AI solution [4][8]. - The company emphasizes the importance of multi-model collaboration as a step towards achieving Artificial General Intelligence (AGI), positioning itself as a leader in the evolving AI landscape [8]. Target Audience and Pricing - The All-in-One MMS is tailored for various enterprise scales, with specific models priced for different user needs, such as the AI Workstation Ultra at USD 250,000 and the Cluster Edition starting from USD 4,000,000 [9][11]. - SuperX serves a global clientele, including enterprises, research institutions, and cloud computing deployments, highlighting its broad market reach [9].

Core Viewpoint - The article discusses the complexities and dynamics within OpenAI, particularly during a crisis involving the board and the return of Sam Altman, highlighting the importance of leadership and decision-making in the tech industry [2][3][4]. Group 1: OpenAI Crisis and Leadership - Bret Taylor, a key figure in OpenAI's board, was initially reluctant to get involved but felt compelled to help after reflecting on the significance of OpenAI's impact on the AI landscape [2][3]. - Taylor emphasized the need for a transparent and fair process to address the crisis, aiming to restore trust among employees and stakeholders [3][4]. - The crisis led to a collective employee response, with a public letter demanding Sam Altman's return, indicating the strong connection between leadership and employee morale [3][4]. Group 2: AI Market Dynamics - The AI market is expected to evolve into three main segments: foundational models, AI tools, and application-based AI, with a particular focus on the potential of AI agents [5][33]. - Foundational models will likely be dominated by a few large companies due to the high capital requirements for training these models, making it a challenging area for startups [34][35]. - The AI tools market presents risks as larger infrastructure providers may introduce competing products, necessitating careful strategic planning for smaller companies [36][37]. Group 3: Application-Based AI and Business Models - The application-based AI market is seen as the most promising, with companies developing AI agents to handle specific business tasks, leading to higher profit margins [37][38]. - The shift towards AI agents represents a significant change in how software is perceived, moving from tools that assist humans to systems that can autonomously complete tasks [41][42]. - The concept of "outcome-based pricing" is gaining traction, where companies charge based on the results delivered by AI agents, aligning business goals with customer satisfaction [44][46].

Core Viewpoint - Grok Imagine has officially launched for all Grok Heavy subscribers, gaining significant attention on social media as an "AI version of Vine" [1][5] Group 1: User Engagement and Popularity - Users have creatively resurrected historical figures like Steve Jobs and Nikola Tesla using Grok Imagine [2] - The platform has inspired humorous content, with users creating funny renditions of Elon Musk as various superheroes [3] - Some videos have garnered over 13 million views, showcasing the platform's viral potential [4] Group 2: Unique Features of Grok Imagine - Grok Imagine differentiates itself from competitors like OpenAI's Sora and Google's Veo 3 by offering a simpler user experience, allowing anyone to create videos from text without complex parameters [10][12] - The platform supports a "Spicy" mode for more creative freedom, appealing to users looking for novelty [13] - Video rendering time has significantly improved, reducing from 60 seconds to 15 seconds, with expectations to further decrease to under 12 seconds soon, all while maintaining quality [15][16] Group 3: Community and Industry Impact - Grok Imagine is deeply integrated with the X social platform, creating a closed loop of generation, sharing, and virality [18] - Industry leaders, including the CEOs of Dell and Shopify, have praised Grok Imagine for its innovative capabilities, indicating a shift in content creation dynamics [19][20] - The domestic AI creation ecosystem is also thriving, with tools like Keling and iFlytek driving creators towards industrial-level production [21] - AI technology is now capable of managing key production elements in short dramas, suggesting a significant industry turning point [23][25] - This technological revolution is poised to reshape the content industry, potentially leading to a surge in AI-generated short dramas in the near future [26]

机器之心报道 机器之心编辑部 众所周知，大型语言模型的训练通常分为两个阶段。 第一 阶段 是「预训练」 ，开发者利用大规模文本数据集训练模型，让它学会预测句子中的下一个词。 第二 阶段是「后训练」 ，旨在教会模型如何更好地理解和执行人类指令。 在 LLM 后训练阶段，似乎是一个强化学习的特殊形式。用于大语言模型（LLMs）微调的强化学习（RL）算法正沿着一条明确的演进路径持续发展。 起初，OpenAI 开创了一种名为 基于 人类反馈的强化学习（RLHF） 的技术，用于改进 ChatGPT。RLHF 的核心是让人类标注员对模型生成的多种响应进行打分， 并选出最优答案作为训练参考。这一过程虽然有效，但也耗时、昂贵且依赖人力，通常需要一支小型但专业的数据标注团队。 DeepSeek 的重要创新在于用 RL 技术自动化了这一环节。算法不再依赖人工逐一评估，而是让模型在探索过程中，通过获得「奖励信号」自主学习正确行为，从 而显著降低了成本，提高了效率，最终能以较低的成本实现高性能。 OpenAI 在 ChatGPT 的训练中采用了 近端策略优化（Proximal Policy Optimization, PPO） 。 ...

SoftBank Group says it’s taking longer than anticipated to get Stargate off the ground, conceding for the first time that the $500 billion AI tie-up in the US with OpenAI is bogging down https://t.co/0m0TTJ2eG0 ...

Core Viewpoint - The article anticipates a significant live event from OpenAI, likely focusing on advancements in reinforcement learning and its implications for inference computing power [1] Group 1 - The event is expected to highlight breakthroughs in reinforcement learning, which could enhance inference applications [1] - There is an emphasis on the readiness of various ASICs and inference chips to support these advancements [1]