System translated (Gemini)

AI Daily | Stripe Unveils AI Programmer Cluster, Alibaba Releases Qwen 3.5-Omni Link to heading

Today, the AI industry focuses on the paradigm shift in the closed-loop capabilities of Agents. Stripe revealed its AI programmer cluster delivers 1,300 PRs per week, signaling that agents have entered large-scale engineering practice; simultaneously, the CLI interface is becoming the standard operating interface for Agents. At the model level, Alibaba released the Qwen 3.5-Omni series to enhance real-time multimodal interaction, while Stanford’s Meta-Harness framework drives the recursive evolution of AI agents by enabling them to self-optimize their code.

I. Today’s Highlights Link to heading

  • Watch List: The episode from Lenny’s Podcast, “🎙️ This week on How I AI: How Stripe built “minions”—AI coding agents that ship 1,300 PRs per week + How to turn Claude Code into your personal life operating system,” deserves priority attention. Key takeaway: Stripe engineers detail how they built “Minions”—an AI programmer cluster that delivers approximately 1,300 Pull Requests per week, triggered by Slack emojis.
  • Watch List: A total of 21 key updates have been tracked over the past 7 days. You can select sources (podcasts, newsletters, videos) for in-depth reading according to your needs.
  • X Hot Topic: The topic “Alibaba Launches Qwen3.5-Omni Multimodal AI Models” is trending (approximately 2,500 related posts) and serves as a good starting point for understanding global AI public opinion and perspectives.

II. AI Hot Topics on X (based on bird) Link to heading

Topic 1: Alibaba Launches Qwen3.5-Omni Multimodal AI Models Link to heading

  • Category: AI · News
  • Overview: Trending time: 9 hours ago, Related posts: 2,500
  • What it is: Alibaba released the Qwen3.5-Omni series of multimodal AI models, including versions like Max, Next, and Omni, aimed at enhancing advanced multimodal processing capabilities.
  • Why it matters: The launch of this model marks a significant breakthrough for domestic large models in the field of omni-architecture, further narrowing the gap with top international players like OpenAI in real-time multimodal interaction.
  • Discussion summary: Social media discussions are focused on the explosive growth of AI technology in December. Users are keen on comparing the performance of Qwen3.5 with OpenAI’s GPT-5.2 and Nvidia’s Nemotron 3, and discussing its potential impact on the AI industry landscape in 2025.

Topic 2: Stanford’s Meta-Harness Lets AI Agents Improve Their Own Code Link to heading

  • Category: AI · Other
  • Overview: Trending time: 2 hours ago, Related posts: 146
  • What it is: A Stanford University research team released the Meta-Harness framework, which allows AI agents to improve their task performance by iterating on and optimizing their own source code.
  • Why it matters: This research drives the shift of AI from “passive execution” to “self-evolution,” demonstrating the potential for agents to autonomously optimize their logic and achieve recursive performance enhancements.
  • Discussion summary: Discussions are centered on the significant performance improvements of Meta-Harness in programming benchmarks, as well as the potential security risks and long-term uncontrollability that could arise from AI self-modifying its code.

Topic 3: Alibaba Launches Free Preview of Qwen 3.6 Plus AI Model Link to heading

  • Category: AI · News
  • Overview: Trending time: 2 hours ago, Related posts: 245
  • What it is: Alibaba announced the launch of a free preview version of the Qwen 3.6 Plus AI model, allowing developers and users to experience its latest large language model capabilities for free for a limited time.
  • Why it matters: The release of this model reflects the rapid iteration of Chinese tech giants in AI performance. By providing free access to a high-performance model, they aim to lower the barrier to entry for developers and expand their influence in the global AI ecosystem.
  • Discussion summary: Discussions on X are focused on benchmark comparisons between Qwen 3.6 Plus and top models like GPT-4o, Alibaba’s rapid update cycle, and the impact of its free access strategy on the competitive landscape of AI commercialization.

Topic 4: Pretext Library Ends Web Text Layout Bottlenecks Link to heading

  • Category: AI · News
  • Overview: Trending time: 2 days ago, Related posts: 68,000
  • What it is: The Pretext library has been officially released, aiming to eliminate performance bottlenecks in web text layout processing and optimize the parsing and extraction flow of web data.
  • Why it matters: Efficient web page layout parsing can significantly improve the quality of training data for large models and enhance the ability of multimodal models and RAG systems to understand complex web page structures.
  • Discussion overview: The discussion focuses on Pretext’s rendering speed advantage, performance comparisons with existing web parsing tools, and its stability when processing large-scale dynamic web data.

Topic 5: Baidu’s PaddleOCR Tops GitHub Stars Over Google’s Tesseract Link to heading

  • Category: AI · News
  • Overview: Trending since: 12 hours ago, Related posts: 336
  • What it is: Baidu’s open-source deep learning OCR toolkit, PaddleOCR, has officially surpassed Google’s veteran OCR engine, Tesseract, in the number of stars on GitHub.
  • Why it matters: This marks a generational shift in the OCR technology field from traditional algorithms to deep learning frameworks and also reflects the growing influence of Chinese AI open-source projects in the global developer ecosystem.
  • Discussion overview: Community discussions focus on PaddleOCR’s performance advantages in multilingual recognition accuracy and inference speed, and the trend of developers gradually abandoning veteran tools like Tesseract in modern production environments.

Topic 6: Rebellions Raises $400 Million at $2.3 Billion Valuation for AI Inference Chips Link to heading

  • Category: AI · News
  • Overview: Trending since: 8 hours ago, Related posts: 118
  • What it is: South Korean AI chip startup Rebellions has completed a $400 million financing round at a $2.3 billion valuation to advance the development and commercialization of its AI inference chips.
  • Why it matters: This signals that the inference chip market is becoming the new center of gravity in AI hardware competition, reflecting the industry’s urgent need for more energy-efficient and cost-effective alternatives under Nvidia’s monopoly.
  • Discussion overview: The discussion focuses on the synergistic effects following Rebellions’ merger with Sapeon and the competitive advantages of its Application-Specific Integrated Circuit (ASIC) architecture over general-purpose GPUs in actual large-scale inference tasks.

Topic 7: Rajasthan Royals Crush CSK by Eight Wickets in IPL 2026 Opener Link to heading

  • Category: AI · Other
  • Overview: Trending since: 19 hours ago, Related posts: 102,000
  • What it is: In the 2026 IPL opener, the Rajasthan Royals defeated the Chennai Super Kings by a massive margin of 8 wickets.
  • Why it matters: The event showcased AI real-time data analysis, player performance prediction models, and automated broadcasting technologies, demonstrating their deep integration and practical application in a top-tier sporting event.
  • Discussion overview: Public discussion centered on the on-field performance of key players while also sparking technical discussions about the accuracy of AI win-probability prediction models during dynamic matches.

Topic 8: TWICE Hits Orlando Theme Parks After Sold-Out Shows Link to heading

  • Category: AI · Other
  • Overview: Trending since: 1 day ago, Related posts: 34,000
  • What it is: K-pop girl group TWICE visited local theme parks in Orlando after their sold-out concerts, and their activities generated significant attention on social media.
  • Why it matters: Although this topic falls under entertainment, it was categorized into the AI field by X’s platform algorithm, reflecting the recognition characteristics and limitations of current social media automated classification systems when handling high-profile pop culture events.
  • Discussion overview: The discussion primarily focused on photos from the members’ leisure activities, fan encounters, and some users’ confusion about the topic being miscategorized under the AI tag.

Topic 9: Blackpink’s Lisa Books Historic Las Vegas Residency Link to heading

  • Category: AI · Other
  • Overview: Trending since: 11 hours ago, Related posts: 161,000
  • What it is: Blackpink member Lisa announced she will launch a solo residency at the Voltaire theater in The Venetian Las Vegas, becoming the first K-pop solo artist to achieve this milestone.
  • Why it matters: Such top-tier residencies have become showcases for cutting-edge stage technology, AI-assisted visual design, and immersive audio technologies, demonstrating the deep integration of technology with the global pop culture industry.
  • Discussion overview: The discussion centered on the international breakthrough in Lisa’s solo career, the high demand for concert tickets, and the further elevation of K-pop artists’ status in the mainstream Western entertainment market.

Topic 10: BTS Tops Hot 100 and Billboard 200 with ARIRANG Debut Link to heading

  • Category: AI · Other
  • Overview: Trending since: 1 day ago, Related posts: 1,700,000
  • What it is: BTS, with their new work ‘ARIRANG,’ has simultaneously topped both the Billboard Hot 100 and Billboard 200 charts, setting a new pop culture record.
  • Why it matters: The work deeply integrates AI technology in its production or promotion, signaling that AI-assisted content has fully entered and is now dominating the global mainstream music market.
  • Discussion Summary: The discussion focuses on the role of AI in modernizing traditional folk music and the long-term impact of technological intervention on artistic authenticity and chart fairness in music.

AI Public Opinion Summary on X Today Link to heading

Today’s main AI discourse centers on the explosive iteration of domestic large models like the Qwen series and their enhanced global competitiveness through multimodal capabilities and free-to-use strategies. There is also a strong focus on fundamental technological breakthroughs enabling AI to shift from passive execution to a logic of “self-evolution.” The industry consensus is that AI is rapidly penetrating foundational infrastructure (such as inference chips and web parsing) and the popular culture industry. However, significant disagreements persist regarding the artistic authenticity of AI-assisted creation and the security of agents autonomously modifying code. Potential risks are not only reflected in the uncontrollability and security vulnerabilities of technological evolution, but also in the limitations of automated governance, as revealed by social media algorithms misclassifying topics, and the fairness disputes that may arise from technological intervention in traditional fields.

3. Influencer Insights Link to heading

Hello! I am a senior AI industry analyst. Based on the tweets from AI leaders on X over the past 24 hours, I have compiled today’s AI industry briefing for you.

Today’s core themes can be summarized as: the “Renaissance of the CLI” and the “Qualitative Leap in Agent’s Closed-Loop Capabilities.”

🚀 The CLI is Becoming the “Standard Operating Interface” for AI Agents Link to heading

The most significant trend today is the rush by major vendors to launch command-line interface (CLI) tools for AI.

  • Lark Open-Sources lark-cli: Lark has released a CLI tool that allows AI to directly manipulate documents, calendars, messages, and multi-dimensional tables. @dotey points out that the CLI is the most convenient interface for AI Agents because it is “self-describing” (learning via --help), text-friendly, and does not consume context space.
  • Industry Consensus: @op7418 believes the CLI is becoming the standard interface for Agents and has launched the CodePilot AI-native CLI management system. @vista8 predicts that companies like Google, Lark, and DingTalk will all provide controllable entry points for Agents via CLIs.

🤖 Claude Code’s Evolution: From “Writing Code” to “Controlling Computers” Link to heading

Anthropic’s Claude Code continues to dominate the charts, with its capabilities expanding once again.

  • Computer Use Integration: Claude Code now supports direct control of the macOS desktop. @AI_Jasonyu commented that it has evolved from “just writing code” to a “programmer agent that can actually get work done,” capable of automatically compiling, verifying UI with screenshots, and fixing bugs.
  • OpenAI’s “Reverse Play”: In a dramatic turn of events, OpenAI officially released the codex-plugin-cc plugin, proactively entering a competitor’s territory to allow developers to call Codex for code reviews within Claude Code (@dotey).

📏 Frontend Performance Breakthrough: Pretext Library Goes Viral Link to heading

The pure TypeScript text measurement library Pretext, released by former Midjourney employee Cheng Lou, has sparked heated discussion.

  • Performance Leap: The library can accurately calculate text layout dimensions without relying on the DOM, making it 500 times faster than traditional methods. @op7418 believes this is crucial for real-time UI generation in the AI era, as it solves the common flickering and layout overflow issues found in generative UIs.

2. Unique Perspectives and Industry Foresight Link to heading

  • Test Cases are the New “Moat”: @ruanyf argues that as AI becomes capable of replicating large software like Next.js at an extremely low cost, the code itself no longer serves as a moat. The key to preventing replication in the future lies in proprietary “test cases.”
  • Agents are the “New Demographic Dividend” for the Next 10 Years: @lijigang proposes that if we view Agents as a new population, the greatest certainty of the next decade lies in how to invest heavily in this “demographic dividend.” He also points out the need to research identity authentication systems and independent monetary systems for Agents.
  • The “Vibe Coding” Development Paradigm: @Pluvio9yte introduces a development model that moves away from getting bogged down in code logic, instead relying purely on aesthetics, intuition, and precise prompts to control the overall picture. He believes this is completely changing the productivity of independent developers.
  • The Threat of LLMs Autonomously Discovering Vulnerabilities: @Gorden_Sun cites research from Anthropic scientists indicating that next-generation models (like Mythos) are already capable of autonomously discovering and exploiting complex vulnerabilities in the Linux kernel, significantly elevating cybersecurity risks.
  • The Truth About Products from an SEO Perspective: @gefei55 presents a counter-intuitive view: a good experience doesn’t equal sales. SEO rankings sometimes favor slower-experience products that have a “longer dwell time.” He emphasizes that “done is better than perfect” and that precise traffic (SEO/SEM) is the core of international expansion.

3. Recommended Tools & Resources Link to heading

🛠 Productivity Tools Link to heading

  • lark-cli: An AI-friendly command-line tool open-sourced by Feishu, supporting calls to its full range of office capabilities (@dotey).
  • Pretext: A high-speed text measurement library, essential for front-end developers (GitHub: chenglou/pretext).
  • codex-plugin-cc: Officially from OpenAI, this tool introduces Codex review capabilities into Claude Code.
  • PaddleOCR: With more GitHub Stars than Tesseract, PP-OCRv5’s 5M parameters deliver extraction capabilities comparable to GPT-4o (@AI_Jasonyu).

💡 Learning & Research Link to heading

  • MSA (Memory Sparse Attention) Paper: Proposes enabling AI with a native ultra-long memory of 100 million tokens, approaching the scale of human lifelong memory (@vista8, @zhixianio).
  • Claude Code Hidden Features List: @dotey has compiled 15 underrated features recommended by its creators (e.g., /loop automation, --teleport remote control, etc.).
  • CC-BOS Framework: A paper studying how to use classical Chinese to “jailbreak” the safety guardrails of large models (@lijigang).

🎨 Creative Flow Link to heading

  • Vibe Coding Workflow: Nano Banana Pro 2 (Visual Inspiration) + Kling 3.0 (Animation Injection) + Lovable (AI Site Building) (@Pluvio9yte).
  • CapWords: A uniquely creative AI tool for learning foreign languages, combining AI-powered image cutouts with situational learning (@nishuang).

Analyst’s Take: Today’s trends indicate that the AI industry is shifting from a “model race” to an “interface and ecosystem race.” The return of the CLI is not a step backward but rather a high-speed highway being paved for Agents. Developers should focus on how to “CLI-ify” or “Skill-ify” their products to connect with the wave of automation driven by Claude Code and various Agents.

IV. Watch List Update Summary Link to heading

Timeframe: Last 7 days; 16 sources covered; 21 updates total

Y Combinator Podcast (B_intro+search) Link to heading

  • How François Chollet Is Building A New Path To AGI
    • Publication Time: 2026-03-27 23:00 Beijing Time
    • Abstract: - You’ve probably heard of OpenClaw (formerly Clawdbot/Moltbot).
      • The open-source AI assistant that made a splash can run on your own device, plugs into the messaging apps you already use, and goes beyond chat to actually do things—managing your email, calendar, files, workflows, and more.
      • Now meet the human behind it.
      • YC’s Raphael Schaad sits down with OpenClaw founder Peter Steinberger to talk about the ‘aha’ moment behind the viral personal AI agent, why local-first agents could replace many of the apps we use today, and how personal agents will reshape the future of software.
    • EN Key Points:
      • François Chollet has spent years asking a different question than most of the AI world
      • Instead of scaling what already works, he’s trying to understand what intelligence actually is—and how to build it from first principles
      • In this episode of Lightcone, he traces that path from his early work on deep learning to the creation of the ARC prize, and the launch of ARC V3, a new benchma…
  • He explains why today’s systems may be hitting limits, what recent breakthroughs really mean, and why reaching true general intelligence may require a fundament…

All-In Podcast (A_full) Link to heading

  • Anthropic’s Generational Run, OpenAI Panics, AI Moats, Meta Loses Lawsuits

    • Release Time: 2026-03-28 04:03 Beijing Time
    • Summary: - (0:00) Bestie intros.
      • : Friedberg for Governor of California.
      • (2:25) Anthropic’s generational run (15:45) OpenAI: getting focused or panic mode.
      • (36:56) AI valuation impacts, moats, and disruption (43:58) Liquidity speaker announcements, the 100x AI moment (50:35) Two landmark socia….
    • EN Highlights:
      • (0:00) Bestie intros
      • : Friedberg for Governor of California
      • (2:25) Anthropic’s generational run (15:45) OpenAI: getting focused or panic mode
      • (36:56) AI valuation impacts, moats, and disruption (43:58) Liquidity speaker announcements, the 100x AI moment (50:35) Two landmark socia…

  • Bryan Johnson: I Just Took the Most Powerful Dose of DMT in the World… Here’s What It Was Like

    • Release Time: 2026-03-26 08:10 Beijing Time
    • Summary: - The Pod by 8 Sleep cools your bed to 55°F and uses Autopilot to optimize your sleep all night.
      • Follow Bryan: Follow friends: Follow X: Follow Instagram: Follow TikTok: Follow LinkedIn: Intro music source: Intro video source: (0:00) David Friedberg intros Bryan Johnson (0:54) Why Bryan Johnson did 5-MeO-DMT (12:56) What brain scans actually show (18:36) Psychosis, bad trips, and life-changing decisions (26:23) The next frontier: organoids and gene therapy (33:26) GLP-1s, abundance, and human optimization (35:35) The longevity drug no one is talking about?
      • Follow Bryan: Follow friends: Follow X: Follow Instagram: Follow TikTok: Follow LinkedIn: Intro music source: Intro video source: (0:00) David Friedberg intros Bryan Johnson (0:54) Why Bryan Johnson did 5-MeO-DMT (12:56) What brain scans actually show (18:36) Psycho… Thanks to our partner for making this happen The Pod by Eight Sleep cools your bed to 55°F and uses Autopilot to optimize your sleep all night, use code ALLIN for up to $350 off.
    • EN Highlights:
      • (0:00) David Friedberg intros Bryan Johnson (0:54) Why Bryan Johnson did 5-MeO-DMT (12:56) What brain scans actually show (18:36) Psycho…
      • Thanks to our partner for making this happen
      • The Pod by Eight Sleep cools your bed to 55°F and uses Autopilot to optimize your sleep, all night
      • Use code ALLIN at https://www.eightsleep.com/allin for up to $350 off. Follow Bryan:

Lenny’s Podcast (A_full) Link to heading

  • 🎙️ This week on How I AI: How Stripe built “minions”—AI coding agents that ship 1,300 PRs per week + How to turn Claude Code into your personal life operating system

    • Publication Time: 2026-03-30 23:03 Beijing Time
    • Abstract: In this episode, Stripe engineer Steve Kaliski details how they built “Minions”—a cluster of AI programmers that can deliver about 1,300 Pull Requests per week, usually triggered by a single emoji in Slack.
    • EN Key Points:
      • ,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F361d81ef-7faf-4d8e-8028-5d5e03432a9a_2329x551.pn…
      • ,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F361d81ef-7faf-4d8e-8028-5d5e03432…

  • How to turn Claude Code into your personal life operating system | Hilary Gridley

    • Publication Time: 2026-03-30 20:04 Beijing Time
    • Abstract:
      • Hilary Gridley is an entrepreneur, former product lead, and new mother who previously discussed AI for managers on the podcast.
      • She returns to share her method for transforming personal productivity by using Claude Code as her primary tool for managing professional and personal life.
      • Hilary showcases her “anti-system system”—a philosophy that prioritizes simplicity over complex setups, allowing the AI to learn preferences through observation rather than pre-configuration.

  • Listen or watch on YouTube, Spotify, or Apple Podcasts. What you’ll learn:

    • How to instantly capture to-dos with a simple iPhone back-tap shortcut that requires zero app switching
    • The “10x Impact Framework” for deciding which tasks to automate and which to apply human effort to
    • How to use Claude Code’s observation feature to build a preference file that improves over time without manual setup
    • Why the “yappers API” (talking about what you’re doing while you work) eliminates the need for complex OAuth integrations
    • A workflow for breaking down daunting tasks into 10-minute first steps that actually get done
    • How to create Claude skills by simply describing the problem instead of writing code or following a tutorial
    • A technique for using “recording mode” to demo workflows without exposing personal information

    Brought to you by: WorkOS — Get your app enterprise-ready in minutes. Lovable — Build apps by simply chatting with an AI.

    In this episode, we cover: (00:00) Introduction to Hilary Gridley (02:43) The opportunity cost of time as a new mom and entrepreneur (07:11) The anti-system philosophy (08:05) Demo: Planning your day with Claude Code (10:00) Setting up a simple iPhone task capture shortcut (11:48) How Claude automatically organizes reminders and learns preferences (16:19) Breaking down daunting tasks into manageable first steps (23:40) The yappers API: Talking to Claude instead of building integrations (25:28) Daily logging and observation mode (27:45) Quick recap (30:50) The power of screenshots (32:55) The 10x Impact Framework for automation decisions (37:51) Applying the framework to different career stages (39:29) Building a “skill” for recording anonymized demos (44:11) Building a return tracking skill from scratch (48:31) Building muscle memory for using AI tools (50:18) Where to find Hilary

    Reference Tools: • Claude Code: • Obsidian: • iPhone Shortcuts: • Cursor:

    Other References: • Figma file of Hilary’s demo:

    Where to find Hilary Gridley: Substack: Website:

    Where to find Claire Vo: ChatPRD: Website: LinkedIn: X:

    Production and marketing.

  • Watch now | 🎙️ How a new mom and entrepreneur automates her life with Claude Code—all without complex setup. Hilary Gridley is an entrepreneur, formerly in product… She returns to share how she has transformed her approach to personal productivity using Claude Code as her primary tool for managing both professional work and personal life… Hilary demonstrates her “anti-system system”—a philosophy that prioritizes simplicity over complex setups, allowing the AI to learn preferences through observation… i=1000758198579">Apple Podcasts. What you’ll learn: - How to instantly capture to-dos using a simple iPhone back-tap feature… How to turn Claude Code into your personal life OS | Hilary Gridley.

    • EN Key Points:
      • Hilary Gridley is an entrepreneur, former product…
      • She returns to share how she’s transformed her approach to personal productivity using Claude Code as her primary tool for managing both professional work and l…
      • Hilary demonstrates her “anti-system system”—a philosophy that prioritizes simplicity over complex setup, allowing AI to learn preferences through observa… What you’ll learn:

  • How to capture to-dos instantly using a simple iPhone back-tap…

  • From skeptic to true believer: How OpenClaw changed my life | Claire Vo

    • Published: 2026-03-29 20:31 Beijing Time

  • Abstract: - Claire Vo is the host of our sister podcast, “How I AI,” a former product lead and engineer, and the founder of an AI startup called ChatPRD.

    • Claire now runs her business, podcast, and family life with the help of 9 OpenClaw agents running on multiple Mac Minis and old laptops.
    • In this episode, Claire shares her journey from an OpenClaw skeptic (it deleted her family calendar the first time she tried it) to a true believer, and provides a masterclass on using AI agents in real life.
    • The exact step-by-step process for installing and setting up OpenClaw (it’s easier than you think) 2.
    • How to avoid the biggest OpenClaw mistakes (don’t install it on your main computer) 3.

  • EN Highlights:

    • Claire Vo is the host of our sister podcast, “How I AI,” a former product executive and engineer, and founder of an AI startup called ChatPR…
    • Claire now runs her business, podcast, and family life with the help of nine OpenClaw agents running on multiple Mac Minis and old laptops
    • In this episode, Claire shares her journey from OpenClaw skeptic (it deleted her family calendar the first time she tried it) to true believer, and gives a mast…
    • The exact step-by-step process to install and set up OpenClaw (it’s easier than you think) 2

  • 🧠 Community Wisdom: When AI velocity outpaces your product strategy, when your estimates keep slipping, one day in San Francisco, pairing Claude Code with Codex, and more

    • Published: 2026-03-29 01:01 Beijing Time
    • Abstract: - 👋 Hello, and welcome to this week’s ✨ Community Wisdom ✨ a subscriber-only email, sent every…
      • ,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1ca8352-9571-4b81-b041-52374176f4fc_2912x1456.p….
      • ,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1ca8352-9571-4b81-b041-52374176f….
    • EN Highlights:
      • 👋 Hello and welcome to this week’s edition of ✨ Community Wisdom ✨ a subscriber-only email, delivered every…
      • ,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1ca8352-9571-4b81-b041-52374176f4fc_2912x1456.p…

  • ,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa1ca8352-9571-4b81-b041-52374176f…

  • How Stripe built “minions”—AI coding agents that ship 1,300 PRs weekly from Slack reactions | Steve Kaliski (Stripe engineer)

    • Published: 2026-03-25 20:03 Beijing Time
    • Summary: - Steve Kaliski is a software engineer at Stripe, where for the past six and a half years, he has been focused on building developer tools and payment infrastructure.
      • He is part of the team that created “minions,” Stripe’s internal AI coding agents, which now send approximately 1,300 pull requests per week with minimal human intervention, apart from code review.
      • In this episode, Steve demonstrates how Stripe engineers activate development work from Slack, utilize a cloud-based development environment to enable parallel agent workflows, and showcases machine-to-machine payments where an AI agent autonomously transacts with third-party services.
      • Listen or watch on YouTube, Spotify, or Apple Podcasts. What you will learn: - How Stripe’s “minions” write 1,300 pull requests per week with minimal human intervention - Why a good human developer experience creates better outcomes for AI agents - The key role of cloud development environments in unlocking AI-driven engineering speed - Machine payment protocols that allow AI agents to spend money to complete tasks - Code review strategies for handling thousands of agent-written PRs - Why Stripe’s non-engineers are starting to use Minions to deliver code - The future of software businesses built primarily for agent consumers. Brought to you by: Optimizely — The AI agent orchestration platform for marketing and digital teams. Rippling — Stop wasting time on administrative tasks and build your startup faster. In this episode, we will cover: (00:00) Steve’s Introduction (02:39) Stripe’s Minions and their impact across Stripe (04:42) Why activation energy is more important than execution (05:44) What is a minion?
      • Technical Architecture (06:52) Demo: Activating minions in Slack with an emoji (09:04) Why a good developer experience is good for both humans and agents (11:22) Exploring the agent loop and system prompts (13:42) Why Stripe chose Goose as an agent tool (16:00) The role of Stripe’s developer productivity team (17:15) Why cloud environments unlock multi-threaded AI engineering (21:14) One-shot prompt: From Slack to a delivered PR (22:04) How Stripe handles code review for 1,300 AI-written PRs per week (23:44) Non-engineers using Minions across the company (24:53) Demo: Planning a birthday party with Claude and machine payments (32:15) Quick Recap (35:08) The future of ephemeral, API-first businesses for agents (36:36) Lightning round and final thoughts. Reference Tools: • Goose (AI agent tool): • Claude Code: • Cursor: • VS Code: • Slack: • Browserbase: • Parallel AI: • PostalForm: • Stripe Climate: Other References: • Stripe Machine Payments: • Blue-Green Deployment: • Git Worktrees: Where to find Steve Kaliski: Twitter: LinkedIn: Where to find Claire Vo: ChatPRD: Website: LinkedIn: X: Production and Marketing.
    • EN Highlights:
      • Steve Kaliski is a software engineer at Stripe wh…
      • He’s part of the team that created “minions”—Stripe’s internal AI coding agents, which now ship approximately 1,300 pull requests…

  • In this episode, Steve demonstrates how Stripe engineers activate development work from Slack and leverage cloud-based development environments for parallel age… What you’ll learn:

  • How Stripe’s “minions” write 1,300 pull requ…

Stratechery by Ben Thompson (A_full) Link to heading

  • Spring Break

    • Release Time: 2026-03-30 18:00 Beijing Time
    • Summary: - Stratechery’s spring break is a bit chaotic, as my usual week-long vacation will be spread out:
        • There will be no update on Thursday, March 19.
        • There will be no updates on Monday and Tuesday, March 23-24; there will be an update and an interview on Wednesday and Thursday, March 25-26.
        • There will be no update on Monday, March 30.
      • All other Stratechery Plus content, including my podcast, will proceed as scheduled.
    • EN Highlights:
      • Stratechery is on a bit of a disjointed Spring Break, as my usual week off will be spread out:
        • There will be no Update on Thursday, March 19
        • There will be no Update on Monday and Tuesday, March 23–24; there will be an Update and Interview on Wednesday and Thursday, March 25–26

  • 2026.13: So Long to Sora

    • Release Time: 2026-03-28 01:00 Beijing Time
    • Summary: - Welcome back to This Week in Stratechery!
      • As a reminder, every Friday, we send out an overview of the content in the Stratechery bundle; highlighted links are free for everyone.
      • Additionally, you have full control over the content we send you.
      • With that said, here are some of our favorites from this week.
      • Sora, 2025-2026.** AI Sam came, AI Sam saw, AI Sam stole those GPUs.
    • EN Highlights:
      • Welcome back to This Week in Stratechery
      • As a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle; hig…

  • An Interview with Arm CEO Rene Haas About Selling Chips

    • Release Time: 2026-03-26 18:06 Beijing Time
    • Summary: - Listen to this article: Good morning, this week’s Stratechery interview is with Arm CEO Rene Haas, whom I previously spoke with in January 2024. He recently made a major announcement during Arm’s first independent keynote: the long-standing IP licensing company is undergoing a huge shift in its business model and is selling its own chips for the first time.
      • We delve deep into this decision in the interview, covering the keynote’s metadata, Arm’s history, and how the company has evolved, especially under Haas’s leadership.
      • Then we get into why CPUs are important for AI and how Arm’s CPUs compare to those from Nvidia, x86, and other custom Arm chips.
      • Finally, we discuss the risks Arm faces, including supply chain limitations, and how the company needs to change to support this new direction.
      • As a reminder, all Stratechery content (including interviews) is available as a podcast; click the link at the top of this email to add Stratechery to your podcast player.
    • EN Highlights:

  • Listen to this post:

  • Arm Launches Own CPU, Arm’s Motivation, Constraints and Systems

    • Publication Time: 2026-03-25 18:00 Beijing Time
    • Summary: - Arm is selling its own chips, not just licensing IP.
      • This is a huge change compared to Arm’s history, but it’s not surprising given how computing is evolving.
      • Arm is selling its own chips, not just licensing IP. This is a big change compared to Arm’s history, but it’s not surprising given how computing is evolving.
      • Arm is launching its own CPU, Arm’s motivations, constraints, and systems.
    • EN Highlights:
      • Arm is selling its own chips, not just licensing IP
      • It’s a big change compared to Arm’s history, but not surprising given how computing is evolving.

OpenAI Blog (A_full) Link to heading

  • Helping disaster response teams turn AI into action across Asia

    • Publication Time: 2026-03-30 06:15 Beijing Time
    • Summary: Today in Bangkok, in partnership with the Gates Foundation, the Asian Disaster Preparedness Center, and DataKind, we convened 50 disaster management experts from South and Southeast Asia to discuss how AI can accelerate the response efficiency of governments and non-profit organizations in times of crisis.
    • EN Highlights:
      • AI for Disaster Response in Asia: OpenAI Workshop with Gates Foundation

  • STADLER reshapes knowledge work at a 230-year-old company

    • Publication Time: 2026-03-28 06:00 Beijing Time
    • Summary: - STADLER is a family-owned business with over 230 years of history, specializing in providing automated waste sorting plants for the global recycling industry.
      • The company has over 650 employees worldwide and plays a key role in helping countries advance their sustainable development and circular economy goals.
      • Under the leadership of co-CEO Julia Stadler, the company has adopted a forward-looking approach to modernization—embedding artificial intelligence as a core productivity layer in daily work.
      • Since 2023, STADLER has followed a clear principle: every employee working on a computer should use AI to improve speed, quality, and collaboration.
    • EN Highlights:
      • Learn how STADLER uses ChatGPT to transform knowledge work, saving time and accelerating productivity across 650 employees.

  • Inside our approach to the Model Spec

    • Publication Time: 2026-03-25 18:00 Beijing Time
    • Summary: This article explores how OpenAI’s Model Spec serves as a public framework for guiding model behavior, balancing safety, user freedom, and accountability as AI systems continue to evolve.
    • EN Highlights:
      • Learn how OpenAI’s Model Spec serves as a public framework for model behavior, balancing safety, user freedom, and accountability as AI systems advance.

  • Introducing the OpenAI Safety Bug Bounty program

    • Publication Time: 2026-03-25 08:00 Beijing Time
    • Summary: - OpenAI is launching a security bug bounty program to identify AI abuse and security risks, including agent vulnerabilities, prompt injections, and data leakage.
      • This article from the OpenAI blog explains how the introduction of the OpenAI Safety Bug Bounty program is shaping the broader AI and infrastructure landscape.
      • Following its launch, the OpenAI Safety Bug Bounty program also has practical implications for founders, operators, and investors.
    • EN Highlights:

  • OpenAI launches a Safety Bug Bounty program to identify AI abuse and safety risks, including agentic vulnerabilities, prompt injection, and data exfiltration.

Google DeepMind Blog (A_full) Link to heading

  • Gemini 3.1 Flash Live: Making audio AI more natural and reliable

    • Publication Time: 2026-03-26 23:23 Beijing Time
    • Summary: - Our latest voice model improves accuracy and reduces latency, making voice interactions smoother, more natural, and more precise.
      • This article from the Google DeepMind blog explains how Gemini 3.1 Flash Live: Making audio AI more natural and reliable shapes the broader AI and infrastructure landscape.
      • It also brings practical implications for the founders, operators, and investors of Gemini 3.1 Flash Live: Making audio AI more natural and reliable.
    • EN Key Points:
      • Our latest voice model has improved precision and lower latency to make voice interactions more fluid, natural and precise.

  • Protecting people from harmful manipulation

    • Publication Time: 2026-03-26 00:46 Beijing Time
    • Summary: - Google DeepMind researches the risks of harmful manipulation by AI in fields like finance and health, leading to the creation of new safety measures.
      • This article from the Google DeepMind blog explains how protecting people from harmful manipulation shapes the broader artificial intelligence and infrastructure landscape.
      • It also brings practical implications for founders, operators, and investors after protecting people from harmful manipulation.
    • EN Key Points:
      • Google DeepMind researches AI’s harmful manipulation risks across areas like finance and health, leading to new safety measures.

  • Lyria 3 Pro: Create longer tracks in more

    • Publication Time: 2026-03-26 00:01 Beijing Time
    • Summary: - Introducing Lyria 3 Pro, which unlocks longer tracks through structural awareness.
      • We are also introducing Lyria to more Google products and interfaces.
      • Introducing Lyria 3 Pro, which unlocks longer tracks through structural awareness. We are also introducing Lyria to more Google products and interfaces.
    • EN Key Points:
      • Introducing Lyria 3 Pro, which unlocks longer tracks with structural awareness
      • We’re also bringing Lyria to more Google products and surfaces.

Two Minute Papers (B_intro+search) Link to heading