DeepMind CEO Hassabis Reaffirms AGI Arrival by 2030

DeepMind CEO哈萨比斯重申AGI将于2030年到来 Google DeepMind首席执行官德米斯·哈萨比斯认为，通用人工智能将在六年内实现，为AI将长达十年的药物研发进程缩短至数天铺平道路。 德米斯·哈萨比斯回顾了自己从国际象棋神童、游戏开发者到神经科学研究员及DeepMind创始人的职业历程，阐述了早年经历如何指引他确立在2030年前实现AGI的20年规划使命。他解释了AlphaFold和Isomorphic Labs等AI模型如何将药物发现周期从十年大幅压缩至数天，并变革生物学、物理学和天气预报等领域的科学研究。此外，哈萨比斯还探讨了高保真AI模拟、将信息视为现实基本实体的观点，以及围绕AI智能体与意识的哲学思考。 Demis Hassabis 回顾了 DeepMind 在 2010 年创立时的四大核心洞察（深度学习、强化学习结合、加速计算与计算神经科学），并重申其制定的 20 年 AGI 路线图进展顺利，预计将在 2030 年左右实现通用人工智能。 (4:40) 回顾早期创业经历时，Hassabis 总结出科技创业应“领先时代 5 年，而非 50 年”的法则；当年 Elixir 试图在初代奔腾电脑上模拟百万人动态系统的尝试证明，过早超前硬件能力将导致执行难以成功。 (2:40) 继 AlphaFold 破解长达 50 年的蛋白质 3D 结构预测难题后，衍生公司 Isomorphic Labs 正致力于通过计算手段直接设计新型化合物分子，目标将原本长达 10 年的药物研发流程压缩至数天。 (9:05) Hassabis 提出信息而非能量或物质才是宇宙最基本的存在形式，并指出机器学习是描述高噪、复杂生物系统的原生语言，正推动虚拟细胞与多领域高保真模拟科学的发展。 (15:00)

DeepMind co-founder and CEO Demis Hassabis argues that artificial general intelligence is progressing on schedule toward a 2030 arrival, serving primarily as an empirical tool to unlock fundamental scientific breakthroughs and simulate complex emergent systems. Strategic Origins and the Path to AGI (0:00) Hassabis traced his career trajectory through chess and video game development at Bullfrog and Elixir Studios as deliberate preparation for AI research, noting his use of early GPUs in the late 1990s and his co-design of Theme Park at age 17. His first startup, Elixir Studios, taught him the operational pitfall of being too far ahead of contemporary technology when Republic: The Revolution attempted to simulate one million dynamic citizens on late-1990s Pentium PCs. Hassabis noted that entering a space when it is obvious is too late, but being decades ahead makes execution nearly impossible. DeepMind was founded in 2009–2010 around the convergence of deep learning, reinforcement learning, neuroscience principles, and accelerated compute. Despite skepticism from an academic establishment focused on 1990s-era rule-based systems, DeepMind mapped out a 20-year roadmap to achieve AGI by roughly 2030. Hassabis maintains that the field remains directly on track to meet that timeline. Transforming Scientific Discovery and Biology (9:05) DeepMind established its dedicated AI for Science division following AlphaGo's 2016 victory, leading to AlphaFold's solution to the 50-year-old challenge of predicting 3D protein structures. To commercialize and extend this capability, Hassabis founded Isomorphic Labs, which applies biochemistry AI models to design chemical compounds that bind selectively to target protein sites while minimizing off-target side effects. Hassabis projects that moving 99% of drug design in silico could compress typical 10-year pharmaceutical timelines down to months, weeks, or days. DeepMind is expanding this paradigm through learning simulators, which Hassabis considers the natural descriptive language for biology—a domain dominated by noisy data and dynamic interactions that resist traditional mathematical equations. Current initiatives include WeatherNext, which outpaces standard meteorological models in speed and accuracy, and a virtual cell project designed to dynamically simulate living systems. Simulating Emergent Systems and Information Theory (12:40) Beyond biology, Hassabis anticipates AI will establish mechanistic interpretability as a formal engineering discipline to unpack neural networks, while high-fidelity AI simulations will formalize human domains like macroeconomics into quantitative sciences. Because researchers cannot run repeated, controlled macro-policy experiments in the real world, AI simulators could allow policymakers to evaluate thousands of probabilistic counterfactuals prior to implementation. From a theoretical standpoint, Hassabis regards information as more fundamental than matter or energy, describing biological structures that resist entropy as information processing systems. He highlights that classical Turing machines—operating through modern neural networks like AlphaFold—can successfully approximate complex quantum biological systems with practical precision without needing to simulate every quantum variable directly. Tool-First Development and Questions of Consciousness (20:00) As AI transitions into an era of autonomous agents, Hassabis advocates prioritizing extremely capable, precise tools before addressing true agency. Drawing on Immanuel Kant's view that the mind constructs perception and Baruch Spinoza's perspective of scientific inquiry as uncovering the language of the universe, Hassabis views AI as a computational lens to study the human brain. On the subject of consciousness, Hassabis cites discussions with the late philosopher Daniel Dennett to separate behavioral imitation from experiential consciousness. While AI systems near AGI may mimic conscious behavior—such as self-awareness and temporal continuity—Hassabis cautions that verifying true experiential consciousness remains difficult because synthetic systems lack biological substrate equivalence with humans. Hassabis positions the final stretch to 2030 AGI as an engineering and scientific effort focused on powerful discovery tools rather than conscious autonomous entities. Unresolved questions remain regarding whether learning simulators can eventually yield explicit fundamental physical equations, and how researchers will evaluate claims of artificial consciousness in the absence of biological benchmarks.

DeepMind 联合创始人兼首席执行官 Demis Hassabis 在与红杉资本的对话中指出，人工智能的终极使命是作为推动基础科学突破的通用工具，并重申通用人工智能（AGI）有望在 2030 年左右实现。 创业节奏与技术结合的历史验证 (0:00) Hassabis 回顾其早期职业生涯指出，从国际象棋神童、游戏开发、计算神经科学到创办 DeepMind，背后存在一条清晰的技能积累主线。他在 17 岁时参与设计的模拟游戏《Theme Park》销量突破 1000 万份，验证了复杂底层 AI 系统能够创造巨大用户价值，并使他在 20 世纪 90 年代末便开始接触早期 GPU 硬件。随后创办 Elixir Studios 尝试制作全动态模拟百万居民国家的《Republic: The Revolution》，但在当时的 Pentium 家用个人电脑上遭遇硬件算力瓶颈。他由此总结出核心创业教训：“领先时代 5 年，而不是领先 50 年”；市场显而易见时入场已为时过晚，而领先半个世纪则几乎不可能依赖现有硬件落地。 逆流确立 AGI 路线与二十年长线规划 (4:40) 2009 年至 2010 年创立 DeepMind 之际，学术界主流（如麻省理工学院和剑桥大学）仍停留在 20 世纪 80 至 90 年代的专家系统与符号逻辑体系，普遍将 AGI 视作纯粹的科幻概念。Hassabis 及其联合创始人基于四项判断押注 AGI：Geoff Hinton 等人带来的深度学习突破、强化学习与深度学习交叉融合的潜力、GPU 等加速计算硬件的崛起，以及计算神经科学揭示的大脑学习机制。DeepMind 在 2010 年制定了为期 20 年的路线图（目标指向 2030 年），其总体纲领是“第一步解决智能（构建 AGI），第二步用智能解决其他一切问题”。Hassabis 强调这一时间规划至今仍在按预期推进。 AI 赋能科学与全流程计算化药物设计 (9:05) 在 2016 年 AlphaGo 战胜李世石后，DeepMind 组建了由 Pushmeet Kohli 领导的“AI for Science”专门部门，致力于将成熟算法应用于重大科学问题。其标志性成果 AlphaFold 解决了困扰学界 50 年的蛋白质三维结构预测难题。为实现端到端药物研发，Hassabis 进一步分拆成立 Isomorphic Labs，利用 AI 计算设计能够精确靶向蛋白质结合位点且规避脱靶副作用的新型化合物。传统药物研发平均耗时约 10 年，他的目标是将 99% 的药物设计过程转移至计算机模拟环境中完成，仅将湿实验室用作最终验证，从而把研发周期缩短至数月、数周甚至数天，使攻克更多疾病与实现个性化医疗成为可能。 学习型模拟器与信息本体论视角 (12:40) 针对传统数学方程难以直接精确表达的复杂动态系统，DeepMind 正在构建“学习型模拟器”。例如气象模型 WeatherNext 在速度与精度上均超越传统气象系统，而“虚拟细胞”项目则试图将机器学习确立为生物学的原生描述语言，以解析生物体内海量含噪且非线性的动态涌现行为。Hassabis 进而提出一种理论观点，认为信息而非能量或物质才是宇宙最基本的构件；AlphaFold 等成果表明，基于现代神经网络的经典图灵机架构无需在底层精确计算每个微观量子态，即可在实用层面近似高度复杂的生物系统。同时，高保真度 AI 模拟未来还有望应用于经济学和社会科学等缺乏可重复实证实验的领域。 安全构建工具与机器意识的哲学反思 (20:00) 在面对高度自主化 AI 的发展方向时，Hassabis 主张现阶段最稳妥有效的路径是先打造极其精准、高能力的智能“工具”，而非过早跨越界限追求自主代理或机器意识。他结合与哲学家 Daniel Dennett 的过往探讨，将意识区分为行为表现与内在体验两个维度；即便接近 AGI 的系统能够在行为层面模拟意识，但人工系统缺乏人类共有的生物基质，使得体验层面的意识难以被确证。在哲学指引上，他借鉴康德关于“心智构建现实”的论点来探索神经认知，同时汲取斯宾诺莎的自然哲学思想，将利用 AI 探索宇宙运行规律视为一种解读“宇宙语言”的崇高探索。 Hassabis 维持了通用人工智能将于 2030 年前问世的预期，并将 AlphaFold 视为 DeepMind 迄今最自豪的里程碑。然而，从工具型智能向具备完全自主性的系统演进时，如何确证机器体验层面的意识，以及如何将宏观社会的复杂系统转化为完全严密的定量科学，仍是尚未彻底解决的开放性课题。

## Background 该内容由红杉资本（Sequoia Capital）于 2026 年 4 月 30 日通过 RSS 及播客平台发布，记录了红杉合伙人 Konstantine Buhler 在 AI Ascent 2026 大会上与 Google DeepMind 联合创始人兼 CEO、2024 年诺贝尔化学奖得主 Demis Hassabis 的深度对谈。官方简介指出，对话聚焦于通往 AGI 的最后冲刺与未来展望，涉及 Hassabis 为何认为 2030 年可实现 AGI、药物研发周期如何从十年压缩至数天，以及为何应将信息而非物质或能量视为宇宙最底层实在等前沿议题。 ## TL;DR Demis Hassabis 回顾了自己从国际象棋神童、游戏开发者到认知神经科学家的跨界历程，阐明其毕生追求就是为了创立 DeepMind 并用“深度学习+强化学习”攻克 AGI。他详细拆解了“AI for Science”的深远布局，指出 AlphaFold 与 Isomorphic Labs 正在将传统长达十年的新药研发周期缩短至数天，并将机器学习定位为解析生物学复杂涌现系统的“微积分”。在哲学与科学底层层面，Hassabis 提出信息比质能更为本质，主张构建 AGI 应坚守“先打造极致工具，再探索意识与主体性”的两步走路线。最后，他重申了 2030 年实现 AGI 的坚定预测，并推荐《真实世界的脉络》作为理解现实与未来科学的重要著作。 ## Atmosphere & style 整场对谈兼具极客的松弛感与学术的深邃感，开场以“巧克力工厂”的幽默调侃自然切入，随即迅速深入高密度的智性交锋。Hassabis 的表达极为冷静、严谨且条理分明，毫无浮躁的商业宣讲气息，而是以跨越神经科学、计算理论与哲学的广阔视野娓娓道来。听感上既充满直面技术现实与工程试错的务实理性，又洋溢着解码宇宙语言与生命本质的终极浪漫，展现出顶尖科学家引领时代的笃定与沉稳。 ## Bulletin - `25:00` Demis Hassabis 重申其对 AGI 落地时间表的连贯预测，认为通用人工智能将在 2030 年左右实现。 - `10:00` DeepMind 旗下 Isomorphic Labs 致力于将药物研发中 99% 的试错转移至计算机内（in silico）完成，有望将新药研发周期从传统的 10 年大幅压缩至数天或数周。 - `17:15` Hassabis 提出信息比能量与物质更基础的宇宙观，强调经典图灵机足以为宏观量子效应建模，机器学习则是生物学这一强涌现系统的“微积分”。 - `20:00` 在走向 AGI 的路径上，Hassabis 坚持分两步走策略：首先集中全力打造极致的智能工具，之后再借助该工具跨越意识与主体性的“卢比孔河”。 ## Key takeaways - `00:00` Hassabis 自少年时期便确立了以 AGI 为毕生追求的目标，早期的游戏开发与神经科学学术背景皆为创立 DeepMind 所积累的技能拼图。 - `02:50` 早年创办 Elixir Studios 研发超前游戏受挫，让 Hassabis 得出深刻创业教训：技术探索应当领先时代 5 年，而非领先 50 年。 - `05:00` DeepMind 在 2010 年创立时便制定了 20 年攻克 AGI 的战略，敏锐押注“深度学习 + 强化学习 + 神经科学启发 + 加速计算”的跨界融合路径。 - `07:15` 攻克 AGI 只是元手段（Meta-way），其终极使命是打造能解决自然科学、疾病医学与宇宙奥秘等人类重大挑战的通用工具。 - `10:00` AlphaFold 攻克蛋白质三维折叠结构难题，标志着生物学迎来了属于自己的大模型爆发时刻，也成为 Hassabis 迄今最自豪的成就。 - `10:00` Isomorphic Labs 正在推动计算驱动的药物设计革命，通过高特异性分子生成与虚拟筛选，使个性化医疗与全疾病攻克成为可能。 - `12:38` AI 发展将催生两大全新科学分支：透视复杂神经网络机理的“可解释性工程学”，以及用于经济学、社会学等强涌现系统的“高保真学习型模拟器”。 - `15:00` 机器学习是生物学的最佳描述语言，DeepMind 正在构建“虚拟细胞”，未来有望从无成本采样的隐式模拟器中反向提炼出生物学的底层基础方程。 - `17:15` 宇宙最底层的本质是信息而非能量或物质，经典图灵机完全能够有效近似并建模宏观生物系统中的微观物理与量子现象。 - `20:00` AGI 演进应坚持分两步走：先聚焦于构建高通用性的辅助工具，再利用其算力与洞察反哺对人类意识、心智及主体性边界的深入研究。 - `21:40` 人工系统与人类之间缺乏生物基质等价性，这导致单纯的行为模仿与真正的主观内在体验之间始终存在难以逾越的鸿沟。 - `23:45` 受康德的心智建构论与斯宾诺莎的宇宙观启发，Hassabis 将打造科学 AI 工具视作人类解码与阅读“宇宙语言”的崇高过程。 - `25:00` Hassabis 推荐大卫·多伊奇的《真实世界的脉络》作为 AGI 达成后的必读书目，并称博弈论大师冯·诺依曼是高风险策略对决中的终极梦幻队友。 ## Detailed walkthrough ### `00:00` 贯穿一生的主线：从国际象棋神童、游戏开发者到神经科学家 对话在红杉资本活动现场轻松开场，主持人 Konstantine Buhler 调侃欢迎 Demis Hassabis 来到他们的“巧克力工厂”，Demis 也幽默回应期待活动后的巧克力。作为 AI 领域的元老级人物（OG）和 2024 年诺贝尔化学奖得主，Demis 拥有看似跨度极大的人生履历：少年时期的国际象棋神童、爆款游戏公司创始人、认知神经科学学者，再到如今掌舵全球顶尖 AI 机构 Google DeepMind。面对主持人关于这些角色看似毫不相干的提问，Demis 坦言虽然这可能有事后归因的成分，但在他心中始终存在一条无比清晰的主线——早在十五六岁时，他就认定 AI 是世界上最重要、最迷人的事业，此后所学、所做的每一步，全都是为了有朝一日创办一家像 DeepMind 这样的公司而提前铺路。 在上世纪 90 年代转向游戏领域并非无心插柳，而是因为那是当时最前沿技术的聚集地。不仅是 AI 算法，图形学硬件也在游戏行业孕育爆发——如今大行其道的 GPU 最早正是为了驱动游戏引擎而诞生，Demis 在 90 年代末便已是最早一批使用初代 GPU 的开发者。他制作的每款游戏都以 AI 为核心玩法：大约 17 岁时，他在传奇工作室牛蛙制作（Bullfrog Productions）参与打造了里程碑式的模拟经营游戏《主题公园》（*Theme Park*）。游戏中成千上万的小人在游乐园中闲逛、游玩并决定购物，底层运转着一套完整的经济学 AI 模型。当看到这款游戏与《模拟城市》（*SimCity*）并驾齐驱并卖出超过 1000 万份、玩家因与底层 AI 互动而欣喜若狂时，Demis 更加坚定了倾尽一生投入 AI 的信念；而随后的神经科学研究，则是为了从大脑运作机制中汲取算法灵感，最终将这些积累汇聚成创立 DeepMind 的原动力。 - 核心志向的确立：约 15、16 岁即选定 AI 为毕生追求，后续求学与创业皆为实现该目标而设计的技能拼图。 - 早期代表作：17 岁主导/参与开发的《主题公园》全球销量突破 1000 万套，首创基于 AI 驱动的自主游客群体行为与经济模拟。 - 跨领域融合：将 90 年代顶尖游戏研发（软硬件图形技术及初代 GPU 运用）、计算神经科学与通用智能算法结合，构成了 DeepMind 的底层逻辑。 ```mermaid flowchart LR accTitle: Demis Hassabis 多领域融合与 AGI 路径 accDescr: 展示 Hassabis 从游戏研发、神经科学到创立 DeepMind 的技术演进历程 subgraph Games["游戏工业（1990年代）"] G1["底层 AI 交互<br/>《Theme Park》"] G2["探索前沿算力<br/>初代 GPU 应用"] end subgraph Neuro["计算神经科学"] N1["大脑工作机制"] N2["强化学习原则"] end subgraph DeepMind["DeepMind 创办与突破"] DM1["深度学习 + 强化学习"] DM2["两步走战略：<br/>1. 攻克通用智能<br/>2. 解决一切科学问题"] end Games --> DM1 Neuro --> DM1 DM1 --> DM2 ``` --- ### `02:50` 第一次创业教训：要领先时代 5 年，而非 50 年 走出大学校门后，Demis 创立了自己的第一家公司 Elixir Studios。当时他在欧洲最负盛名的牛蛙工作室学到了精湛的制作工艺，创办 Elixir 的实质，其实是通过游戏研发“走后门”为自己的 AI 研究筹措资金，试图将最尖端的算法探索与天马行空的创意结合起来。这种自由探索的基因也延续到了如今 DeepMind 对前沿蓝天研究（Blue-sky Research）的支持方式上。然而，第一次创业也给他上了刻骨铭心的一课——“你应当领先你的时代 5 年，而不是领先 50 年”。 在 Elixir Studios 时期，团队试图打造一款名为《共和国：革命》（*Republic: The Revolution*）的雄心之作。游戏构想了一个被独裁者统治的完整虚拟国家，玩家可以通过政治、文化等多种途径推翻政权。游戏试图在当时奔腾（Pentium）处理器的家用 PC 上，实时模拟出一座拥有上百万自主行动居民的庞大动态城市，渲染每栋建筑并让每个市民具备独立的 AI 决策逻辑。在 90 年代末有限的硬件条件下实现这一设想显得过于超前，最终导致了严重的研发困境。Demis 从中深刻体会到：当一项技术变得显而易见时，入局已经太迟；但如果领先时代多达半个世纪，残酷的基础设施现实会让你根本无法将其做成。 --- ### `05:00` 2009-2010 创立 DeepMind：手握秘密的少数派与 20 年 AGI 蓝图 到了 2009 年前后，Demis 意识到创办 DeepMind 的时机已经成熟——这一次，他们大概只比主流认知早了 10 年，而非难以跨越的 50 年。当时，Geoffrey Hinton 等学者刚刚在学界提出现代深度学习的雏形，但几乎没人意识到这是一场技术革命；与此同时，Demis 团队在强化学习领域造诣颇深。在当时的学术界，深度学习与强化学习是各自孤立的两个细分分支，除了少数玩具级问题外极少有人尝试将其混合。凭借敏锐的直觉，Demis 坚信二者的结合将释放出不可思议的能量。加之看到 GPU 等加速计算硬件的爆发势头，以及自身在神经科学博士和博士后研究中领悟到的人脑运作原理，团队坚信强化学习具备一路拓展放大至 AGI 的潜力。 在那个年代宣称要构建 AGI 或“强 AI”（Strong AI），在主流学术界和工业界看来完全是天方夜谭。Demis 回忆道，当他们提出要做 AGI 时，学术界同行几乎是在“翻白眼”，冷嘲热讽称“大家都知道这行不通”。因为 90 年代的人们曾在传统符号主义上撞得头破血流，Demis 在 MIT 做博士后期间，那里依然是专家系统和一阶逻辑语言系统的中心堡垒，剑桥亦然。然而，正因为传统老路已经走入死胡同，反而让 Demis 更加笃定自己正在通往正确的方向：“至少就算我们要失败，我们也是以一种全新的、不同于 90 年代前辈的方式失败。” 这种“失败也要失败得有独创性”的信念，坚定了早期核心团队的信心。 - 远见验证：2010 年创立 DeepMind 时，团队便设定了“20 年攻克 AGI”的战略时间表；站在当前节点回望，整个领域的演进节奏几乎与他们当年的预测完全吻合。 - 破除偏见：当时学界因传统专家系统的溃败而对 AGI 避之不及，DeepMind 反其道而行之，将“深度学习 + 强化学习 + 神经科学启发 + 加速硬件”确立为破局四要素。 - 经典使命宣言：DeepMind 成立之初写下的两步走战略极为纯粹——“第一步，解决智能问题（即构建 AGI）；第二步，用它来解决其他一切问题”。 --- ### `07:15` AI for Science：构建终极元工具，从 AlphaGo 凯旋次日启航 在 Demis 的构想中，研发通用人工智能从来不是终点，而是通往更大目标的阶梯。他将科学、医学以及人类对宇宙奥秘的深层理解视为真正的个人热情所在。对他而言，打造 AGI 是一种高级的“元任务”（Meta-way）——先锻造出人类历史上最强大的通用智力工具，再回过头来用它在自然科学领域引发颠覆性突破，正如后来一举破解蛋白质三维折叠难题的 AlphaFold。此外，AI 也是探索人类自身意识、梦境与创造力本质的绝佳参照物与对照实验工具。 为了确保这一使命不被商业浪潮冲淡，DeepMind 从成立早期就将科学探索深植于组织基因中。早在近十年前，DeepMind 就正式成立了由 Pushmeet Kohli 领衔的“AI for Science”专门部门。Demis 透露了一个极具象征意义的历史细节：这个科学部门正式立项启动的时间点，恰恰是团队结束在韩国首尔与李世石的世纪人机大战、带着 AlphaGo 获胜的荣耀刚返回伦敦的第二天。Demis 当时一直在克制等待，直到算法能力在围棋等极度复杂的搜索空间中证明了自己的威力，他便立刻按下按钮，全面开启了用 AI 解码基础科学未知领域的宏大征程。 ### `10:00` 蛋白质折叠之后：Isomorphic Labs 与计算驱动的药物研发革命 在攻克围棋之后，DeepMind 迎来了关键的转折点——团队确信底层通用技术已经足够成熟，可以转向解决具有重大现实意义的科学挑战。Demis Hassabis 强调，将人工智能应用于攻克疾病、延长健康寿命以及推动医学变革，始终被他视为 AI 最有价值的使命；紧随其后的则是材料科学、气候环境与清洁能源等领域。当谈及生物学何时能迎来类似大语言模型与代码生成那样的“爆发时刻”时，Hassabis 直言，AlphaFold 解决长达半个世纪的蛋白质三维结构折叠难题，本身就已经是生物学领域的这种时刻。 然而，解析蛋白质结构仅是漫长药物研发链路中的第一步。为此，DeepMind 孵化了独立子公司 Isomorphic Labs，专注于生物化学与化学设计领域。知道了目标蛋白质的精确三维外形与表面位点后，下一步是自动设计能够与靶点紧密结合的小分子化合物，同时必须严格避免与其它非靶点结合以消除毒副作用。Isomorphic Labs 的终极愿景是将药物研发中 99% 的试错与探索工作全部转移至计算机内（*in silico*）完成，把昂贵且耗时的湿实验（wet lab）仅保留在最后一步用于验证。 ```mermaid flowchart LR accTitle: 计算生物学药物研发新流程 accDescr: 对比传统实验驱动流程与以计算和模拟为主导的新型药物研发流程 subgraph TRAD["传统药物研发模式（平均耗时约 10 年）"] direction TB T1["盲目湿实验试错"] --> T2["化合物多轮修饰"] --> T3["漫长动物与临床验证"] end subgraph NEW["In Silico 新型研发模式（未来缩减至数周或数天）"] direction TB N1["AlphaFold 解析靶点结构"] --> N2["算法自动设计高特异性化合物"] N2 --> N3["99% 计算机内探索与虚拟筛选"] N3 --> N4["湿实验仅作最终靶向验证"] end ``` 这种研发范式的转变将彻底重构生物制药行业： - **研发周期坍缩**：将传统平均长达 10 年的新药研发周期大幅压缩至数月、数周，甚至未来有望缩短至几天以内。 - **全疾病攻克与个性化医疗**：当设计分子的成本与门槛指数级下降，所有疾病靶点理论上都将变得可被攻克，基于基础药物针对个体基因差异定制的“个性化医疗”（Personalized Medicine）也将真正成为可能。 --- ### `12:38` AI 催生新科学：从机械可解释性工程到复杂社会的模拟器 当被问及 AI 是否会像当年的工业革命催生热力学那样创造全新的基础学科并改写人类的教育体系时，Hassabis 给出了肯定的判断。他指出，未来至少会诞生两个全新的科学分支。首先是**针对 AI 系统本身的科学与工程学**。当今构建的前沿神经网络极其复杂，未来其复杂度甚至会媲美人类的大脑与心智，人类需要全新的分析工具来透视黑盒内部的工作机理。现阶段的机械可解释性（Mechanistic Interpretability）仅是这门新兴工程科学的雏形，未来还有广阔的未知有待建立。 其次，**AI 将作为强大的“模拟器”（Simulators）解锁人类过去无法严谨研究的复杂学科**。Hassabis 提到自己早期开发游戏时，所有作品的核心不仅有 AI，更是一套套精密的模拟系统。这种模拟能力正是攻克经济学、社会学等传统人文学科的钥匙： - **涌现系统的困境**：像经济学这样的宏观系统与生物学高度相似，属于强涌现系统，充满了海量弱信号与非线性关联，在现实世界中几乎无法进行严格受控的重复实验（例如央行为了观察结果将基准利率上调 0.5%，在现实中只能执行一次，无法进行上千次回滚与平行对照）。 - **高保真模拟采样**：一旦拥有极度精确的学习型模拟器，研究者便能在其中进行严密的、成千上万次的高频采样与压力测试，使人类在充满高度不确定性的决策领域中获得近乎硬科学的推演确定性。 --- ### `15:00` 机器学习是生物学的“微积分”：从天气预测到“虚拟细胞” 构建高精度的世界模型与模拟系统，核心在于打造“学习型模拟器”（Learning Simulators）。对于那些底层机理过于复杂、人类物理公式无法精确描述的动力学系统，传统的手工编码模拟器往往误差巨大。DeepMind 此前开发的天气预测模型展示了这一路径的威力，其精度与推演速度大幅超越了气象学界现有的数值模拟方案。现场主持人风趣地追问这是否意味着已经可以“改变天气”，Hassabis 笑着回答：“我们还做不到，而且我也不确定改变天气是不是个好主意，首要任务是先更好地理解它。” 更宏大的挑战在于生命科学，DeepMind 正在着手构建所谓的“虚拟细胞”（Virtual Cell）——一个庞大、高动态的涌现系统。Hassabis 提出了一个深刻的科学隐喻：**“正如数学是物理学的描述语言一样，机器学习正是生物学的绝佳描述语言。”** 物理学可以通过简洁优美的方程式推导，但生物学和自然界系统中充斥着数以亿计的微弱相关性、噪声与海量数据，早已超出人脑认知极限；现有数学工具要么无法应对这种维度，要么表达能力不足以刻画高度动态的涌现特征。 未来科学甚至可能出现一种全新的反向发现路径： - **从隐式直觉到显式公式**：AI 模拟器先通过海量数据建立起一个“隐式模拟器”（Implicit Simulator）；由于该模拟器可以被无限次无成本采样，科学家有望从中提取出显式的核心方程式。 - **生物学的麦克斯韦方程组**：如果这些涌现系统背后确实隐藏着类似电磁学中麦克斯韦方程组一样本质的基础定律，借助高保真模拟器的提炼，人类终将揭示它们的存在。 --- ### `17:15` 信息是宇宙最底层的实在：做“图灵的捍卫者” 在探讨宇宙深层机理时，Hassabis 分享了他更具哲学意味的物理观。爱因斯坦通过质能方程 $E=mc^2$ 确立了能量与质量的等价性，而 Hassabis 认为，**“信息”（Information）与质量、能量之间存在着同等深度的等价转换关系**。物质结构的组织形式，尤其是生物体抵御热力学熵增的生命活动，其本质都是信息处理系统。他甚至认为，信息是比能量和物质更为基础的宇宙本原，这颠覆了 20 世纪初经典物理学将质能视为首要实在的传统范式。如果宇宙的本质即信息，那么致力于组织、理解和重构信息的人工智能，其哲学地位将远比人们所想象的更为深邃。 谈及图灵机与计算的极限，Hassabis 将自己和 DeepMind 的探索者们自豪地称为“图灵的捍卫者”（Turing's Champions）： - **图灵机的通用威力**：阿兰·图灵是他一生中最崇敬的科学偶像，图灵机证明了任何在数学上可计算的事物，都可以通过相对简洁的机器描述来执行，人类大脑本质上大概率也是一台近似图灵机。 - **经典计算建模量子现象**：面对外界关于是否需要量子计算机才能完整模拟微观物理世界的争论，Hassabis 指出，AlphaGo 特别是 AlphaFold 已经做出了强有力的证明——以现代深度神经网络为表征的经典图灵机，完全有能力在宏观表征层面精确建模原本被视为强量子特性的复杂微观系统（如蛋白质折叠过程）。 ### `20:00` 经典计算模拟量子系统与“AI 工具论”的两步走策略 讨论从微观物理的建模切入。处理极微小的粒子时，人们通常直觉地认为必须把水分子键等所有的量子效应全部纳入考量；然而事实证明，通过在经典计算系统上进行巧妙的设计与思考，其实能够获得某种近似的最优解。这意味着，许多原以为必须依赖量子系统才能建模或运行的问题，只要思路正确，完全可以在现有的经典系统上模拟出来。 主持人顺势抛出一个更深层的根本问题：哈萨比斯一直以来都将 AI 描述为一种“终极工具”，就像几个世纪以来的望远镜、显微镜或星盘一样。但如果一台机器几乎能为世间万物建模——甚至包括量子系统，它何时会不再仅仅是一个工具？这一天究竟会不会到来？ 哈萨比斯对此给出了非常明确且务实的阶段性规划。他强烈认为，在构建通用人工智能（AGI）的征途上，最明智的做法是**分两步走**： 1. **第一步：打造极致的工具**。首先集中精力构建一个极其智能、实用且精密的工具，这本身就已经是极其深刻且影响深远的成就； 2. **第二步：跨越意识与主体性的“卢比孔河”**。尽管当前我们正身处“智能体时代”（Agent Era），工具正变得越来越自主化，但诸如“它是否真正具备自主意识（Agency）”或“它是否有知觉（Consciousness）”这类更深远的问题，应当作为第二步来探讨。届时，人类完全可以借助第一步所打造出的强大智能工具，反过来协助解决这些高维谜题，并更好地理解人类自身的大脑与心智，给意识下一个更精确的定义。 ```mermaid flowchart LR accTitle: AGI 演进两步走策略 accDescr: 展示从构建高精度智能工具到探索意识与主体性的两阶段路径 subgraph Phase1["第一阶段：工具属性"] A["高精度专用工具<br/>(如 AlphaFold)"] --> B["自主智能体时代<br/>(Agent Era)"] B --> C["高通用性 AGI 工具"] end subgraph Phase2["第二阶段：跨越卢比孔河"] C -.->|"作为探索研究工具"| D["理解人类大脑与心智"] D --> E["精确定义意识与主体性"] end ``` --- ### `21:40` 意识的定义困境：行为表现、主观体验与基质等价性 谈及意识的严格定义，哈萨比斯谦虚地表示，几千年的哲学史已经讨论得非常充分，自己很难凭空增添太多革命性定义。不过，他指出了构成意识不可或缺的几个核心要素——这些要素或许是**必要条件**，但绝非充分条件： - **自我觉察（Self-awareness）**； - **区分“自我”与“他者”的认知边界（Self and other）**； - **在时间维度上的连续性体验（Continuity over time）**。 在谈到具体的判定标准时，哈萨比斯回忆起多年前与著名哲学家丹尼尔·丹尼特（Daniel Dennett，近期不幸逝世）的一场长谈。他分析道，判断一个系统是否具有意识，通常存在两个核心维度： 1. **行为维度（Behavioral）**：系统表现得像一个有意识的系统吗？随着 AI 越来越逼近 AGI，人工系统完全可能在行为表象上做到这一点。 2. **基质等价性与体验维度（Experiential & Substrate Equivalence）**：人类之所以毫不怀疑彼此拥有意识，一方面是因为彼此的行为举止像有意识的生命，另一方面则是因为我们运行在完全相同的物理/生物基质（Substrate）上。基于奥卡姆剃刀的简约原则（Parsimonious），假设你正在经历我所经历的主观体验是最合理的推论。 哈萨比斯强调，对于人工系统而言，我们**永远无法获得基质上的等价性**。因此，在行为表现与真正的主观内在体验之间，始终存在一条难以彻底弥合的鸿沟。或许在实现 AGI 之后，人类能找到某种方法来探测这种主观体验，但在当下，这甚至超出了“AI for Science”的讨论范畴。 --- ### `23:45` 哲学底色：康德的心智建构与斯宾诺莎的宇宙语言 主持人提及哈萨比斯最喜欢的两位哲学家——代表义务论的康德（Immanuel Kant）与持决定论宇宙观的斯宾诺莎（Baruch Spinoza），并询问他如何将这两种截然不同的思想体系与自己的世界观联系起来。 哈萨比斯解释了他钟情于这两位哲学家的深层原因： - **康德（心智与现实）**：在攻读认知神经科学博士学位期间，康德关于“心智创造了我们所感知的现实”的论断深深触动了他。哈萨比斯坚信这一点基本属实。他最终追求的是理解“现实的本质”，但现实必须通过心智来解构与诠释，因此深入研究大脑与思维的运作机制就成了通向本质的必经之路。 - **斯宾诺莎（宇宙的灵性维度）**：斯宾诺莎的思想为他提供了一种近乎灵性维度的宏大视角。利用科学——以及哈萨比斯手力推的 AI 工具——去探索宇宙，本质上就是在以一种极为深刻的方式去理解宇宙运作的底层奥秘。 哈萨比斯动情地总结道，当他从事科学研究、开发 AI 并打造这些前沿工具时，他的内心感受就是：“**我们在以某种方式，阅读宇宙的语言**（Reading the language of the universe）。” --- ### `25:00` 快问快答：2030 年 AGI 预测、经典著作与梦幻团队 访谈进入最后的快问快答环节，哈萨比斯在没有提前获知问题的情况下作出了敏捷且极具个人特色的回答： - **关于 AGI 实现的年份预测**：哈萨比斯毫不犹豫地给出了 **2030 年**，并表示自己多年来在这个时间节点上的判断一直非常连贯一致。 - **AGI 达成之日必读的一本书、诗歌或论文**：他推荐了大卫·多伊奇（David Deutsch）的**《真实世界的脉络》（*The Fabric of Reality*）**。哈萨比斯将其视作自己最喜爱的著作，并坦言自己希望在 AGI 问世后的工作（Post-AGI work），就是利用 AGI 去彻底解答该书中提出的那些宇宙终极问题。 - **DeepMind 迄今最自豪的时刻**：尽管 DeepMind 创造了许多历史性成就，但哈萨比斯毫无悬念地选择了 **AlphaFold**。 - **高风险回合制策略游戏（如《文明》《Polytopia》）的队友选择**：如果能从人类历史上挑选一位科学家作为并肩作战的队友，哈萨比斯排除了爱因斯坦、图灵和牛顿，最终选择了**约翰·冯·诺依曼（John von Neumann）**。他笑着解释，在这种高难度博弈中，你最需要的是一位博弈论大师（Game theorist），而冯·诺依曼无疑是其中最顶尖的天才，“他绝对会是一个极棒的队友”。 ## Notable quotes > `02:50` 你应当领先你的时代 5 年，而不是领先 50 年。 — Demis Hassabis > `05:00` 至少就算我们要失败，我们也是以一种全新的、不同于 90 年代前辈的方式失败。 — Demis Hassabis > `05:00` 第一步，解决智能问题（即构建 AGI）；第二步，用它来解决其他一切问题。 — Demis Hassabis > `15:00` 我们还做不到，而且我也不确定改变天气是不是个好主意，首要任务是先更好地理解它。 — Demis Hassabis > `15:00` 正如数学是物理学的描述语言一样，机器学习正是生物学的绝佳描述语言。 — Demis Hassabis > `23:45` 我们在以某种方式，阅读宇宙的语言。 — Demis Hassabis > `20:00` 第一步：打造极致的工具。集中精力构建一个极其智能、实用且精密的工具，这本身就已经是极其深刻且影响深远的成就；第二步：跨越意识与主体性的‘卢比孔河’。 — Demis Hassabis > `21:40` 对于人工系统而言，我们永远无法获得基质上的等价性。 — Demis Hassabis ## Mentioned Demis Hassabis, Konstantine Buhler, Sequoia Capital, Google DeepMind, Bullfrog Productions, Theme Park, SimCity, Elixir Studios, Republic: The Revolution, Geoffrey Hinton, Pushmeet Kohli, AlphaGo, AlphaFold, Isomorphic Labs, Alan Turing, Albert Einstein, Daniel Dennett, Immanuel Kant, Baruch Spinoza, David Deutsch, The Fabric of Reality, John von Neumann, Polytopia, Civilization ## Video description Demis Hassabis, co-founder and CEO of Google DeepMind and 2024 Nobel laureate in chemistry for AlphaFold, joins Sequoia partner Konstantine Buhler at AI Ascent 2026 for a wide-ranging conversation about the path to AGI and what comes after. He explains why he believes AGI is achievable by 2030, why drug discovery could collapse from ten years to days, and why we should think of information, not matter or energy, as the most fundamental substance in the universe. Also: what Einstein would tell us about the limits of today's models, and why the next year or two will be critical for humanity. [Source](https://pscrb.fm/rss/p/traffic.megaphone.fm/CPUAI1446613297.mp3)

[2:50] Demis Hassabis: 你应当领先你的时代 5 年，而不是领先 50 年。 [5:00] Demis Hassabis: 至少就算我们要失败，我们也是以一种全新的、不同于 90 年代前辈的方式失败。 [5:00] Demis Hassabis: 第一步，解决智能问题（即构建 AGI）；第二步，用它来解决其他一切问题。 [15:00] Demis Hassabis: 我们还做不到，而且我也不确定改变天气是不是个好主意，首要任务是先更好地理解它。 [15:00] Demis Hassabis: 正如数学是物理学的描述语言一样，机器学习正是生物学的绝佳描述语言。 [23:45] Demis Hassabis: 我们在以某种方式，阅读宇宙的语言。 [20:00] Demis Hassabis: 第一步：打造极致的工具。集中精力构建一个极其智能、实用且精密的工具，这本身就已经是极其深刻且影响深远的成就；第二步：跨越意识与主体性的‘卢比孔河’。 [21:40] Demis Hassabis: 对于人工系统而言，我们永远无法获得基质上的等价性。

[2:40] Demis Hassabis: you want to be five years ahead of your time, not 50 years ahead. [2:40] Demis Hassabis: when it's obvious to everyone, it's too late. But if you're 50 years ahead, then there's probably no way you can get it to be successful. [4:40] Demis Hassabis: Our original mission statement at DeepMind was step one, solve intelligence, i.e. build AGI. Step two, use it to solve everything else.

[0:00] Dennis, thank you so much. Exciting to be here. Thanks everyone for coming. It's great to be here. We're so honored to have you at our chocolate factory. Yes, I just heard about that. Looking forward to the chocolate afterwards. Excellent. Well, Dennis, we're going to jump right in. We have one of the OGs in every way. Original thinkers, founders, visionaries in all things AI. A true believer, true scientist in Dennis. We're going to spend the beginning of the conversation about the early days, then early days of DeepMind, and then we'll get into the science and open up the room for some questions. So let's jump right in. So Demis, you were a chess prodigy. You also were a founder of a gaming company. You're a neuroscientist, and you're the founder of DeepMind, now leader of a really big consequential company. Those seem like pretty unrelated pieces, but you've said that there's a common thread. Can you walk us through that? There is a common thread and maybe I made it into a common thread So it could be a post hoc you know sort of shaping But I wanted to do AI for a long time So I kind of decided it was the most important thing I could possibly and most interesting thing I could do in my teenage years And then I picked things to study or do that I felt eventually would help me build a company like DeepMind So I had that as a plan from about 15, 16 years old. I had a detour into games because actually that was in the 90s. This is that's where all the most cutting edge technology was being done. Obviously, not just in AI, but in graphics, especially, including hardware. Of course, the GPUs we all use today, they were designed for graphics engines. And we used the, I mean, I was using the very first GPUs back in the day in the late 90s. So there was a lot of really cutting edge technology. And then all the games I made, including all the games I did for Bullfrog, but also my own company, Elixir Studios, all involved AI as the main gameplay component. so probably the most you know well-known game I made was theme park when I was about 17 and that was a simulation of an amusement park and thousands of little people came into your theme park and played on your rides and decided what to buy from the shop so there was a whole kind of economics AI model underneath and it was one of the first games of its type along with SimCity and when I saw it sold you know 10 million plus copies and when I saw how delighted people were to interact with the AI you know that was one of the things that made me think about spending my whole career on it And then of course neuroscience is to get inspiration from how the brain works and different algorithmic ideas from that And then just bringing all those different things together for the start of DeepMind when the timing we felt was right. And then, of course, we use games as the early proving ground for our AI ideas. So we've got a room full of founders here, and you can relate because you're a founder not just once, but twice. Take us back to the first time, Elixir Studios. What was that like? It's not the startup that you're most known for, but it was one that you had incredible success with. How did you lead that and what did it teach you about building? Well, look, I started Elixir Studios straight out of college, and I was lucky enough to work at Bullfrog Productions, which, for those of you who know games, it was a kind of legendary game studio in the early days of the game industry, probably the best one in Europe. And I wanted to do something that combined, pushed AI. So effectively, I was funding AI back in those days through the backdoor, through games development, and then pushed the forefront of that and combine it with cutting edge creativity And I think that still relevant today with the way we do our blue sky research but maybe the biggest lesson i learned was you want to be five years ahead of your time not 50 years ahead so we tried to do a game called republic at elixir studios which simulated a whole country and then the idea of the game is you could sort of overthrow i think there was a dictator in charge of the country in any number of different ways and we basically simulated living, breathing cities. And this is, bearing in mind, this is like the late 90s on a Pentium. So we had to get all the graphics and all the AI for a million people working on, you know, a home PC at the time. So it was a little bit ambitious. And maybe it was too ambitious. And it caused some issues. And I took that lesson with me of like, you want to be ahead of your time, you don't want to be obviously, when it's obvious to everyone, it's too late. But But if you're 50 years ahead, then there's probably no way you can get it to be successful. All right. So speaking of not being too far ahead of your time, it was 2009 and you decided there would be AGI. Yes. Maybe it was only 10 years ahead of our time that time, better than 50 years. So tell us about, again, room full of founders here. Tell us about 09. How did you convince the first few brilliant talents? [5:00] You pulled in really high caliber employees, early team members. How do you convince them to believe in what seemed like total sci-fi at the time? Well, there were some interesting threads that I think we picked up on. I think we thought we were five years ahead, but maybe we were more like 10. But it was, you know, deep learning had just been invented by Jeff Hinton and colleagues sort of in academia, but almost no one had really realized it was a big deal. we knew a lot about reinforcement learning and we felt there was huge progress to be made by combining those two techniques which almost had never been mixed together really certainly not any kind of anything other than toy problems in the academic subjects they were quite two quite siloed parts of AI then we could see the compute the GPUs at the time were going to be really useful of course we use TPUs now but the accelerated computing industry was was was going to be very helpful. And then we also felt at the end of my PhD in postdoc and some of the other people I got together were computational neuroscientists, that we had enough ideas and principles from the brain that could be useful including the idea that reinforcement learning could eventually scale to AGI So we felt we had these ingredients and we almost felt like we were keepers of a secret because no one either in academia or industry really believed that any big progress was possible. In fact, a lot of the people in academia used to literally roll their eyes up at us when we were sort of suggested we would work on AGI or strong AI, it was sometimes called at the time, because it was like, well, we know this doesn't work. So everyone tried it in the 90s and I did my postdoc at MIT which was the which is the sort of center point for expert systems and first of all the logic language systems I mean it seems amazing to think that now but I was already feeling like that was a chaotic then but they they you know that's still how it was done both in Cambridge in the UK and also in MIT these big centers of traditional AI and it felt like but actually that convinced me even more that we were onto something because at least if we were going to fail, we would fail in a different way than people had failed, you know, to get to AGI in the nineties. So that felt like it was worth doing, uh, no matter what, even if obviously it was research, we didn't know for sure it would be successful, but at least we would, we would fail in an original way if it didn work Was there any common sticking point in that early belief Was there something that you had to prove either to yourself or to your early followers to get them on board Well it was I mean we had put it this way I would have been spent my life on AI no matter what happened. So as it's turned out, it's gone, you know, it's sort of gone on the absolutely amazing side of the optimistic side of what we thought still actually within what we were predicting in 2010. We thought it would be a 20 year mission. And I think we're basically exactly on track as a field for that. And obviously we played our part in that. But even if it hadn't transpired that way, and it was still now a niche subject, that's what I would still be doing. Because I felt it was the most important technology ever, if it was obvious to me. Our original mission statement at DeepMind was step one, solve intelligence, i.e. build AGI. Step two, use it to solve everything else. So I always thought it was the most important technology that could ever be invented, but also the most interesting one. So as a tool for science, as an interesting artifact in itself, and actually as one of the best ways to understand our own minds, you know, like the nature of consciousness, dreaming, creativity, all of these questions I had as a neuroscientist, I felt one of the things that was missing was an analysis tool like AI but also a comparison that you could do sort of a controlled experiment study and compare two different systems against each other Let's talk about AI for science. You've been early to that. You've been a believer and you've been really a purist about this. This is the driving mission. What about the way you set up DeepMind and set the culture has positioned it to be on the constant forefront of AI for science? Well, that was the ultimate goal, at least for me, my personal passion. There's my own drive to build AI, which was to advance science and medicine and our understanding of the world. It's my expression of that mission was to sort of do it in a meta way, right? Build the ultimate tool and then come back when that was ready and use it to make breakthroughs in science. Things like AlphaFold that we've done and I think many more things. So we've always had that at the heart of what we've been trying to do at DeepMind. Actually, we've had an AI for science group division led by Pushmeet Kohli that has existed for nearly a decade now. Actually, pretty much the day after we got back from Seoul and the AlphaGo match, which is sort of 10 years to the month now, is when we started, formally started the AI for science efforts because I was waiting for the algorithms to be powerful. [10:00] and the ideas to be general enough. And for me, you know, cracking go was that point that time that we thought, OK, now we're ready to really apply these ideas to important real world problems, starting with these big scientific challenges. So we've always had that in mind as the most beneficial use of AI, like what could be better than using it to, you know, cure diseases and give us healthier lifespans and to help with medicine, followed obviously by other really important areas like material science and the environment and energy and these kinds of topics, which I think AI is also going to have a huge part to play in the next few years. And how does AI break through in biology? You're deeply involved with isomorphic. This is an area of deep passion. You have been a purist on the potential of AI to cure diseases from the very beginning. When do we have the type of moment that we've had in language and coding, but in biology? Yeah, well, I mean, I'd argue we've already had one of those moments with alpha folds. So you know it a 50 year grand challenge protein folding and the 3D structure proteins is incredibly important thing to know about if you want to design medicines or if you want to understand biology Of course it only one part of the drug discovery process It an important part but it only one part So Isomorphic Labs which is our you know latest spin out having a lot of fun running that as well is to build adjacent technologies in more biochemistry and chemistry space that can actually design the compounds automatically to kind of fit and bind to the right part of the protein. So we now know the protein, the shape of the protein, we know that what's on the surface of the protein and what we have to target. But now we've got to build the right compound that, of course, binds strongly to where you want it to bind on the target of interest, but doesn't bind to anything else, ideally, because that would be a toxic side effect. So the dream is to do almost all the exploration, which is 99% of the work and the time in silico, and then save the the wet lab step just for the validation step. Right. So that would be, you know, I think if we can do that and I think we can get there in the next few years, I think we could reduce drug discovery times instead of that for taking like, you know, an average of 10 years down to months, maybe even weeks and perhaps even days one day And and then I think then all disease could be in reach And I think things like personalized medicine will become possible like personalized variations off of base medicines So I think the whole medical area drug discovery areas is going to be revolutionized in the next few years Brilliant. You talked a lot about AI for science. Do you think that at some point AI will create new sciences, a la Industrial Revolution and thermodynamics? Will there be something new, taught fundamentally in our education system? And if so, what would it be like? Well, I think there's several things along those lines that I think is going to happen. So first of all, the understanding and the analysis of AI systems themselves, I think, is going to become a whole science, a kind of engineering science. These are incredibly interesting artifacts that we are building. and they're incredibly complex as well. As complex, eventually they'll be as complex as the human mind and the brain. And so they'll need to be studied so we can understand fully, way beyond where we are today, how these systems work. So I think there's a whole kind of field, Mechinterp is part of that but there a lot more I think that we can do to analyze these systems So that would be a science But I think also AI itself will maybe unlock new sciences which is maybe what you getting at The one I'm particularly excited about is AI for simulations. So I love simulations. All the games I wrote not only had AI, but they were simulations. And I think simulations is the way we can address some of the what we maybe think of social sciences like economics and other more humanistic subjects because it's very difficult to do control studies in that you know why aren't they just sciences like physics today because the problem is they're emergent systems just like biology actually and it's very hard to do repeated controlled experiments you know if you're raising stress rates by half a percent you have to do it in the real world and then see what happens you can have theories, but you can't run it thousands of times. But if you could simulate things really accurately, then maybe there's sort of new sciences to be done where you can sort of rigorously sample from a very accurate simulator. And then I think that will allow us to make much better decisions in these today, very uncertain domains. What will it take to get to those extremely accurate simulations? World models, what kind of science is necessary? [15:00] Yeah, well, look, I mean, I'm thinking a lot about that in, we do a ton of that work, like learning simulators, basically what it would be. So, you know, these are in domains where you can't, we don't know the mathematics of it well enough or it's perhaps too complex. We can't just write, we can't just write a directly down a special case simulator. It's just not accurate enough, doesn't capture all the variables. We're doing that, we've done it with weather. We have the most accurate kind of weather simulator in the world, Weather Next, and it's far faster than what the meteorologists use. Can you change the weather yet? No, we can't. We can't, no. And I'm not sure that would be a good idea. But the first step is to understand it better. But then even biology, you know, we're working on a kind of what I call a virtual cell. So a hugely dynamical emergent system. And I think biology is is perfect. Sort of machine learning is perfect description language for biology in the same way maths is for physics, because I think in biology and in lots of these natural systems, you have loads of weak signals, weak correlations, tons of data, far too much that any human mind can analyze. But there are connections and correlations and interesting causalities within that mass of data So I think it sort of it always struck me that machine learning is the perfect tool to describe those kinds of systems where until today you know mathematics hasn been able to do that either because we can't manage it as top mathematicians because it's too complex or the expressive power of math is not enough to understand these sort of highly emergent dynamical systems. Is it also because of the messiness and stochastic nature? Yeah, sure. And I mean eventually you could – by the way, once you learn these simulators, it may be there's – this is maybe another branch of – new branch of science. You could maybe extract some equations from – once you have the simulator. So you have this sort of implicit simulator or intuitive simulator and then maybe you could extract explicit equations from that because – partly because you could also sample it as many times as you want. As fundamental as Maxwell's or something like that. Maybe. I don't know if that exists for such emergent systems, but if they do exist, I don't see why we won't be able to find them with these methods. That would be amazing. You've talked about this theory that the basic building block of everything in the universe could be information This is more theoretical How do you think about that and what does that mean for a traditional classical Turing computer Well look I think you can of course all the famous, you know, equals mc squared and all the stuff Einstein did and energy and matter are kind of equivalent. But I actually think information has a kind of equivalency in the same way. So you can think of, you know, the organization of matter and structure and especially things like biology that are resisting entropy as basically information processing systems at their heart. So I think one can convert all of those three kind of quantities into each other. But I have this feeling information is most fundamental. So it's a little bit the opposite way around to the classic physicist thought in the 1920s and things where, you know, it's sort of energy and matter primary. I actually think it's a better way to understand the world, the universe, is to think about it as information first. And if that's true, and I think there's quite a lot of evidence for that, then of course, AI is even more sort of profound in a sense than we think. And it's already pretty profound because it's also about organizing information and understanding information and constructing informational objects So AI in my opinion is all about information processing so I think there something sort of very deeply connected with with uh these different areas If you think of it through the lens of information processing as the primary way to think about it And do you think a classical Turing machine will be able to compute everything? Well, I sometimes think, you know, I sometimes sort of think about what we're doing and refer to ourselves as Turing's champion because Turing machines, I think Alan Turing is one of my all time, you know, favorite scientific heroes. I think what he did obviously laid the foundations for the computers for computer science, but also AI. And I think it's one of the most profound results ever is the Turing machine result. You know, everything that is computable can be computed by a relatively simple description of a machine. So I think our brains are likely to be approximate Turing machines. And I think it's interesting to think about the connection between Turing machines and quantum computers and quantum systems. But I think at least what we've shown with things like AlphaGo and especially AlphaFold is that a classical Turing machine, obviously in the guise of a modern neural network, it can model what was thought to be in the case of protein folding, it's a quantum system, you know, at some sense. [20:00] It's dealing with very small particles, and one might think you'd have to take into account all the quantum effects of the water bonds and all sorts of things. But it turns out you can get to an approximate optimal sort of solution on a classical system. So it may turn out that a lot of things that we think that would need a quantum system to model or run might be modelable on a classical system if you thought about in the right way. So you've talked about AI consistently as a tool, like a telescope or a microscope, astrolabe through the centuries. But when you think about a machine that can model almost anything, let's say it can even model quantum systems, like you pointed out, when does it stop becoming a tool? And will that ever happen? Well, my strong feeling is we should, in this sort of mission and journey to build AGI, those of us on that journey, many people in this room, I feel like it would be best to build a tool first an incredibly intelligent and useful and precise tool and then cross the next sort of Rubicon That already profound enough And of you know of course the tool could start becoming more and more autonomous and agent like that we all seeing We're in the midst of that, the agent era now. But then there's a further step of like, you know, does it have agency? Is it conscious? These sorts of questions, which are also going to be questions we're going to need to address. But I would recommend we do that as a second step, perhaps using the tool in the first step to help us with those next profound questions. And ideally, also, we could understand our own brain and minds better and define things like consciousness a lot more precisely than we can today. Do you have estimations of what that definition of consciousness might look like? No, I mean, I haven't got much to add beyond that thousands of years of philosophy hasn't said already. But I mean, it's very clear to me that it's obvious some components are going to be needed. They're probably necessary, but not sufficient. Things like self-awareness and, you know, the idea of self and other, some kind of continuity over time. So some of these things are clearly needed for anything that might look like consciousness. But I mean, obviously it's an open question as to what the full definition is And I talked to many of the great philosophers about that Daniel Dennett obviously sadly passed away recently but we had a long conversation a few years back about this And I think, you know, one of the issues is how does a system behave? Does it behave like a conscious system? So that's, you know, you could argue some of the AI systems might end up being able to do that as they get close to AGI. But then there's still the question of why do you, you know, why do we think each other are conscious? one is the way we're behaving we're behaving like conscious beings but the other thing is we're running on the same substrate so i think if both those things are true then it's parsimonious to imagine you're you're experiencing the same thing i'm experiencing which is why we don't have that debate uh about you know normally about are each other conscious but i think we'll obviously we'll never have the substrate equivalence with an artificial system um so they'll always be a i think it will be hard to completely close that gap. So you can look at it behaviorally, but what about experientially? There are probably some ways to do that post-AGI, but maybe it's a bit about out of scope today, even for AI for science discussion. Brilliant. So we're going to open the room to questions in just a moment get your questions ready But you brought up philosophers You mentioned Kant and Spinoza as two of your favorite philosophers Kant is this deontological highly duty philosopher Spinoza almost has this deterministic view of the universe. How do you kind of connect those two beliefs, and where is your thinking of how the world works? Well, the reason I like those two, they stuck out for me, is that I think Kant, when I was doing my PhD in neuroscience, you know, his sort of statements about the mind creates reality, right? I think that's basically true. And so another reason to study the mind, right, and how the brain works. And I'm interested ultimately in the nature of reality. So we have to understand how the mind is interpreting that. And so I think that's for me what I took from Kant. And then Spinoza, it's more about the you could almost call spiritual dimension of like, well, if you're trying to understand the universe using science, in my case, as the tool, you're sort of understanding some deep mystery about how the universe works. in a really kind of deep way. And that's what I feel we're doing and I'm doing when I do my science and we work on AI and we're building these tools is somehow we're kind of reading [25:00] language of the universe. Beautiful. What a beautiful way to say what you do every day. Demis, scientist, orator, and philosopher. We will, before we wrap, do a couple of rapid fire questions. Thank you for finishing. He's not seeing these yet. Over, under, on distribution, year of AGI. Oh, wow. Or reject premise of question. No, 2030. I've been pretty consistent about that. Okay, 2030. Must read book, poem, or paper for when we achieve AGI? Oh, wow. For when we achieve, once we achieve it. Well, my favorite book is The Fabric of Reality by David Deutsch. So I think that still holds. I'd hope to answer the questions in that book with the AGI. That's my post-AGI work. Brilliant. Proudest moments so far in DeepMind? uh oh wow um we've been lucky to have a lot i mean probably alpha fold okay yeah now a couple games questions if you were engaged in a high stake strategy game turn-based strategy game okay uh civ polytopia serious games uh and you could select one scientist from history we're thinking the the einsteins the turing's newton's who would you select to be on your team On my team? On your team. Oh, gosh. Probably von Neumann, I think. Great choice. Yeah, you want a game theorist, I think. And I think he's the best. Makes sense. I feel like he'd be a good teammate. All right. Well, Demis, you do it all. Thank you so much for being with us. Please join me in taking a time. Thanks, everyone. Thank you.