← 首页
01SEP
AI前沿每日脉动
AI Frontier Pulse · 中英双语版 Bilingual Edition
2026.09.01 · 周二刊
10 位 Builder19 条推文1 期播客0 篇博客
ChatGPT WorkApplied AIOpen WeightsTokens per WattTrust & Access
Richard Liu · 2026
Curated by Richard Liu · Snapshot follow-builders-2026-09-01-v1
今日头条 · Model And Harness Co-design01 / 15
今天最强信号：GPT-5.6 Sol 只用两项运行设置变化，就在 ARC-AGI-3 达到 SoTA。模型能力不再能脱离 harness、compaction 与跨 context 执行单独评估。The strongest signal: two runtime setting changes put GPT-5.6 Sol at the top of ARC-AGI-3. Model capability can no longer be separated from harness design, compaction, and cross-context execution.
@thsottiauxX 原文2.3K ❤ · 42 RT · 1.8K 💬
Thibault Sottiaux：If you still haven't tried Codex but have considered it in工作产品化 · Source Excerpt
Thibault Sottiaux 原文要点：If you still haven't tried Codex but have considered it in the past, what's the one thing holding you back?
Source-grounded excerpt from Thibault Sottiaux: If you still haven't tried Codex but have considered it in the past, what's the one thing holding you back?
@thsottiauxX 原文2.3K ❤ · 42 RT · 1.8K 💬
Nan Yu：Some personal news—I’m joining OpenAI to work on Codex and工作产品化 · Source Excerpt
Nan Yu 原文要点：Some personal news—I’m joining OpenAI to work on Codex and ChatGPT.
Source-grounded excerpt from Nan Yu: Some personal news—I’m joining OpenAI to work on Codex and ChatGPT.
@thenanyuX 原文3.1K ❤ · 37 RT · 249 💬
02 / 15
行业分析 · Science, Evals, And Credibility02 / 15
Guillermo Rauch：Your next design system is… Markdown.Builder 动态 · Source Excerpt
Guillermo Rauch 原文要点：Your next design system is… Markdown.
Source-grounded excerpt from Guillermo Rauch: Your next design system is… Markdown.
@rauchgX 原文2.4K ❤ · 103 RT · 72 💬
Garry Tan：Circleback is so much better than Granola, it's not even cBuilder 动态 · Source Excerpt
Garry Tan 原文要点：Circleback is so much better than Granola, it's not even close Granola still doesn't even support multiple person disambiguation
Source-grounded excerpt from Garry Tan: Circleback is so much better than Granola, it's not even close Granola still doesn't even support multiple person disambiguation
@garrytanX 原文603 ❤ · 12 RT · 62 💬
03 / 15
产品设计 · Experience Becomes A Force Multiplier03 / 15
AI 降低执行成本，却没有消除品味与经验：资深创始人的判断继续复利，设计工具也必须把审美、易用性和乐趣一起产品化。AI lowers execution cost without eliminating taste or experience. Seasoned judgment compounds, while design tools must productize aesthetics, ease, and fun together.
@garrytanX 原文403 ❤ · 33 RT · 72 💬
Garry Tan：I just made some new GBrain evals that help prove that my Builder 动态 · Source Excerpt
Garry Tan 原文要点：I just made some new GBrain evals that help prove that my retrieval-for-AI-agent open source layer is SOTA for reading memory back without LLM-in-loop And I've also added evals…
Source-grounded excerpt from Garry Tan: I just made some new GBrain evals that help prove that my retrieval-for-AI-agent open source layer is SOTA for reading memory back without LLM-in-loop And I've also added evals…
@garrytanX 原文403 ❤ · 33 RT · 72 💬
Guillermo Rauch：Coding tokens are basically infrastructure.低延迟模型 · Source Excerpt
Guillermo Rauch 原文要点：Coding tokens are basically infrastructure.
Source-grounded excerpt from Guillermo Rauch: Coding tokens are basically infrastructure.
@rauchgX 原文250 ❤ · 7 RT · 34 💬
04 / 15
设计系统 · Harness, Design.md, Routing, Mobile04 / 15
Madhu Guru：If you’re a PM, there is immense alpha in understanding thBuilder 动态 · Source Excerpt
Madhu Guru 原文要点：If you’re a PM, there is immense alpha in understanding the model frontier for your specific product and use cases.
Source-grounded excerpt from Madhu Guru: If you’re a PM, there is immense alpha in understanding the model frontier for your specific product and use cases.
@realmadhuguruX 原文140 ❤ · 10 RT · 8 💬
Aaron Levie：Now that the base open weights AI models are getting far bBuilder 动态 · Source Excerpt
Aaron Levie 原文要点：Now that the base open weights AI models are getting far better, and post training infra is becoming more mature and commercialized, there are going to be all new plays for comp…
Source-grounded excerpt from Aaron Levie: Now that the base open weights AI models are getting far better, and post training infra is becoming more mature and commercialized, there are going to be all new plays for comp…
@levieX 原文119 ❤ · 13 RT · 17 💬
Dan Shipper：Anthropomorphization of AI is good when it serves to help Builder 动态 · Source Excerpt
Dan Shipper 原文要点：Anthropomorphization of AI is good when it serves to help us use (understand, predict, etc) AI better.
Source-grounded excerpt from Dan Shipper: Anthropomorphization of AI is good when it serves to help us use (understand, predict, etc) AI better.
@danshipperX 原文87 ❤ · 10 RT · 25 💬
Aaron Levie：As AI security events pick up, it’s going to be critical tAgent 安全 · Source Excerpt
Aaron Levie 原文要点：As AI security events pick up, it’s going to be critical that we have the most sophisticated AI agents for being able to detect and prevent security issues.
Source-grounded excerpt from Aaron Levie: As AI security events pick up, it’s going to be critical that we have the most sophisticated AI agents for being able to detect and prevent security issues.
@levieX 原文77 ❤ · 13 RT · 16 💬
05 / 15
深度主题 · Product And Engineering Signals05 / 15
没有博客并不等于没有深度：今天的深度来自工程、安全、组织与商业化信号在同一天汇合。No blog does not mean no depth: today's depth comes from engineering, safety, organization, and commercialization signals converging.
follow-buildersNo blog post in this snapshot
今日无新增博客：深度信号来自产品与工程现场No Blog Today, But Strong Product Signals
今天中央 feed 没有新增博客，但高互动推文已经形成清晰主线：产品入口、工程基础设施、模型治理和组织能力正在同时变化。
The central feed has no new blog post today, but the top tweets still form a clear arc across product surfaces, engineering infrastructure, model governance, and organization design.
follow-buildersSnapshot follow-builders-2026-09-01-v1
06 / 15
播客深度 · Compute Meets The Physical World06 / 15
PODCAST DEEP DIVE
AI 的下一轮竞争不只发生在模型层，也发生在冷却、电网、数据中心、供应链与每瓦 token 数构成的物理系统里。The next AI race is not only at the model layer, but across cooling, grids, data centers, supply chains, and tokens per watt.
Training DataYouTube 完整播客
Rich Sutton and Khurram Javed: Why AI Models Stop Learning, and How to Start It AgainPodcast Deep Dive
今天的播客把产品、基础设施和商业化放在同一条线上：AI 不只是模型能力，也是速度、成本和交付系统。
Today's podcast connects product, infrastructure, and commercialization: AI is not only model ability, but speed, cost, and delivery systems.
Training DataYouTube 完整播客2026-08-18
07 / 15
播客理念 · Industrial Compute At Scale07 / 15
算力上线即被消耗New Compute Is Consumed Immediately
新增容量上线即被工作负载吸收，过早放慢建设通常会带来负面意外。
New capacity is consumed immediately, and slowing the build too early repeatedly creates negative surprises.
ComputeYouTube 完整播客
关键指标：每瓦 token 数The Metric Is Tokens Per Watt
定制芯片与模型工作负载协同设计，目标是在电力约束下最大化每瓦可生成的 token。
Custom silicon and model workloads are co-designed to maximize tokens generated per watt.
EfficiencyYouTube 完整播客
数据中心必须新增电力Data Centers Must Add Power
大规模 AI 基础设施需要同步投资发电、输电、变电站与配套容量。
Large-scale AI infrastructure requires new generation, transmission, substations, and supporting capacity.
GridYouTube 完整播客
算力策略是 All of the AboveThe Strategy Is All Of The Above
多种云、芯片、数据中心与供应路径并行推进，减少单点约束。
Clouds, chips, data centers, and supply paths advance in parallel to reduce single-point constraints.
PortfolioYouTube 完整播客
08 / 15
快讯速览 · Builder Signals08 / 15
Madhu Guru：If you’re a PM, there is immense alpha in understanding thBuilder 动态 · Source Excerpt
Madhu Guru 原文要点：If you’re a PM, there is immense alpha in understanding the model frontier for your specific product and use cases.
Source-grounded excerpt from Madhu Guru: If you’re a PM, there is immense alpha in understanding the model frontier for your specific product and use cases.
@realmadhuguruX 原文140 ❤
Aaron Levie：Now that the base open weights AI models are getting far bBuilder 动态 · Source Excerpt
Aaron Levie 原文要点：Now that the base open weights AI models are getting far better, and post training infra is becoming more mature and commercialized, there are going to be all new plays for comp…
Source-grounded excerpt from Aaron Levie: Now that the base open weights AI models are getting far better, and post training infra is becoming more mature and commercialized, there are going to be all new plays for comp…
@levieX 原文119 ❤
Dan Shipper：Anthropomorphization of AI is good when it serves to help Builder 动态 · Source Excerpt
Dan Shipper 原文要点：Anthropomorphization of AI is good when it serves to help us use (understand, predict, etc) AI better.
Source-grounded excerpt from Dan Shipper: Anthropomorphization of AI is good when it serves to help us use (understand, predict, etc) AI better.
@danshipperX 原文87 ❤
Aaron Levie：As AI security events pick up, it’s going to be critical tAgent 安全 · Source Excerpt
Aaron Levie 原文要点：As AI security events pick up, it’s going to be critical that we have the most sophisticated AI agents for being able to detect and prevent security issues.
Source-grounded excerpt from Aaron Levie: As AI security events pick up, it’s going to be critical that we have the most sophisticated AI agents for being able to detect and prevent security issues.
@levieX 原文77 ❤
Peter Yang：Trust is going to be the biggest barrier (and driver) of pBuilder 动态 · Source Excerpt
Peter Yang 原文要点：Trust is going to be the biggest barrier (and driver) of personal agent adoption.
Source-grounded excerpt from Peter Yang: Trust is going to be the biggest barrier (and driver) of personal agent adoption.
@petergyangX 原文28 ❤
Dan Shipper：Pragmatists stay winning in the age of AIBuilder 动态 · Source Excerpt
Dan Shipper 原文要点：Pragmatists stay winning in the age of AI
Source-grounded excerpt from Dan Shipper: Pragmatists stay winning in the age of AI
@danshipperX 原文22 ❤
09 / 15
数据洞察 · Snapshot09 / 15
今日数据概览Today Stats
收录 Builder：10 位
总推文数：19 条
播客节目：1 期
博客文章：0 篇

最高互动：Thibault Sottiaux：If you still haven't tried Codex but have considered it in · 2.3K ❤
第二高互动：Nan Yu：Some personal news—I’m joining OpenAI to work on Codex and · 3.1K ❤
The snapshot includes 10 builders, 19 tweets, 1 podcast, and 0 blog post.
follow-buildersSnapshot follow-builders-2026-09-01-v1
5 条关键洞察5 Key Takeaways
ChatGPT Work 正从任务助手变成可调度上下文、应用与协作的新计算入口。
ChatGPT Work is becoming a computing surface that orchestrates context, apps, and collaboration.
模型越强，企业越需要连接数据、监管、人类决策点与现实反馈回路的 applied AI 层。
Stronger models increase demand for an applied-AI layer spanning data, regulation, human decisions, and real-world feedback.
开放权重从开发者偏好升级为产业领导力与生态扩散战略。
Open weights are moving from developer preference to an industrial strategy for leadership and ecosystem diffusion.
算力竞争的关键约束转向物理世界：电力、冷却、建设速度与每瓦 token 数。
Compute competition is increasingly constrained by power, cooling, build speed, and tokens per watt.
跨应用执行放大了价值，也把 Gmail、日历与办公数据的信任问题推到产品核心。
Cross-app execution expands value while making trust around Gmail, calendars, and office data a core product issue.
10 / 15
本周之声 VOICE 01
只要允许模型持续推理，并借助 canonical compaction 跨多个 context window 工作。
“You just have to allow it to reason and work over multiple context windows.”
@thsottiauxX 原文2.3K ❤ · 42 RT · 1.8K 💬
11 / 15
本周之声 VOICE 02
最好的路径，是赋能科学家，而不是试图由我们自己解决所有问题。
“The best way to do this is for us to empower scientists.”
@rauchgX 原文2.4K ❤ · 103 RT · 72 💬
12 / 15
本周之声 VOICE 03
我总会先创建 design.md，写清颜色、字体以及其他设计准则。
“I always create a design md file with guidelines for color, typography, and more.”
@levieX 原文119 ❤ · 13 RT · 17 💬
13 / 15
播客之声 PODCAST VOICE
AI 是一场“50 年的一夜成功”。
“AI's fifty year overnight success.”
Kevin Kelly · AI & I by EveryYouTube 完整播客
14 / 15
AI前沿每日脉动
AI Frontier Pulse · 2026.09.01
本期收录 10 位 Builder · 19 条推文 · 1 期播客 · 0 篇博客
ChatGPT Work · Applied AI · Open Weights · Tokens per Watt · Trust & Access
感谢阅读 · Thank You For Reading
Richard Liu · AI前沿每日脉动 · 2026 · Snapshot follow-builders-2026-09-01-v1