System translated (Gemini)

AI Daily | Anthropic Confirms New Model Claude Mythos, OpenAI Pushes for Codex Pluginization Link to heading

Today, the AI industry is focused on the strategic rivalry between Anthropic and OpenAI. Anthropic has confirmed its new-generation model, Claude Mythos, which shows significant performance improvements in programming and security. OpenAI, meanwhile, is entering a period of strategic consolidation, deepening its focus on the productivity ecosystem by shutting down non-core businesses and launching the Codex plugin system. On the technical front, the industry consensus is shifting from pure “reasoning” to “Agentic Thinking,” emphasizing a paradigm shift for AI from thinking to taking practical action.

I. Today’s Highlights Link to heading

  • Watch List: The post “STADLER reshapes knowledge work at a 230-year-old company” from the OpenAI Blog is worth prioritizing. Key points: - STADLER is a family-owned company 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 co-CEO Julia…
  • Watch List: A total of 30 key updates have been captured in the last 7 days. You can select sources (Podcasts/Newsletters/Videos) for in-depth reading based on your needs.
  • X Hot Topic: The topic “Anthropic Confirms Leaked Claude Mythos AI Model Tests” is trending with high engagement (approx. 44,000 related posts), serving as a good entry point for understanding global AI public opinion and viewpoints.

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

Topic 1: Anthropic Confirms Leaked Claude Mythos AI Model Tests Link to heading

  • Category: AI · News
  • Overview: Trending time: 20 hours ago, Related posts: 44,000
  • What it is: Anthropic has officially confirmed the authenticity of a test version of a new AI model named “Claude Mythos” that was previously circulating online.
  • Why it matters: This signals that Anthropic is advancing the next major iteration of its large language model, aiming to maintain its leading position in the AI performance hierarchy with the new version.
  • Discussion summary: Discussions on X are focused on the specific positioning of Mythos (whether it’s a codename for Claude 4 or a version optimized for specific tasks), its performance in benchmarks, and whether it can surpass OpenAI’s latest models.

Topic 2: Trump Extends Iran Deadline to April 6 Amid Missile Strikes on Israel Link to heading

  • Category: AI · News
  • Overview: Trending time: 2 days ago, Related posts: 448,000
  • What it is: Following missile strikes on Israel by Iran, Trump announced an extension of the ultimatum deadline for Iran to April 6.
  • Why it matters: Such major geopolitical conflicts provide a key window for observing the real-world application of AI in monitoring and early warning, disinformation detection, and autonomous defense systems.
  • Discussion summary: The discussion focuses on the misleading impact of AI-generated fake war imagery on public opinion, the bias in how social media algorithms push conflict-related information, and the role of AI in missile interception technology.

Topic 3: Hegseth Removes Four Army Officers from Promotion List Amid Bias Claims Link to heading

  • Category: AI · Other
  • Overview: Trending time: 11 hours ago, Related posts: 76,000
  • What it is: U.S. Secretary of Defense nominee Hegseth removed four Army officers from the promotion list, citing concerns of bias (primarily related to DEI policies).
  • Why it matters: This event reflects a sharp shift in the U.S. administration’s direction from “diversity” to “meritocracy,” a trend that is profoundly impacting the governance logic, talent selection criteria, and political discourse of Silicon Valley and the AI field.
  • Discussion summary: The debate centers on whether this move is a restoration of military professionalism or a form of political vetting, and the potential impact of this ideological shift on future defense technology and AI strategic leadership.

Topic 4: Z.ai Unveils GLM-5.1, Nipping at Claude Opus 4.6 in Coding Benchmarks Link to heading

  • Category: AI · News
  • Overview: Trending time: 12 hours ago, Related posts: 2,400
  • What it is: Zhipu AI (Zhipu AI) has released the GLM-5.1 model, whose performance in coding benchmarks is now closely approaching that of Claude Opus 4.6.
  • Why it matters: This marks a significant milestone, indicating that domestically developed large models have reached top-tier international performance on highly difficult tasks like programming, further narrowing the technology gap with leading companies such as OpenAI and Anthropic.
  • Discussion Summary: The discussion focuses on the actual code generation quality of GLM-5.1, the risk of overfitting in benchmarks, and its potential to replace existing tools in developers’ actual workflows.

Topic 5: AI Researcher Impressed by China’s Robotics Surge After Month-Long Trip Link to heading

  • Category: AI · News
  • Summary: Trending Time: 22 hours ago, Related Posts: 360
  • What happened: After a month-long field trip to China, an AI researcher expressed shock at the explosive growth of China’s robotics industry, the speed of technological iteration, and the advantages of its hardware supply chain.
  • Why it matters: This highlights China’s global competitiveness in the field of Embodied AI hardware and signals an acceleration in the transition of robots from laboratories to large-scale industrialization.
  • Discussion Summary: The discussion centers on the “Shenzhen Speed” of Chinese hardware manufacturing, its extremely high cost-performance advantage, and the potential gap between Western countries and China in terms of supply chain integrity for robotics.

Topic 6: Stripe Launches Projects CLI to Simplify Service Provisioning for Developers Link to heading

  • Category: AI · News
  • Summary: Trending Time: 4 hours ago, Related Posts: 296
  • What happened: Stripe has released a new Projects CLI tool aimed at simplifying the service configuration and resource provisioning process for developers building applications via the command line.
  • Why it matters: This tool lowers the barrier for AI developers to integrate payment and related infrastructure, helping to accelerate the transition from AI prototypes to commercial products.
  • Discussion Summary: Community discussion has mainly focused on the significant improvement in development efficiency this tool provides and whether Stripe is transitioning from a payment giant to a more comprehensive developer infrastructure platform.

AI Public Opinion Summary on X Today Link to heading

Today’s main AI narrative focuses on the comprehensive escalation of the global tech race across both software algorithms and hardware supply chains. There is a broad consensus on the rapid iteration of large model performance and China’s large-scale advantages in the field of embodied intelligence. However, on a technical level, public opinion shows cautious skepticism about the authenticity of model benchmarks. On the governance level, there are serious disagreements regarding the ideological reshaping triggered by the shift in U.S. defense policy from “diversity” to “meritocracy.” Potential risks not only point to the misleading of public perception by AI-driven misinformation in geopolitical conflicts, but also involve the profound impact that drastic shifts in political winds could have on the stability of global AI strategic leadership and the developer ecosystem.

3. Influencer Insights Link to heading

Hello! I am an AI industry analyst. Based on the activities of AI leaders and senior developers on X over the past 24 hours, I have compiled today’s AI industry briefing for you.

Today’s core keywords are: “From Reasoning to Action (Agentic Thinking)”, “OpenAI’s Pre-IPO Cleanup”, and “Anthropic’s Next Behemoth”.

🚀 Paradigm Shift from “Reasoning Thinking” to “Agentic Thinking” Link to heading

The most in-depth industry discussion today comes from a long post by Justin Lin (@JustinLin610), former tech lead for Alibaba’s Qwen. He points out that the next stage for large models is not simply about making them “think longer” (Reasoning), but about “thinking in order to act”.

  • Core Idea: True agents need to address questions like “When to stop thinking and start acting?” and “How to process environmental feedback?”
  • Industry Resonance: @dotey provided a detailed interpretation of this trend, believing that future competitive advantages will come from environmental design, sampling infrastructure, and interface design for multi-agent collaboration.

🛡️ Anthropic Leaks New Model “Claude Mythos” Link to heading

Due to a CMS configuration error, Anthropic leaked a new model tier codenamed Capybara — Claude Mythos.

  • Performance: Positioned above Opus, it is the fourth-tier “strongest model,” significantly outperforming Opus 4.6 in programming and cybersecurity tests.
  • Risk: Official documentation warns that it poses an “unprecedented cybersecurity risk,” potentially heralding a wave of AI-driven exploits (@dotey, @op7418).

🧹 OpenAI’s “Pre-IPO Cleanup” and Codex Pluginization Link to heading

Within a week, OpenAI has successively cut the Sora app, a built-in checkout feature, and adult chat functionalities, aiming to clean up high-cost, low-growth, or legally risky businesses before its IPO (@dotey).

  • Strategic Shift: Focusing resources on productivity tools. OpenAI Codex officially launched its Plugins system today, supporting tools like Slack, Figma, and Notion, directly competing with the Claude Code ecosystem (@op7418).

🦞 OpenClaw (Crawfish) and the CLI Agent Craze Link to heading

Autonomous agents, represented by OpenClaw, continue to gain popularity, even prompting major companies (like Taobao and WeChat) to quickly follow suit or integrate them (@AI_Jasonyu, @zhixianio).

  • CLI Trends: Stripe launched the Stripe Projects CLI, and Cline introduced Kanban task management. Agents are comprehensively taking over terminal operations.

2. Unique Perspectives & Industry Foresight Link to heading

  • Observations on the “Asymmetry” of China’s AI Ecosystem: After an in-depth investigation of China, VC investor José Maria Macedo pointed out that China’s hardware supply chain advantage (the Shenzhen model) is incredibly formidable, but the software sector suffers from a “V2 bubble,” meaning most products are iterations of existing ones rather than original creations. He believes China’s top founders are often the “mavericks” from outside the major tech companies (translated by @dotey).
  • Testing is the New Moat: @ruanyf argues that in an era where AI can replicate any large-scale software at a low cost (e.g., replicating Next.js for $1100), the code itself is no longer a moat. Comprehensive test cases will become the key barrier to prevent easy replication by AI.
  • Foundation Models “Know What They Don’t Know”: @vista8 shared a recent paper from Apple, which points out that foundation models inherently have the ability to assess the confidence of their answers. However, RLHF (Reinforcement Learning from Human Feedback) can undermine this “honesty,” causing models to become “blindly confident” in their attempt to please humans.
  • Identity and Monetary Systems for Agents: @lijigang raises two forward-looking questions: How to design an identity card system for agents (cross-model identity authentication)? And how to give agents a “wallet” (fiat currency, BTC, or a universal token)?

3. Recommended Tools & Resources Link to heading

🛠️ Productivity Tools Link to heading

  • Type4Me: A local voice input tool for macOS that supports text optimization via large models, claiming to be an alternative to Typeless (recommended by @dotey).
  • Cline Kanban: An open-source, parallel multi-agent orchestration tool that supports task management with cards and branch visualization (recommended by @op7418).
  • CodePilot 0.40.0: An agent product that fully embraces the CLI, with built-in AI-friendly tools like FFmpeg and NetEase Cloud Music (recommended by @op7418).
  • Vue Lynx: Evan You (@Huxpro) independently completed this cross-platform framework integration in two weeks with the help of AI, demonstrating an effective methodology for AI-driven development.

📚 Learning Resources Link to heading

  • Official OpenAI Prompt Pack: In-depth prompt templates covering professions such as HR, engineers, and executives (recommended by @Pluvio9yte).
  • Claude Code Cheat Sheet: Quickly master the commands for Anthropic’s command-line tool (recommended by @Gorden_Sun).
  • GEO Prompt Open Source Library: Teacher Yao Jingang shared 27 sets of Generative Engine Optimization (GEO) prompts (recommended by @vista8).

⚠️ Special Reminders Link to heading

  • Claude Code Version Warning: @Pluvio9yte warns that version 2.1.81 has a cache anomaly that causes a sharp increase in token consumption. It is recommended to roll back to the stable version 2.1.77.
  • Claude Quota Adjustment: Anthropic has adjusted the quota consumption rate during peak hours (8 PM to 2 AM Beijing time). Users should avoid running heavy tasks during this period (@Pluvio9yte).

Analyst’s Comment: The current AI competition has shifted from a “model parameter race” to an “engineering and agentification race.” The contraction of OpenAI and the expansion of Anthropic present a stark contrast, while developers are rapidly turning the terminal (CLI) into the main battleground for agents. It is advisable to pay attention to investment opportunities in infrastructure related to Agentic RL (Agentic Reinforcement Learning).

IV. Watch List Update Summary Link to heading

Time frame: Last 7 days; 16 sources covered; 30 updates in total

a16z Podcast (A_full) Link to heading

  • Patrick Collison on Stripe’s Early Choices, Smalltalk, and What Comes After Coding
    • Release Time: 2026-03-24 18:00 Beijing Time
    • Abstract: - Michael Truell, CEO of Cursor, sits down with Patrick Collison, CEO of Stripe and investor in Anysphere, to talk about Collison’s history using Smalltalk and Lisp, the MongoDB and Ruby decisions Stripe is still sticking with 15 years later, why he’d spend more time on API design if he could do it all over, and whether AI is actually showing up in the economic productivity data.
  • This episode originally aired on Cursor’s podcast.
  • Follow Patrick Collison on X: Follow Michael Truell on X: Follow Cursor: Check out everything a16z is doing with AI here, including articles, projects, and more podcasts.
  • Please note that the content herein is for informational purposes only; it should not be considered legal, business, tax, or investment advice or used to evaluate any investment or security; and it is not intended for any investors or potential investors of any a16z fund.
  • a16z and its affiliates may maintain investments in the companies discussed.
  • EN Highlights:
  • Michael Truell, CEO of Cursor, sits down with Patrick Collison, CEO of Stripe and an investor in Anysphere, to talk about Collison’s history with Smalltalk a…
  • This episode originally aired on Cursor’s podcast.
  • Follow Patrick Collison on X:    https://twitter.com/patrickc
  • Follow Michael Truell on X: https://twitter.com/mntruell

Y Combinator Podcast (B_intro+search) Link to heading

  • How François Chollet Is Building A New Path To AGI
    • Release Date: 2026-03-27 23:00 Beijing Time
    • Summary: - You may have heard of OpenClaw (formerly known as Clawdbot/Moltbot).
      • The viral sensation is an open-source AI assistant that runs on your own device, connects with messaging apps you already use, and goes beyond chat to actually execute tasks like managing your email, calendars, files, workflows, and more.
      • Now meet the person behind it.
      • YC’s Raphael Schaad sat down with Peter Steinberger, the creator of OpenClaw, to discuss the “aha” moment behind the viral personal AI agent, why local-first agents could replace many of today’s apps, and how personal agents will reshape the future of software.
    • EN Highlights:
      • 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

    • Publication Time: 2026-03-28 04:03 Beijing Time
    • Abstract: - (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…

  • Are Psychedelics the Key to Living Forever? (ft. Bryan Johnson)

    • Publication Time: 2026-03-26 08:10 Beijing Time
    • Abstract: - The Pod by 8 Sleep can cool your bed to 55°F and uses Autopilot to optimize your sleep all night.
      • Follow Bryan: Follow friends: Follow on X: Follow on Instagram: Follow on TikTok: Follow on LinkedIn: Intro Music Source: Intro Video Source: (0:00) David Friedberg intros Bryan Johnson (0:54) Why Bryan Johnson does 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-1, abundance, and human optimization (35:35) The longevity drug no one is talking about?
      • Follow Bryan: Follow friends: Follow on X: Follow on Instagram: Follow on TikTok: Follow on LinkedIn: Intro Music Source: Intro Video Source: (0:00) David Friedberg intros Bryan Johnson (0:54) Why Bryan Johnson does 5-MeO-DMT (12:56) What brain scans actually show (18:36) Psycho… Thanks to our partners for making this happen The Pod by Eight Sleep cools your bed to 55°F and uses Autopilot to optimize your sleep, all night. Get up to $350 off with code ALLIN.
    • 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: <a href=“https…

  • Four CEOs on the Future of AI: CoreWeave, Perplexity, Mistral, and IREN

    • Publication Time: 2026-03-24 02:07 Beijing Time

  • Summary: - (0:00) Intro live from Nvidia GTC (0:37) CoreWeave CEO, Michael Intrator (32:58) Perplexity CEO, Aravind Srinivas (1:07:11) Mistral CEO,….

    • It all happens at the NYSE - Follow the besties:.
    • This piece from the All-In Podcast explains how four CEOs: CoreWeave, Perplexity, Mistral, and IREN, are shaping the broader artificial intelligence and infrastructure landscape.
    • EN Highlights:
      • (0:00) Intro live from Nvidia GTC (0:37) CoreWeave CEO, Michael Intrator (32:58) Perplexity CEO, Aravind Srinivas (1:07:11) Mistral CEO,…
      • It all happens at the NYSE - https://nyse.com Follow the besties: https://x.com/ch…

  • How Matt Mahan Thinks He Can Save California

    • Published: 2026-03-23 09:03 Beijing Time
    • Summary: - (0:00) Matt Mahan: Why He’s Running for Governor (1:51) How California Went From Bad to Worse (12:05) Public Sect….
      • lang=en”>.
      • lang=en <a href=".
    • EN Highlights:

Lenny’s Podcast (A_full) Link to heading

  • 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 focused on building developer tools and payment infrastructure.
      • He was part of the team that created “minions,” Stripe’s internal AI coding agents that now ship approximately 1,300 pull requests weekly with very little human intervention aside from code reviews.
      • In this episode, Steve demonstrates how Stripe engineers activate development work from Slack, utilize cloud-based development environments for parallel agent workflows, and showcases machine-to-machine payments where AI agents autonomously transact 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 weekly 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 non-engineers at Stripe are starting to use Minions to ship 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 intro ( 02:39 ) Stripe’s Minions and their impact across Stripe ( 04:42 ) Why activation energy is more important than execution ( 05:44 ) What are minions?

  • Technical Architecture (06:52) Demo: Activating Minions in Slack with Emojis (09:04) Why Good Developer Experience Benefits Both Humans and Agents (11:22) Browsing Agent Loops 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 Delivered PR (22:04) How Stripe Handles Code Reviews for 1,300 AI-Written PRs Per Week (23:44) Non-Engineers Using Minions Across the Company (24:53) Demo: Curating a Birthday Party with Claude Code 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 Referenced 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 Key Takeaways:
      • 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…
      • i=1000757255000”>Apple Podcasts What you’ll learn:

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

  • State of the product job market in early 2026

    • Published: 2026-03-24 20:45 Beijing Time
    • Summary: - Each week, I answer reader questions about building products, driving growth, and accelerating careers.
      • Welcome to our biannual product job market report—our fourth, and surprisingly, our most optimistic yet.
      • Despite headlines about layoffs and AI taking jobs, we’re actually seeing many promising signs in tech hiring, as well as some interesting new trends: - Product Manager job openings are at their highest in over three years - AI isn’t slowing demand for software engineers (at least not yet) - AI job postings are exploding overall - Design job postings have stabilized - The increasing importance of the Bay Area - Remote work opportunities continue to decline - Despite ongoing layoffs, the total number of tech jobs continues to grow. While these numbers are promising, I know many people are struggling to find work right now.
      • More job openings don’t mean people find jobs faster.
      • For anyone experiencing this, first, I’m sorry.
    • EN Key Takeaways:
      • 👋 Hey there, I’m Lenny
      • Each week, I answer reader questions about building product, driving growth, and accelerating your career
      • For more: Lenny’s Podcast | Lennybot | <a href=“https://www…
      • via=lenny”>public speaking course , and interview prep copilot <a class…

  • State of the product job market in early 2026

    • Published: 2026-03-24 17:02 Beijing Time
    • Summary: - <a class=“image-link image2 is-viewable-img” href=".
      • ,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7fd46e1f-918c-491c-a014-9a564084d36f_3016x3016.p….
      • ,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7fd46e1f-918c-491c-a014-9a564084d….
      • Today’s episode is our fourth biannual update on the tech job market (the most optimistic yet) and breaks down seven key trends shaping the job landscape…
    • EN Key Points:
      • <a class=“image-link image2 is-viewable-img” href=“https://substackcdn.com/image/fetch/$s_
      • ,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7fd46e1f-918c-491c-a014-9a564084d36f_3016x3016.p…
      • ,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7fd46e1f-918c-491c-a014-9a564084d…
      • Today’s episode is our fourth biannual update on the tech job market (the most optimistic yet) and breaks down seven key trends shaping the job landscape…

  • 🎙️ This week on How I AI: How Microsoft’s AI VP automates everything with Warp

    • Published: 2026-03-23 23:02 Beijing Time
    • Summary: - How Microsoft’s AI VP Automates Everything with Warp | Brought to you by Marco Casalaina: - Rovo — The AI that understands your business - Lovable — Build apps by simply chatting with an AI. Marco Casalaina, VP of Core AI Products at Microsoft, shows how he uses Warp, Microsoft 365 Copilot, and ChatGPT to dynamically launch lightweight “micro-agents” that handle everything from Azure management to document scanning to video compression.
      • Listen now on YouTube • Spotify • Apple Podcasts. Biggest takeaways: - Automation enables you to focus on higher-value activities.
      • As Marco puts it: “While these agents are doing what I need them to do, I’m doing something else.” While scanning his daughter’s homework, he was able to help her with her math problems instead of fighting with the scanning software.
        • Hardware automation is an underappreciated AI use case.
      • Marco demonstrates how Warp controls his scanner via the scanner CLI tool NAPS2.
    • 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…

  • ,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3d3e328b-a17a-4acf-b36b-49acc28915f1_1600x156.pn…

  • How Microsoft’s AI VP automates everything with Warp | Marco Casalaina

    • Publication Time: 2026-03-23 20:01 Beijing Time
    • Abstract: - Marco Casalaina, Microsoft’s VP of Core AI Products and an AI futurist, demonstrates how to use AI tools to automate administrative tasks that typically consume valuable time.
      • Instead of using Warp as a coding assistant (its primary marketing purpose), Marco utilizes it to manage Azure resources, scan documents, compress videos, and more.
      • He shows how these “micro-agents” can reduce friction in everyday workflows, allowing him to focus on higher-value activities.
      • Marco also demonstrates how Microsoft 365 Copilot and ChatGPT can create triggered workflows to respond to emails or check schedule information, emphasizing how the line between consuming and building AI agents is becoming blurred.
      • Listen or watch on YouTube, Spotify, or Apple Podcasts What you will learn: - How to use Warp to manage Azure resources and assign permissions without navigating complex web interfaces - Techniques for automatically scanning and processing documents directly from the terminal - Methods for analyzing and compressing video files using AI-generated FFmpeg commands - How to create simple rules to significantly improve AI performance for specialized tasks - Methods for building triggered workflows in Microsoft 365 Copilot to automatically reply to emails - How to configure ChatGPT to perform scheduled tasks, such as checking for new content - Strategies for creating consistent AI interactions using AutoHotkey shortcuts Provided by: Rovo — The AI that understands your business Lovely — Build applications by simply chatting with an AI In this episode, we will cover: ( 00:00 ) Introduction to Marco Casalaina ( 02:14 ) Why Marco chooses Warp for administrative tasks ( 03:57 ) Demo: Managing Azure resources and permissions with Warp (06:00) How CLI tools eliminate GUI friction for complex tasks (07:18) Creating rules to improve AI performance for specialized tasks (10:28) Demo: Document scanning automation (13:00) Automating the merging of odd and even pages with Python (15:04) The value of temporary AI solutions vs. permanent AI solutions
    • EN Highlights:
      • Marco Casalaina , VP of Core AI Products and AI Fu…
      • Rather than using Warp as a coding assistant (its primary marketed purpose), Marco leverages it to manage Azure resources, scan documents, compress videos, and…
      • He shows how these “micro-agents” can reduce friction in everyday workflows, allowing him to focus on higher-value activities

  • Marco also demonstrates how Microsoft 365 Copilot and ChatGPT can create triggered workflows that respond to emails or check for information on a schedule, high…

  • The art of influence: The single most important skill that AI can’t replace | Jessica Fain (Webflow, ex-Slack)

    • Posted: 2026-03-22 20:31 Beijing Time
    • Summary: - Jessica Fain is a product leader at Webflow and former Chief of Staff to the CPO at Slack, where she worked with April Underwood and many past podcast guests including Stewart Butterfield, Annie Pearl, Tamar Yehoshua, and Noah Weiss.
      • Throughout her career, she has learned how executives actually make decisions and why most people completely misunderstand the process.
      • Why executive calendars are “like strobe lights” and why the first 30 seconds of a meeting are so important 3.
      • Why executives typically optimize for global maxima, while you are usually optimizing locally 4.
      • When a leader says something that seems wrong, the best question Jessica uses is: “That’s so interesting.
    • EN Highlights:
      • Jessica Fain is a product leader at Webflow and former Chief of Staff to the CPO at Slack, where she worked alongside April Underwood and ma…
      • She’s spent her career learning how executives actually make decisions—and why most people completely misunderstand the process.

We discuss… - Why great ideas often don’t get buy-in 2 - Why executive calendars are “like strobe lights” and why the first 30 seconds of a meeting matter so much 3

Stratechery by Ben Thompson （A_full） Link to heading

  • 2026.13: So Long to Sora

    • Posted: 2026-03-28 01:00 Beijing Time
    • Summary: - Welcome back to This Week in Stratechery!
      • As a reminder, every week on Friday, we send out an overview of the content in the Stratechery bundle; highlighted links are free for everyone.
      • In addition, you have full control over what we send you.
      • With that said, here are a few of our favorites from this week.
      • Sora, 2025-2026. ** AI Sam came, AI Sam saw, AI Sam stole those GPUs.
    • EN Highlights:
      • <img alt=”” class=“wp-image-18637” height=“956” src=“https://i0.wp.com/stratechery.com/wp-content/uploads/20…
      • resize=1700%2C956&ssl=1” width=“1700” />
      • 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

    • Posted: 2026-03-26 18:06 Beijing Time

  • Summary: - In this interview, we delve deep into this decision, including the keynote’s metadata, Arm’s history, and how the company has evolved, especially under Haas’s leadership.

    • Then we explore 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 adapt to support this new direction.
    • As a reminder, all Stratechery content (including interviews) is available in podcast format; click the link at the top of this email to add Stratechery to your podcast player.

    • An Interview with Arm CEO Rene Haas About Selling Chips. Link to heading

    • 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 significant change from Arm’s history, but it’s not surprising given the evolution of computing.
      • Arm is selling its own chips, not just licensing IP. This is a big change compared to Arm’s history, but not surprising given how computing is evolving.
      • Arm Launches Its Own CPU, Arm’s Motivation, 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.

  • Spring Break

    • Publication Time: 2026-03-23 18:00 Beijing Time
    • Summary: - Stratechery’s spring break is a bit disjointed, as my usual week off 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. I will return to my normal publishing schedule on Tuesday, March 31.
      • All other Stratechery Plus content, including my podcast, will proceed as scheduled.
      • Stratechery’s spring break is a bit disjointed, as my usual week off will be spread out: I will return to my usual posting schedule on Tuesday, March 31.
      • Stratechery’s spring break is a bit disjointed, as my usual week off 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 during the spring break.
    • 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

OpenAI Blog (A_full) Link to heading

  • 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 company with over 230 years of history that specializes in automated waste sorting plants for the global recycling industry.
      • With over 650 employees worldwide, the company plays a crucial role in helping countries advance their sustainability and circular economy goals.
      • Under the leadership of co-CEO Julia Stadler, the company has taken a forward-thinking approach to modernization—embedding artificial intelligence as a core productivity layer into daily operations.

  • Since 2023, STADLER has followed a clear principle: every employee who works on a computer should use AI to improve speed, quality, and collaboration.

  • “In many teams, people spend too much time turning raw knowledge into usable output—summarizing, translating, drafting.

  • 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: - At OpenAI, we believe that AI should be fair, safe, and free, so that more people can use it to solve difficult problems, create opportunities, and benefit in areas like health, science, education, work, and daily life.
      • We believe that democratizing access to AI is the best path forward: instead of concentrating the benefits or control of AI in the hands of a few, we want to make it accessible for more people to understand and help shape AI.
      • This is the core reason for the existence of the OpenAI Model Spec.
      • It defines how we want models to follow instructions, resolve conflicts, respect user freedom, and operate safely within the extremely broad range of queries users make every day.
      • More broadly, we are trying to make expected model behavior explicit: not just during our training process, but in a form that users, developers, researchers, policymakers, and the wider public can actually read, scrutinize, and debate.
    • 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 Safety Bug Bounty program to identify AI abuse and safety risks, including agentic vulnerabilities, prompt injection, and data exfiltration.
      • This article on the OpenAI blog explains how the introduction of the OpenAI Safety Bug Bounty program is shaping the broader AI and infrastructure landscape.
      • Following the launch of the OpenAI Safety Bug Bounty program, it also brings 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.

  • Helping developers build safer AI experiences for teens

    • Publication Time: 2026-03-24 19:00 Beijing Time
    • Summary: - OpenAI is releasing prompt-based teen safety policies for developers using gpt-oss-safeguard to help mitigate age-specific risks in AI systems.
      • This article on the OpenAI blog explains how helping developers build safer AI experiences for teens is shaping the broader AI and infrastructure landscape.
      • Following “Helping developers build safer AI experiences for teens,” it also brings practical implications for founders, operators, and investors.
    • EN Highlights:
      • OpenAI releases prompt-based teen safety policies for developers using gpt-oss-safeguard, helping moderate age-specific risks in AI systems.

  • Powering product discovery in ChatGPT

    • Publication Time: 2026-03-24 17:00 Beijing Time
    • Summary: - ChatGPT is introducing richer, visually immersive shopping powered by the Agentic Commerce Protocol, supporting product discovery, side-by-side comparisons, and merchandise….

  • This article from the OpenAI blog explains how driving product discovery in ChatGPT is shaping the broader AI and infrastructure landscape.

  • Following the support for product discovery in ChatGPT, it also has practical implications for founders, operators, and investors.

    • EN Key Points:
      • ChatGPT introduces richer, visually immersive shopping powered by the Agentic Commerce Protocol, enabling product discovery, side-by-side comparisons, and merch…

  • Update on the OpenAI Foundation

    • Published: 2026-03-24 17:00 Beijing Time
    • Summary: - The OpenAI Foundation announced plans to invest at least $1 billion in treating diseases, economic opportunity, AI resilience, and community programs.
      • This article from the OpenAI blog explains how the update on the OpenAI Foundation is shaping the broader AI and infrastructure landscape.
      • Following the update on the OpenAI Foundation, it also has practical implications for founders, operators, and investors.
    • EN Key Points:
      • The OpenAI Foundation announces plans to invest at least $1 billion in curing diseases, economic opportunity, AI resilience, and community programs.

  • Creating with Sora Safely

    • Published: 2026-03-23 08:00 Beijing Time
    • Summary: [Translation pending] - To address the novel safety challenges posed by a state-of-the-art video model as well as a new social creation platform, we’ve built Sora 2 and the Sora app with safety at the foundation.
      • Our approach is anchored in concrete protections.
      • To address the novel safety challenges posed by a state-of-the-art video model as well as a new social creation platform, we’ve built Sora 2 and the Sora app wi… Our approach is anchored in concrete protections.
    • EN Key Points:
      • To address the novel safety challenges posed by a state-of-the-art video model as well as a new social creation platform, we’ve built Sora 2 and the Sora app wi…
      • Our approach is anchored in concrete protections.

Google DeepMind Blog (A_full) Link to heading

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

    • Published: 2026-03-26 23:23 Beijing Time
    • Summary: - Our latest speech 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 is shaping the broader AI and infrastructure landscape.
      • It also has practical implications for 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 is researching the risks of harmful manipulation by artificial intelligence in fields like finance and health, leading to the development of new safety measures.
      • This article from the Google DeepMind blog explains how protecting people from harmful manipulation is shaping the broader landscape of AI and infrastructure.
      • Beyond protecting people from harmful manipulation, it also has practical implications for founders, operators, and investors.
    • EN Highlights:
      • 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 can unlock longer tracks through structural awareness.
      • We are also introducing Lyria to more Google products and interfaces.
      • Introducing Lyria 3 Pro, which can unlock longer tracks through structural awareness. We are also introducing Lyria to more Google products and interfaces.
    • EN Highlights:
      • 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

Lex Fridman (B_intro+search) Link to heading