人物访谈 on Peng Tan's AI Bloghttps://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/与AI领域的专家学者和从业者的深度对话在人工智能时代保持人性：Michael Edmondson 博士的八大人类核心特质框架https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/staying_human_ai_edmondson/Wed, 26 Aug 2026 15:59:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/staying_human_ai_edmondson/基于 Michael Edmondson 博士著作《在人工智能时代保持人性》的解读：在 AI 普及的时代，个人与组织如何通过强化人类核心特质来应对技术变革。Karpathy （Gist）: LLM Wiki实操理念说明https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/llm_wiki_gist/Sat, 04 Apr 2026 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/llm_wiki_gist/Karpathy基于大语言模型的个人知识库的新的方法的实操理念说明，我搬运过来，翻译了一下。Karpathy:基于大语言模型的个人知识库https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/llm_knowledge_base/Fri, 03 Apr 2026 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/llm_knowledge_base/Karpathy基于大语言模型的个人知识库的新的方法，我搬运过来，翻译了一下。Anthropic CEO Dario Amodei博客-《技术的青春期》https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/anthropic-ceo-dario-adolescence-of-technology/Tue, 10 Feb 2026 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/anthropic-ceo-dario-adolescence-of-technology/Dario Amodei（The Adolescence of Technology），系统阐述强大AI对人类社会的深远影响，既肯定其变革潜力，也重点警示未妥善管控的风险，引发全球广泛关注。Dan Koe: 如何在一天内彻底改变你的人生https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/dankoe_changelifeinoneday/Sat, 07 Feb 2026 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/dankoe_changelifeinoneday/Dan Koe阐述了通过重塑身份来实现人生转变的核心逻辑，强调真正的改变并非源于行为的被迫调整，而是源于内在自我的根本重新定义。Karpathy: 2025 年大语言模型年度回顾https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/andrej_karpathy_2025_review/Sat, 20 Dec 2025 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/andrej_karpathy_2025_review/Karpathy大神在2025年年底的时候多大模型这一年的发展做了一个年度总结,我搬运过来，翻译了一下。Sam Altman: 10 Yearshttps://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/samaltman/ten_years/Thu, 11 Dec 2025 11:11:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/samaltman/ten_years/Sam Altman关于 OpenAI 十周年回顾的全文翻译。在这篇里程碑式的回顾中，Sam Altman 复盘了 OpenAI 从 15 人的“理想主义实验室”进化为全球 AGI 领航者的非线性历程，重点强调了 2017 年在 RLHF（基于人类反馈的强化学习）与无监督情感神经元上的技术奇点，以及“迭代式部署”策略对行业标准的重塑，并最终抛出了一个极具技术压迫感的预测：在未来十年内（2035 年前），人类几乎确定将构建出超级智能（Superintelligence）。沈向洋https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/shen_xiangyang/Fri, 28 Nov 2025 20:00:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/shen_xiangyang/沈向洋的观点搜集Block CTO Dhanji R. Prasannahttps://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/block-cto_dhanji/Tue, 25 Nov 2025 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/block-cto_dhanji/Block CTO Dhanji R. R. Prasanna在各个场合的谈话整理。芮勇博士介绍https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yongrui_lenovo/Thu, 20 Nov 2025 16:10:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yongrui_lenovo/芮勇博士介绍深入理解AI Agent：芮勇博士的九个核心问题解析https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yongrui_agent_nine_question/Thu, 20 Nov 2025 16:10:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yongrui_agent_nine_question/联想芮勇博士对AI Agent的未来提出了九个深刻而关键的问题Jason Wei在各个场合的观点https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/jason_wei/Sun, 19 Oct 2025 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/jason_wei/Jason wei在各个场合的观点Pavel Durov 的成功密码：深度解析自由、自律与商业哲学的融合https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/telegram_ceo_/Wed, 15 Oct 2025 16:10:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/telegram_ceo_/Pavel Durov 的成功密码：深度解析自由、自律与商业哲学的融合Groq创始人Jonathan Ross访谈核心洞察：计算力、能源与AI的未来https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/groq_founder_jonathan-ross/Tue, 14 Oct 2025 23:01:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/groq_founder_jonathan-ross/Groq创始人Jonathan Ross访谈核心洞察：计算力、能源与AI的未来强化学习的奠基人的惊人警告：为什么说LLM可能是一条死胡同？https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/richard_sutton/Thu, 02 Oct 2025 09:10:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/richard_sutton/强化学习的奠基人惊人警告：为什么说LLM可能是一条死胡同？AI的最终审判：为什么资本主义、GDP和人类智力都将在1000天内失效？https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/emadmostaque/Fri, 19 Sep 2025 22:10:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/emadmostaque/AI的最终审判：为什么资本主义、GDP和人类智力都将在1000天内失效？Emad Mostaque在“Tom Bilyeu”YouTube频道上的访谈，主要探讨了人工智能（AI）将如何从根本上颠覆全球经济。Mostaque认为，在不久的将来，AI不仅会取代大量人类工作，还会使资本主义和国内生产总值（GDP）等传统经济结构变得过时。他提出了一种新的**“心智”（M.I.N.D.）框架来衡量经济健康度，该框架关注物质资本、智能、网络效应和多样性，而不仅仅是GDP。他警告称，人类认知劳动的价值将变为负数，并提出需要重新定义货币和财富的流动方式**，例如通过新型数字资产，以应对AI驱动的生产力爆炸所带来的社会和经济动荡，并确保人类生存与发展。Ray Dalio在各个场合的观点https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/raydalio/Sun, 14 Sep 2025 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/raydalio/Ray Dalio:在各个场合的观点Anthropic CEO Dario Amodei访谈https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/anthropic_ceo_dario/Mon, 04 Aug 2025 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/anthropic_ceo_dario/Dario Amodei访谈,按“技术 → 市场 → 组织 → 风险 → 行动清单”的逻辑铺开。AI从业者必读Geoffrey Hinton：2025 WAIC主题演讲《数字智能是否会取代生物智能》https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/geoffrey_hinton/Tue, 29 Jul 2025 09:10:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/geoffrey_hinton/Geoffrey Hinton在2025年世界人工智能大会上关于数字智能与生物智能对比的核心观点梳理李飞飞博士的生平与洞见https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/feifeilee/Wed, 02 Jul 2025 22:03:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/feifeilee/李飞飞博士的生平与洞见Andrej Karpathy在各个场合的观点https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/andrej_karpathy/Thu, 19 Jun 2025 22:30:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrej_karpathy/andrej_karpathy/Andrej Karpathy深入解析了软件从1.0到3.0的范式演进，将LLMs类比为公用事业、晶圆厂和操作系统，并探讨了在当前局限下构建'钢铁侠战衣'式部分自主AI应用的策略Ilya Sutskever在各个场合的观点https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/ilyasutskever/Wed, 18 Jun 2025 23:03:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/ilyasutskever/Ilya Sutskever在各个场合的观点Jim Fan在各个场合的观点https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/jimfan/Thu, 08 May 2025 23:03:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/jimfan/Jim Fan在各个场合的观点Shunyu Yao: The Second Half (下半场)https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/shunyu_yao/Fri, 11 Apr 2025 23:03:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/shunyu_yao/AI的下半场samaltmanhttps://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/samaltman/samaltman/Mon, 01 Jan 2024 00:00:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/samaltman/samaltman/samaltman相关资源链接Stephen Wolfram：计算宇宙的拓荒者https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/stephen_wolfram/Mon, 01 Jan 2024 00:00:00 +0800https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/stephen_wolfram/Stephen Wolfram的生平、贡献和思想，包括计算不可约性理论和Mathematica等工具<link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrawwu/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/andrawwu/</guid><description><p><a href="https://www.andrewng.org/">https://www.andrewng.org/</a></p></description></item><item><title/><link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/deepmind_ceo_demis-hassabis/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/deepmind_ceo_demis-hassabis/</guid><description><p><a href="https://www.youtube.com/watch?v=1XF-NG_35NE">https://www.youtube.com/watch?v=1XF-NG_35NE</a> ==&gt; <a href="https://gemini.google.com/u/0/app/969d731e099abefe?ref=glasp">https://gemini.google.com/u/0/app/969d731e099abefe?ref=glasp</a></p></description></item><item><title/><link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/elonmusk/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/elonmusk/</guid><description/></item><item><title/><link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/harrison_chase/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/harrison_chase/</guid><description><h1 id="harrison-chase">Harrison Chase</h1> <h2 id="1-个人简介">1. 个人简介</h2> <h2 id="2-个人经历">2. 个人经历</h2> <p>Ambient agents listen to an event stream and act on it accordingly, potentially acting on multiple events at a time 环境代理会监听事件流，并据此采取相应的行动，且有可能同时对多个事件进行处理。</p></description></item><item><title/><link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/linyuanqing/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/linyuanqing/</guid><description><p>林元庆作为Aibee创始人及AI领域资深专家，近年来在多场合阐述了对生成式AI及物理世界智能化的深刻见解，其观点可概括为以下核心脉络：</p> <hr> <h3 id="一生成式ai的定位物理世界大模型的基石">一、生成式AI的定位：物理世界大模型的基石</h3> <ol> <li> <p><strong>从感知到理解的技术跃迁</strong><br> 林元庆认为，生成式AI不应局限于文本或图像的表面创作，而需与物理世界的深度理解结合。他提出“物理世界AI大模型”概念，强调其核心是<strong>像素级的3D语义理解能力</strong>，能够通过海量视频数据自学习物理规律（如行人过马路时的行为预判），而非依赖人工标注的监督学习。这一过程需要生成式AI具备动态场景重建与多模态推理能力。</p></description></item><item><title/><link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/luqi/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/luqi/</guid><description><h1 id="当前大模型的春秋战国时代">当前大模型的春秋战国时代</h1> <h1 id="当前scaling-law的情况">当前scaling law的情况</h1> <h2 id="当前scaling-law的进展">当前scaling law的进展</h2> <h2 id="当前scaling-law的进展-1">当前scaling law的进展</h2></description></item><item><title/><link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yann_lecun/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yann_lecun/</guid><description><h2 id="lecun-的成就">LeCun 的成就</h2> <p>Yann LeCun是人工智能领域的杰出学者，被誉为“卷积神经网络之父”，与Geoffrey Hinton和Yoshua Bengio并称为“深度学习三巨头”，共同获得了2018年图灵奖。他在学术上取得了诸多重大成就，具体如下：</p></description></item><item><title/><link>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yaoshunyu/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://77f3dc70.hobbytp-github-io.pages.dev/zh/celebrity_insights/yaoshunyu/</guid><description><p>姚顺雨最新动态</p> <p>姚顺雨（Shunyu Yao），27岁，清华大学姚班（计算机科学精英班）毕业生、普林斯顿大学计算机科学博士，曾是OpenAI智能体产品（如Operator和Deep Research）的核心贡献者。他于2024年8月加入OpenAI，主要研究语言智能体、强化学习（RL）和代理（Agent）技术。</p></description></item></channel></rss>