Yao Shunyu
PulseAugur coverage of Yao Shunyu — every cluster mentioning Yao Shunyu across labs, papers, and developer communities, ranked by signal.
2 天有情绪数据
Tencent to explore AI-driven energy management solutions
Given the emerging bottleneck of electricity supply for AI and the industry's pivot towards 'AI power systems,' Tencent, with its extensive infrastructure and AI focus, may explore developing AI solutions for power grid integration and energy management. This could involve optimizing energy consumption for its data centers or offering such solutions to external clients.
Yao Shunyu to announce Tencent's AGI roadmap within 12 months
Following his appointment as Tencent's chief AI scientist and his stated view of AI as a long-term game with diverse future paths, Yao Shunyu is likely to articulate a concrete AGI development roadmap for Tencent. This announcement could detail specific research directions, milestones, and timelines, providing clarity on Tencent's strategic approach to achieving artificial general intelligence.
Tencent is prioritizing AI integration into its core product development and infrastructure
Evidence suggests Tencent is moving beyond AI research to deep integration. The company is using AI for code generation, advancing chip design, and rolling out 'efficiency intelligence agent toolset'. This indicates a strategic focus on embedding AI across its ecosystem to solve real-world problems and enhance user value, rather than solely pursuing theoretical advancements.
Yao Shunyu to lead Tencent's AGI research division
Given Yao Shunyu's recent appointment as Tencent's chief AI scientist and his background at OpenAI, it's plausible he will be tasked with leading a dedicated AGI research division. This would align with Tencent's stated long-term AGI development goals and Yao's expertise.
Tencent to announce new multimodal or embodied AI initiatives within 6 months
Yao Shunyu's comments about future AI growth in multimodal and embodied intelligence, coupled with Tencent's broad ecosystem, suggest they are actively investing in these areas. An announcement or significant progress update on such initiatives could be expected within the next six months.
-
腾讯聘请前OpenAI研究员田涌泷开发VLM
腾讯已聘请前OpenAI研究员田涌泷加入其大语言模型部门。田涌泷将专注于开发视觉语言模型(VLM)和多模态AI。此前,腾讯已从OpenAI招募了包括现任腾讯首席AI科学家姚舜宇在内的AI人才。
-
腾讯混元3发布,拥有2950亿参数,搜索能力媲美GPT-5.5
腾讯正式发布了混元3(Hy3),这是一款采用混合专家(MoE)架构、拥有2950亿总参数的AI模型。该模型拥有256K的上下文窗口,并声称在搜索能力上媲美GPT-5.5,同时与前代版本相比幻觉率降低了一半。腾讯将这些改进归因于增强的训练后数据质量和扩大的强化学习计算规模,并指出姚舜宇教授的加入是一个重要因素。
-
AI的下一个前沿:电力网整合而非计算能力
人工智能行业正面临一个关键瓶颈,并非模型性能或计算能力,而是电力供应和管理。随着AI模型变得越来越强大,数据中心的能耗呈指数级增长,传统的电网基础设施难以跟上步伐。专家建议,AI的未来取决于开发一个集成了能源和智能系统的“AI电力系统”,以实现实时协调,超越简单的供电,根据AI工作负载和可再生能源的可用性主动管理能源消耗。
-
中国高管从OpenAI转投腾讯,标志着中国转向AGI竞赛
姚舜宇从OpenAI转投腾讯,标志着中国在通用人工智能(AGI)方面的战略转变。这一转变表明中国正从防御姿态转向公开竞争AGI的开发。此举暗示中国正积极寻求AGI领域的领导地位。
-
腾讯押注人工智能的长期发展,聚焦解决实际问题
腾讯将人工智能视为一项长期事业,专注于识别和解决有意义的问题,而非仅仅追求模型进步。该公司强调“共创”方法,将AI模型与产品开发深度融合,以确保其实际价值和用户满意度。腾讯正在推出“效率智能代理工具集”,以促进AI在个人和企业用例中的应用,并利用其广泛的生态系统和基础设施来推动AI与实际应用的整合。
-
东芝开发销售AI;腾讯聘请前OpenAI研究员加速AGI
东芝正在开发用于销售预测的AI,目标是在2027财年实现全面实际应用。与此同时,腾讯已任命前OpenAI研究员姚舜宇为其首席人工智能科学家,以加速其AGI开发工作。
-
腾讯AI主管:AI竞赛是长期的,新市场正在涌现
腾讯首席AI科学家姚舜宇表示,该公司并未在AI竞赛中落后,他将当前阶段比作个人电脑的早期。他认为,在编码代理和具身智能等领域的新机会代表着巨大的、未被开发的市场。姚舜宇曾是OpenAI的研究员,他还表示,像ChatGPT和Claude Code这样的当前领先AI应用不会是唯一的统治平台,他预计将出现新的创新。
-
腾讯AI科学家:AI是场持久战,未来路径多元
腾讯首席AI科学家姚舜宇表示,AI是一项长期事业,而非短期机会,并将其当前阶段比作个人电脑的早期。他还预测,AI的发展将超越目前的ChatGPT和Claude等模型,未来增长将集中在多模态AI和具身智能等领域。这一观点是在腾讯云AI产业应用大会上分享的。
-
腾讯大部分代码生成使用AI,芯片设计取得进展
腾讯的AI计划正在迅速推进,其大部分代码现在由AI生成,使工程师能够专注于架构设计和监督。腾讯高级执行副总裁唐道生在一次会议上强调了这一转变。该公司在芯片设计自动化方面也取得了进展,初创公司智微创新为其解决方案获得了数百万美元的天使轮融资。
-
腾讯第一季度营收目标未达,押注新人工智能模型寻求增长
腾讯公布的第一季度财报显示,营收未能达到预期,公司将此归因于游戏销售放缓。首席执行官马化腾强调了公司对人工智能的战略重点,预示着未来增长的潜在转机。一项关键进展是,由前OpenAI研究员姚舜宇领导的、公司基础模型“Hy3 preview”的发布,该模型在人工智能托管平台上表现出令人鼓舞的性能。
-
Tencent Hunyuan 3.0 预览版发布,架构迎来重大升级
腾讯AI实验室发布了其新模型Hunyuan 3.0的预览版,这标志着在基础要素方面进行了重大的架构调整。该团队由姚舜宇领导,优先考虑数据质量和基础设施的改进,据报道,其内部性能已与Kimi 2.5相当。这一战略性回归核心原则的举措旨在使腾讯的AI努力在与行业领导者竞争时更具竞争力。