[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$ftM3JTSKZygvnUtlESM5sxtZ_5lqftmmqG7f-yiF9HTQ":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1229,"DeepMind哈萨比斯：从AlphaGo到全能AI，AI需跨越理解与工具使用鸿沟","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">在最新一期的《Release Notes》播客节目中，DeepMind的首席执行官德米斯·哈萨比斯深入探讨了公司技术的最新进展及其长远规划。他提到，从AlphaGo在游戏领域的辉煌成就，到如今Deep Think向通用人工智能（AGI）迈进的尝试，这一历程彰显了AI从专门化向全面化发展的必然趋势。在哈萨比斯看来，具备思考与规划能力是通往AGI不可或缺的一环，而DeepMind的最终愿景，是打造一款集语言理解、多媒体处理、物理推理及内容生成于一体的全能模型——Omni Model。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">节目中，哈萨比斯还介绍了DeepMind在物理世界理解方面的新进展，以Genie 3为例，该模型能够构建出逻辑自洽的虚拟世界，这对于AGI在现实环境中的操作与理解至关重要。然而，他也坦诚指出，当前AI系统在特定领域表现出色，但在面对简单任务时却可能栽跟头，这种不一致性正是阻碍其迈向AGI的关键因素。他进一步强调，未来的通用AI系统需具备理解用户真实意图的能力，并将这一理解转化为可优化的奖励机制，这背后涉及到元认知或“元强化学习”的深入研究。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">哈萨比斯还特别提到了工具使用在AI系统能力构建中的重要性。他认为，一个真正的思考型系统，应当能够在思考过程中主动调用并利用各种工具。展望未来，他透露DeepMind旗下的Genie、Veo、Gemini等模型将逐步融合，最终进化为全能模型Omni Model，这将是AGI所应具备的综合能力的集大成者。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】ITBEAR科技资讯 \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.msn.cn\u002Fzh-cn\u002Fnews\u002Fother\u002Fdeepmind%E5%93%88%E8%90%A8%E6%AF%94%E6%96%AF-%E4%BB%8Ealphago%E5%88%B0%E5%85%A8%E8%83%BDai-ai%E9%9C%80%E8%B7%A8%E8%B6%8A%E7%90%86%E8%A7%A3%E4%B8%8E%E5%B7%A5%E5%85%B7%E4%BD%BF%E7%94%A8%E9%B8%BF%E6%B2%9F\u002Far-AA1Ktxga?ocid=msedgntphdr&cvid=1502c407b8c44ed084e126f1e8b85c37&ei=21\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\"> http:\u002F\u002Fu5a.cn\u002F1k6zu\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002F7918d778d2814fababa6016af1be64f4\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F08\u002Fthumbs\u002F7918d778d2814fababa6016af1be64f4\u002FAI领域.jpg",0,1,219,"2025-08-15 17:12",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3Afb6257ab-209b-4ae8-be25-949081e1671c%3A0.wav?Expires=1755255144&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=19KkrVg8I%2Bd3kssYDIpRVtOEy74%3D",3354060,"fb6257ab-209b-4ae8-be25-949081e1671c","2025-08-15 17:10","Demis Hassabis of DeepMind: From AlphaGo to General AI, AI Needs to Bridge the Gap Between Understanding and Tool Use","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">In the latest episode of the \"Release Notes\" podcast, the CEO of DeepMind, Demis Hassabis, delved into the company's latest technological advancements and long-term plans. He mentioned that from the remarkable achievements of AlphaGo in the field of games, to the current attempts of Deep Think to move towards general artificial intelligence (AGI), this journey demonstrates the inevitable trend of AI's development from specialization to comprehensiveness. In Hassabis' view, the ability to think and plan is an indispensable part of reaching AGI, and DeepMind's ultimate vision is to create a comprehensive model called Omni Model that integrates language understanding, multimedia processing, physical reasoning, and content generation.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In the program, Hassabis also introduced DeepMind's new progress in understanding the physical world, taking Genie 3 as an example. This model can build logically consistent virtual worlds, which is crucial for AGI to operate and understand in real environments. However, he also frankly pointed out that current AI systems perform well in specific fields but may stumble on simple tasks, and this inconsistency is a key factor hindering their path to AGI. He further emphasized that future general AI systems need to have the ability to understand users' true intentions and convert this understanding into an optimized reward mechanism, which involves in-depth research on metacognition or \"meta-reinforcement learning.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Hassabis also specifically mentioned the importance of tool usage in building AI system capabilities. He believes that a true thinking system should be able to actively call and use various tools during the thinking process. Looking ahead, he revealed that DeepMind's models such as Genie, Veo, and Gemini will gradually merge, eventually evolving into the comprehensive model Omni Model, which will represent the culmination of the comprehensive abilities that AGI should possess.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【News Source】ITBEAR Technology News \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.msn.cn\u002Fzh-cn\u002Fnews\u002Fother\u002Fdeepmind%E5%93%88%E8%90%A8%E6%AF%94%E6%96%91-%E4%BB%8Falphago%E5%88%B0%E5%85%A8%E8%83%BDai-ai%E9%9C%80%E8%B7%A8%E8%B6%8A%E7%90%86%E8%A7%A3%E4%B8%8E%E5%B7%A5%E5%85%B7%E4%BD%BF%E7%94%A8%E9%B8%BF%E6%B2%9F\u002Far-AA1Ktxga?ocid=msedgntphdr&cvid=1502c407b8c44ed084e126f1e8b85c37&ei=21\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\"> http:\u002F\u002Fu5a.cn\u002F1k6zu\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is reprinted by this site to provide readers with more information and news. The content involved does not constitute investment or consumption advice. If there are any questions about the facts of the article, please verify with the relevant parties. The views of the article are not the views of this site and are for reference only.)\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A296c3520-7cd4-40ad-a99b-f27c40abae5f%3A0.wav?Expires=1774838495&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=3lap4HXy7heUoxPy0GIi3USMDXg%3D","296c3520-7cd4-40ad-a99b-f27c40abae5f",4189064]