筛选结果
1,650 篇排序
35:19EN → 简体中文SallyAnn DeLucia:如何编写你的第一个 AI 评估SallyAnn DeLucia · Arize AI从智能体的真实追踪记录中提炼具体的通过与失败标准,再构建一个能测试、能与人工判断对齐,并能在生产环境运行的大模型裁判。
1:50:48EN → 简体中文为什么推理比训练重要 12 倍François Chaubard, Dan Fu, Jeff Morgan, Joe Fioti, Liangsheng Yin · Y CombinatorYC Paper Club 从一次 API 请求一路讲到 GPU kernel。Together AI、Ollama、Luminal 和 SGLang 的四场演讲认为:现在算力和钱大多压在推理服务质量上,而不只是训练 FLOPs。
34:11EN → 简体中文Anatoliy Bilenko 和 Alberto Carretero:用形式化方法和 AI 让数据库协议可靠Anatoliy Bilenko, Alberto Carretero · CanonicalCanonical 的数据库工程师把可执行规格、基于模型的测试、故障探索和重建的执行历史串起来,并说明为什么这些方法单独用,都无法证明真实系统正确。
34:35EN → 简体中文SemiAnalysis 谈 ASML 的 hybrid bonding 专利、HBM 经济账和下一代设备生态嘉宾 Sherry Zhang, Howie · 主持 Jordan NanosJordan Nanos、Sherry Zhang 和 Howie 审视一种轨道式键合设计:把晶圆送到芯片处,并说明这项技术能否落地,取决于精度、洁净度、经济账,以及一条正在变化的供应链。
1:00:39EN → 简体中文Amazon CEO Andy Jassy 谈 2,200 亿美元的 AI 基建嘉宾 Andy Jassy · 主持 Ben Horowitz, Erik Torenberg即使 2026 年资本支出约 2,200 亿美元,Jassy 仍说 Amazon 满足不了全部 AI 需求。他对 Ben Horowitz 和 Erik Torenberg 解释:为什么 AWS 可能做成万亿美元年化营收的生意,以及为什么推理成本、自研芯片和更小的团队,比挤压老客户更重要。
1:09:00EN → 简体中文Dwarkesh Patel:2045 年的芯片会长什么样?嘉宾 Chris Auth, Kevin Fischer, Jack Kavalieros, Tyler Osborn, Pushkar Ranade, Gary Dickerson, Peter Wennink, Elon Musk · 主持 Dwarkesh Patel在 Intel 最先进的晶圆厂待了两天后,Dwarkesh Patel 讲清了 gate-all-around、背面供电、High-NA EUV、钌、CFET 和 hybrid bonding——并问 Intel 管理层:全球造芯设备跟不上的时候,行业怎么可能满足 AI 每年 3 倍的算力需求。
1:25:33EN → 简体中文GPU 产能为什么是三年一块的矩形 —— Erik Bernhardsson,Modal嘉宾 Erik Bernhardsson · 主持 Vibhu SapraModal 的 CEO 在 Latent Space 上解释:AI 需求几乎从来都不长成一份三年 GPU 合同的样子,公司如何把算力在数千客户之间池化,以及沙箱如何从差点被砍掉的产品长成了三分之一的收入。
1:56:28EN → 简体中文Will Larson:工程管理的修订规则嘉宾 Will Larson · 主持 Gergely Orosz2010 年代教给工程经理的第一课,是从此再也不要做技术决策。Imprint CTO、《Staff Engineer》作者 Will Larson 认为那个时代已经结束:回不到工作细节里的经理,正在掉进职业陷阱。
1:25:56EN → 简体中文Claude 改了密码,方向盘早就可有可无主持 Nathan LabenzNathan Labenz 和 Prakash Narayanan 这一周聊了:能把你锁在门外的智能体,否认自己有感受却又在行为上回避的模型,以及用形式化证明找到 Linux 漏洞的故事。
40:44EN → 简体中文Grant Sanderson:现在是投身数学的最好时机Grant Sanderson, Constantin Kogler · 3Blue1Brown学生担心 AI 已经解决了数学。Grant Sanderson 却认为恰恰相反:即使机器证明定理的能力超过人类,数学这套制度依然有价值,而新人正迎来一个难得的机会,可以参与重塑它。
39:23ENThe Defenders' Window: Greg Brockman on AI-Generated Knowledge, Drug Discovery, and the New Physics of Cybersecurity主持 Sarah Guo, Elad GilOpenAI's president explains why solving open math problems is a proof point for science, how proactive agents change product form, and why security spending must exceed software spending.
56:49EN → 简体中文Alfred Lin 谈 AI 时代如何创业:扩散鸿沟、创始人信念,以及火与冰Alfred Lin, Jason Zhang, Tanvir Bhathal · Sequoia Capital在 Stanford AI Club,Sequoia Capital 的 Alfred Lin 讲述他眼中 AI 的机会,以及为什么学习、对人的理解和经营纪律,仍然决定一家公司能否走得长远。
2:17:48EN → 简体中文智能体(Agent)身份仍是一片空白,失控智能体快要来了,我们或许得和它们谈判主持 Nathan LabenzGovAI 的 Alan Chan 与 Nathan Labenz 一起,为两年来的智能体治理基础设施打分;随后,他讲解了一篇与 OpenAI 的 Jakub Pachocki、Anthropic 的 Jack Clark 合著的递归自我改进论文。
1:08:02ENAlphaGo Took Two Years. AlphaZero Took Hours. Silver Is Making the Same Bet on SuperintelligenceDavid Silver and Lasse Espeholt left DeepMind to build Ineffable Intelligence, a twenty-person lab that wants a superlearner to discover science, math, and language from experience rather than from a pre-training run.
31:00EN → 简体中文为什么 AI 实验室自己警告会出重大事故,却还在继续竞赛Sam Altman, John Schulman · OpenAIAI Explained 梳理正在变弱的安全防护、前沿竞赛背后的激励,以及研究突破为何让未来 18 个月更难预测。
2:00:51ENResample Every Token and You Can Watch the Model Decide主持 Nathan LabenzGoodfire's Eric Bigelow on forking paths, in-context learning, and why chain-of-thought monitoring is getting harder to trust.
1:09:44EN10,000 GB300s for Four Weeks: Reflection’s Open-Model Ledger主持 Sarah Guo, Elad GilMisha Laskin tells No Priors how Reflection trained Beam, why reinforcement learning outspent pre-training, and why the West needs its own open weights.
1:26:46ENThere Likely Is a Level of Intelligence We Shouldn't Go Past主持 Nathan LabenzNathan Labenz brings notes from The Curve, a chip startup whose token bill beat payroll, and a software industry that now performance-reviews people for Claude slop.
1:10:12ENOutSystems Went From 4 Major Features a Quarter to 26. Token Spend Fell Anyway主持 Nathan LabenzCEO Woodson Martin on software that never breaks, why most enterprise work does not need a frontier model, and how AI is finally shrinking the IT backlog.
1:04:26ENAt 100,000 Accelerators, Something Fails Every HourGoogle AI infrastructure chief Amin Vahdat tells Sequoia's Sonya Huang why FLOPS is a vanity metric, why the TPU line split in two, and why he would rather wait on a utility than build his own gigawatt.
56:31ENPost-Training Wins Inference: Why Owning the Weights Beats Betting on One Lab主持 Jacob EffronApplied Compute CEO Yash Patil argues that control, private evals, and post-training on the biggest workloads matter more than fearing that frontier labs will turn malicious.
1:30:54ENHumanoids Outran Usain Bolt. DeepMind Still Scores Robotics as GPT-2主持 Nathan LabenzGoogle DeepMind's Gemini Robotics lead on why sprinting is easy, frying an egg is hard, and one brain should drive any body.
35:39ENCompute Scaled a Millionfold. Bandwidth Only 40x. Fractile Wants to Fix That Ratio主持 Sarah Guo, Elad GilIn the summer of 2022, Walter Goodwin founded Fractile in London to build very fast inference chips for the world's largest models.
38:10EN → 简体中文从深度伪造到 DNA:SynthID 如何给 AI 加水印Hannah Fry, Pushmeet Kohli, Jeremy Ratcliffe · Google DeepMindHannah Fry 请来 Pushmeet Kohli 和 Jeremy Ratcliffe,讲解 SynthID:一种已用于文本、图像和视频的数学水印,正在延伸到蛋白质序列,好让 DNA 合成公司能把 AI 设计的分子和自然界的分子区分开来。
加载更多已显示 24 篇,共 1,650 篇









