This great conversation with @sk7037 of @OpenAI is also available on Spotify, Apple Podcasts and here on YouTube (note: we had sound qualities in this episode, apologies, but the conversation is well worth a listen)
这场与 @OpenAI 的 @sk7037 的精彩对谈也可在 Spotify、Apple Podcasts，以及这里的 YouTube 上收听/观看（注：这一期的音质有些问题，抱歉，但这场对话仍然非常值得一听）
"We can't build fast enough": my conversation with Sachin Katti (@sk7037), Head of Industrial Compute at @OpenAI about Stargate, Jalapeno, data center financing, and all things compute. 00:00 — Cold open: “One of the largest things humanity has ever built” 00:30 — Welcome: Sachin Katti, Head of Industrial Compute at OpenAI 01:44 — Is this the biggest infrastructure buildout in history? 03:41 — Why OpenAI is building a new industrial muscle 04:54 — What an AI data center actually is 05:27 — “Factories turning electrons into tokens” 06:35 — Why AI data centers need liquid cooling everywhere 08:10 — The power problem: grids, generation, transmission, substations 10:43 — Behind-the-meter power and gas turbines 11:02 — Why nuclear “can’t come soon enough” 11:49 — Jalapeño: why OpenAI is designing its own AI chips 13:19 — Tokens per watt: the new metric that matters 13:38 — Why inference may now dominate AI compute 14:58 — Is OpenAI overbuilding compute? 16:47 — Why OpenAI thinks the bigger risk is not building fast enough 17:55 — Communities, jobs, water, and the local data-center debate 21:16 — How OpenAI chooses data-center sites 22:25 — What “industrial compute” means inside OpenAI 25:59 — Sachin’s path: Stanford, startups, Intel, OpenAI 28:05 — OpenAI’s compute portfolio: Microsoft, hyperscalers, neoclouds 29:37 — Stargate explained 31:21 — Abilene, Oracle, and the next wave of AI data centers 32:48 — How massive AI compute gets financed 34:05 — How OpenAI designed Jalapeño so quickly 35:59 — AI is starting to help design AI chips 36:20 — MRC: the networking problem behind 100,000 GPUs 38:47 — Bottlenecks: transformers, turbines, electricians, supply chains 40:29 — Guaranteed capacity: intelligence as a supply unit 42:08 — Will AI data centers move to space?
“We can't build fast enough”：我与 OpenAI Industrial Compute 负责人 Sachin Katti（@sk7037）的对谈，内容涵盖 Stargate、Jalapeno、data center（数据中心）融资，以及所有与 compute（算力）相关的话题。00:00 — 开场：“人类有史以来建造过的最大规模事物之一” 00:30 — 欢迎：OpenAI Industrial Compute 负责人 Sachin Katti 01:44 — 这是历史上最大的基础设施建设吗？ 03:41 — 为什么 OpenAI 正在打造新的工业化能力 04:54 — AI data center（AI 数据中心）到底是什么 05:27 — “把电子转化为 token（词元）的工厂” 06:35 — 为什么 AI 数据中心处处都需要液冷 08:10 — 电力问题：电网、发电、输电、变电站 10:43 — 表后电力与燃气轮机 11:02 — 为什么核能“越早到来越好” 11:49 — Jalapeño：为什么 OpenAI 正在设计自己的 AI 芯片 13:19 — 每瓦 token 数：新的关键指标 13:38 — 为什么 inference（推理）现在可能开始主导 AI 算力 14:58 — OpenAI 是否在过度建设算力？ 16:47 — 为什么 OpenAI 认为，更大的风险是建设速度不够快 17:55 — 社区、就业、用水，以及本地围绕数据中心的争论 21:16 — OpenAI 如何选择数据中心选址 22:25 — “industrial compute” 在 OpenAI 内部意味着什么 25:59 — Sachin 的经历：Stanford、startups、Intel、OpenAI 28:05 — OpenAI 的算力组合：Microsoft、hyperscalers、neoclouds 29:37 — Stargate 解析 31:21 — Abilene、Oracle，以及下一波 AI 数据中心 32:48 — 超大规模 AI 算力如何获得融资 34:05 — OpenAI 如何如此快速地设计出 Jalapeño 35:59 — AI 已开始帮助设计 AI 芯片 36:20 — MRC：10 万张 GPU 背后的网络难题 38:47 — 瓶颈：transformers、turbines、电工、供应链 40:29 — 保证容量：把 intelligence（智能）作为一种供给单位 42:08 — AI 数据中心会迁移到太空吗？