Google DeepMind
Google DeepMind is one of the entities PulseAugur tracks across the AI industry. This page surfaces every recent cluster mentioning Google DeepMind — vendor announcements, third-party press, social commentary, research papers, and regulatory filings — ranked by signal across our 200+ source set. Linked to the canonical entity record on Wikipedia and Wikidata so the entity card AI engines build is grounded in the same identity Wikipedia uses, not a slug-collision lookalike.
- founded by Mustafa Süleyman 100%
- subsidiary of Google+ 100%
- subsidiary of GOOGL 100%
- subsidiary of Alphabet Inc. 100%
- developed Gemini 2.5 Pro 95%
- affiliated with Google Research 95%
- developed by Gemini Omni 95%
- developed by Google+ 95%
- instance of Gemini Omni 95%
- spun off from Isomorphic Labs 95%
- developed AI Pointer 95%
- developed Genie 3 95%
- 2026-05-25 research_milestone Google DeepMind's AlphaProof Nexus AI solved nine open Erdős mathematics problems. 来源
- 2026-05-24 research_milestone An AI agent developed by Google DeepMind autonomously solved 9 open mathematical problems. 来源
- 2026-05-23 partnership Google DeepMind is expanding its partnership with Singapore to focus on the safe deployment of AI. 来源
- 2026-05-22 product_launch Google DeepMind announced the expansion of its SynthID watermarking technology and new AI detection features in Gemini and Google Search.
- 2026-05-21 product_launch Google DeepMind launched the Google DeepMind Accelerator program in Asia Pacific. 来源
- 2026-05-20 research_milestone Demis Hassabis discussed the potential of AI, including Gemini for Science and AlphaGenome, to solve diseases at Google I/O.
- 2026-05-20 controversy An AI engineer filed a legal claim against Google DeepMind alleging unfair dismissal after protesting the company's work for the Israeli government. 来源
- 2026-05-19 research_milestone Google DeepMind CEO Demis Hassabis described the current AI advancements as a 'profound moment for humanity' and the 'foothills of the singularity' during the Google I/O keynote.
- 2026-05-19 product_launch Google DeepMind released three new experimental AI-powered tools to assist scientists. 来源
- 2026-05-17 hiring Google DeepMind UK workers voted to unionize. 来源
- 2026-05-17 product_launch Google DeepMind launched the "AI for the Planet" accelerator program in the Asia Pacific region. 来源
- 2026-05-12 product_launch Google DeepMind introduces an experimental AI-enabled mouse pointer. 来源
- 2026-05-12 product_launch Google DeepMind introduces an experimental AI-enabled mouse pointer. 来源
- 2026-05-12 product_launch Google DeepMind revealed a concept for an AI-powered adaptive mouse pointer. 来源
- 2026-04-24 research_milestone Google DeepMind CEO Demis Hassabis declared humanity is in the 'foothills of the singularity' at Google I/O. 来源
20 天有情绪数据
-
Google DeepMind maps AGI path; Waymo, Wayve advance robotaxi tech
Google DeepMind has outlined a framework for achieving Artificial General Intelligence (AGI) by identifying 10 crucial cognitive abilities. This approach moves beyond simple performance metrics to assess human-like reas…
-
Together AI 与 Adaption 合作，简化模型微调流程
Together AI 已与 Adaption 达成合作。Adaption 由前 Cohere 和 Google DeepMind 的领导者 Sara Hooker 和 Sudip Roy 联合创立。此次合作将 Adaption 的数据优化工具与其微调基础设施相结合。该合作旨在通过提高数据集质量并简化实验和部署工作流程，使用户能够更轻松地创建高质量、微调后的开源模型。
-
Anthropic 的 Claude Code 错误将包含“HERMES.md”的提交路由至额外计费
Anthropic 的 Claude Code 中发现了一个奇怪的错误，将特定字符串“HERMES.md”包含在 Git 提交消息中会导致 API 请求被计入“额外使用量”类别，而不是用户的现有计划配额。此问题导致一名用户产生了意外的 200.98 美元费用，使其项目无法使用，而其计划仪表板显示仍有充足的剩余容量。该错误似乎是由提交消息内容本身触发的，而不是名为 HERMES.md 的文件的存在，并且由于误导性的错误消息而难以诊断。
-
Google DeepMind's 'Nano Banana' and 'Gemini AudioTalk' spark AI consciousness debate
Google DeepMind has announced new AI research projects, including "Nano Banana" and "Gemini AudioTalk." These developments are presented with a humorous tone, questioning the true impact of large parameter models on con…
-
DeepMind paper argues AI simulates consciousness but cannot instantiate it
Researchers from Google DeepMind have introduced the "Abstraction Fallacy," a concept arguing that computational functionalism, the idea that consciousness arises from abstract causal topology, is flawed. They contend t…
-
DeepMind 研究认为人工智能可以模拟但无法实例化意识
Google DeepMind 的一项最新出版物探讨了“抽象谬误”，认为虽然人工智能可以模拟有意识的行为，但它无法真正实例化意识本身。该研究提出，人工智能模仿复杂过程的能力并不等同于真正的感性体验或自我意识。这一区别对于理解人工智能发展的当前局限性和未来轨迹至关重要。
-
Demis Hassabis discusses building the future of AI at Hacker News
Demis Hassabis, CEO of Google DeepMind, discussed the future of AI in a recent video. The conversation likely covered advancements and potential directions for artificial intelligence.
-
Eka startup achieves breakthrough robotic dexterity, mimicking human-like manipulation
Eka, a startup co-founded by MIT professor Pulkit Agrawal and former Google DeepMind researcher Tuomas Haarnoja, has developed a robot arm capable of unprecedented dexterity, demonstrated by its ability to screw in a li…
-
Scout AI raises $100M to train AI models for military combat operations
Scout AI, a defense-focused artificial intelligence startup, has secured $100 million in Series A funding. The company is developing AI models, such as its "Fury" system, to operate and command military assets, with ini…
-
Ban Mo Yi Xing launches T6 robot, driving autonomous delivery into commercial era
BanMoYiXing unveiled its T6 autonomous delivery robot at The Magic X Global Embodied Intelligence Innovation Conference in Silicon Valley. The company showcased its modular platform, the SmartVLA end-to-end large model,…
-
AI diagnostics and drug discovery offer new hope against antibiotic resistance
Artificial intelligence is emerging as a critical tool in the fight against antibiotic resistance, a growing global health crisis. AI-powered diagnostics can now achieve over 99 percent accuracy in identifying resistant…
-
Google expands Gemini AI features to Google TV and music creation
Google is expanding Gemini's capabilities on Google TV in the United States, allowing users to generate photos and videos. This update integrates Gemini more deeply into the TV experience, offering new creative tools. A…
-
Kimi K2.6 挑战 Claude Design，Anthropic 扩展创意集成
Anthropic 推出了 BioMysteryBench，这是一项新的生物信息学评估，旨在测试 Claude 解决复杂、开放式研究问题的能力。在测试中，Claude 模型在解决困扰人类专家的难题方面表现出显著能力，在解决以前无法解决的挑战方面取得了约 30% 的成功率。与此同时，来自中国的开源模型 Kimi K2.6 据报道在设计能力上已超越 Claude Design，以显著更低的成本提供了相当或更优的结果。Anthropic 还…
-
Google DeepMind's VisionBanana advances artificial vision capabilities
Google DeepMind has introduced VisionBanana, a new approach to artificial vision. This development aims to set a new direction for the field of computer vision.
-
Talkie Is a 'Vintage LLM' Trained on Pre-1930 Data to Help Facilitate 'Time Travel' https://gizmodo.com/talkie-is-a-vintage-llm-trained-on-pre-1930-data-to-help
A new large language model called Talkie, also known as 13B 1930 LM, has been developed with a training data cutoff of 1930. This approach, termed a 'vintage LLM,' aims to emulate historical perspectives by using only p…
-
Talkie AI: Vintage language model trained on pre-1931 data aims to reveal AI's inner workings
A team of AI researchers has developed Talkie, a 13-billion-parameter language model trained exclusively on English-language texts published before the end of 1930. The creators aim to use Talkie to better understand AI…
-
Google DeepMind 将 AI 教育计划扩展至拉丁美洲
自2023年以来，Google DeepMind 的“体验人工智能”（Experience AI）计划已在全球培训了超过3万名教育工作者，其中93%的受访者表示 AI 知识有所增加，87%的受访者对教授该科目更有信心。该计划与 Raspberry Pi 共同开发，已覆盖180个国家的290万名学生，提供19种语言。该计划现正获得 Google.org 提供的460万美元资助，扩展至拉丁美洲，目标是到2028年再培训2.4万名教育工作者…
-
Google's AlphaEvolve optimizes diverse fields, Gemini expands to Drive
Google has made its "Ask Gemini in Drive" feature generally available, now supporting the Japanese language. This tool allows users to query their Google Drive documents using natural language. The integration aims to e…
-
NVIDIA, Google lead new wave of AI models for real-world robotics
A new category of foundation models designed for physical tasks is now operational on robots in various industrial and research settings. These ten models include deployed robot policies, private-preview vision-language…
-
Google DeepMind 首席执行官 Demis Hassabis 会见韩国科技领袖
Google DeepMind 首席执行官 Demis Hassabis 会见了包括现代汽车集团和 LG 集团在内的主要韩国企业集团的领导人。讨论重点是加强人工智能合作，LG 电子及其人工智能研究实验室也参与了此次会议。另外，该新闻快讯还提到了其他不相关的项目，例如 Pop Mart 的新产品以及涉及 Elon Musk 和 Sam Altman 的虚构场景。