System translated (Gemini)

AI Daily | Agentic Thinking Replaces the Reasoning Paradigm, Claude Code Enters Autonomous Mode Link to heading

Today, the AI industry is showing a trend of shifting from “reasoning” to “agentic thinking.” Claude Code’s launch of an autonomous mode marks a shift in AI programming from assisted autocompletion to autonomous execution; the OpenClaw ecosystem is accelerating its expansion through WeChat integration and a plugin marketplace. On the technical front, Luma and Google have each released new multi-modal models. Industry analysis warns that as the cost of AI code replication decreases, test cases are becoming the new technical barrier.

I. Today’s Highlights Link to heading

  • Watch List: “The Algorithm That Made Me Cry” from Two Minute Papers is worth prioritizing. Key points: - ❤️ Check out Lambda and sign up for their GPU Cloud here:.
  • Free ray tracing course:.
  • If you would like to support us, our Patreon:.
  • Watch List: A total of 26 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 “Andrew Chen Says AI Text and Photos Cross Uncanny Valley” is trending (approx. 108 related posts) and serves as an entry point for understanding global AI public opinion and viewpoints.

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

Topic 1: Andrew Chen Says AI Text and Photos Cross Uncanny Valley Link to heading

  • Category: AI · Other
  • Overview: Trending Time: , Related Posts: 108
  • What it is: a16z partner Andrew Chen posted that AI-generated text and photos have successfully crossed the “uncanny valley,” reaching a level indistinguishable from human creation.
  • Why it matters: This signifies that the realism of generative AI has moved from “unsettling” to “natural,” heralding the large-scale entry of AI content into the mainstream consumer market and transforming the human-computer interaction experience.
  • Discussion overview: The discussion centers on whether AI has truly eliminated the uncanny feeling, and the profound implications of this high fidelity for information authenticity, copyright definition, and the viability of human creators.

Topic 2: OpenClaw Creator Teams Up with Telegram for AI Agent Boost Link to heading

  • Category: AI · News
  • Overview: Trending Time: 3 hours ago, Related Posts: 178
  • What it is: The developers of OpenClaw announced a partnership with Telegram to enhance AI Agent features and user experience on the platform.
  • Why it matters: This move deeply integrates AI agents with a social platform boasting a massive user base and a mature bot ecosystem, helping to accelerate the adoption of AI applications and the growth of decentralized interactions.
  • Discussion overview: The discussion focuses on whether Telegram will become the primary entry point for AI agents and how the open-source nature of OpenClaw will enhance developer efficiency in its ecosystem.

Topic 3: Iran Fires Missiles at Israel Amid Ceasefire Push Link to heading

  • Category: AI · News
  • Overview: Trending Time: 1 day ago, Related Posts: 311,000
  • What it is: Iran launched a large-scale missile attack against Israel, leading to a dramatic escalation of geopolitical tensions in the Middle East.
  • Why it matters: The conflict not only directly affects the stability of the global technology supply chain and the AI chip market but also serves as a real-world test for AI-driven automated defense systems and intelligence analysis technology.
  • Discussion overview: Discussions on X focus on the technical performance of interception systems (like the Arrow system), the conflict’s impact on tech stocks like Nvidia, and the proliferation of AI-generated fake war imagery on the platform.

Topic 4: OpenAI Engineer Reassures Codex App’s Strong Future Amid Claude Code Rivalry Link to heading

  • Category: AI · News
  • Overview: Trending Time: 19 hours ago, Related Posts: 2,100
  • What it is: In response to competitive pressure from the release of Anthropic’s Claude Code, an OpenAI engineer publicly reaffirmed confidence in the future of Codex applications and its ecosystem.
  • Why it matters: This reflects the intensifying competition in the AI coding tools (AI Coding Agents) space, indicating that developer workflows are evolving from simple code completion towards deeply automated native tools.
  • Discussion overview: Discussions on X focus on whether Claude Code’s command-line integration advantage will disrupt the current IDE plugin model, and whether OpenAI is secretly preparing a more powerful, specialized coding model in response.

Topic 5: Developers Share Agent-Building Guides and Evals in 2026 Push Link to heading

  • Category: AI · News
  • Overview: Hotness: 3 hours ago, Related Posts: 369
  • What happened: The developer community has begun intensively sharing AI Agent construction guides and evaluation frameworks, aiming to prepare for the technological explosion in 2026.
  • Why it’s important: This signifies that AI applications are shifting from simple conversational interactions to complex autonomous task execution. Standardized evaluation systems are key for Agents to move towards industrial-scale production.
  • Discussion overview: Core discussions revolve around how to define effective Agent evaluation metrics (Evals), and how to balance system controllability and security while pursuing high autonomy.

Topic 6: Runable Ditches Bloated MCP for Simple CLI in AI Agents Link to heading

  • Category: AI · News
  • Overview: Hotness: 9 hours ago, Related Posts: 425
  • What happened: Runable announced it is abandoning Anthropic’s MCP (Model Context Protocol) in favor of a simpler CLI (Command Line Interface) mode for building AI Agents.
  • Why it’s important: This move challenges the current mainstream trend of standardizing AI tool integration, sparking deep thought about the trade-off between protocol complexity and practicality in AI Agent development.
  • Discussion overview: Community discussions focus on whether MCP suffers from over-engineering (Bloated) and if, in the pursuit of standardization, simpler CLI solutions better meet developers’ actual needs for efficiency and flexibility.

Topic 7: Developers Turn Claude Code into Autonomous Coding Partner Link to heading

  • Category: AI · News
  • Overview: Hotness: 22 hours ago, Related Posts: 319
  • What happened: Developers are transforming Anthropic’s command-line tool, Claude Code, into an autonomous coding partner capable of independently performing tasks such as writing, testing, and fixing code.
  • Why it’s important: This marks the evolution of AI programming tools from simple code completion to “AI Agents” with autonomous decision-making capabilities, significantly raising the ceiling for software development automation.
  • Discussion overview: Discussions center on the reliability of Claude Code in handling complex engineering tasks, its potential to replace human programmers, and its security and permission control when running in a terminal environment.

Topic 8: Luka Dončić’s 43 Points Lead Lakers Past Pacers in High-Scoring Thriller Link to heading

  • Category: AI · Other
  • Overview: Hotness: 1 day ago, Related Posts: 124000
  • What happened: X’s AI assistant Grok generated a false news summary about NBA star Luka Dončić leading the Lakers to victory over the Pacers, attracting significant user attention.
  • Why it’s important: This incident highlights the “hallucination” problem that AI can encounter when processing real-time news, as well as the challenges to accuracy and reliability of automatically generated content on social media platforms.
  • Discussion overview: Discussions focus on Grok’s factual error (Dončić is not a Lakers player), the risk of AI spreading misinformation, and the current lack of effective moderation of AI-generated content by platforms.

Topic 9: Rodri Leaves Door Open for Real Madrid Transfer Amid City Contract Talks Link to heading

  • Category: AI · Other
  • Overview: Hotness: 23 hours ago, Related Posts: 117000
  • What happened: Manchester City’s midfield maestro Rodri is open to a future transfer to Real Madrid amid contract extension talks with his club.
  • Why it’s important: Although this topic was categorized by the system as being in the AI domain, it is in fact top sports news. Its popularity reflects the classification logic and data weighting of social media algorithms when handling non-technical breaking hot topics.
  • Discussion overview: Discussions focus on the possibility of Rodri returning to Spain, the tactical impact on Manchester City if they lose their key player, and Real Madrid’s attractiveness in the transfer market.

Topic 10: Suno Launches Expressive v5.5 AI Music Generator Link to heading

  • Category: AI · News
  • Overview: Hotness: 3 hours ago, Related Posts: 612
  • What happened: AI music creation platform Suno released version v5.5, with a focus on enhancing the emotional expression and detail in generated music.
  • Why it’s important: This update signifies the evolution of AI music from mere melody generation to “humanized” creation with nuanced emotions, further narrowing the performance gap between AI and human singers.
  • Discussion overview: Discussions focus on the astonishing realism of vocal performances, the potential impact on professional musicians, and a comparison with competitor Udio in terms of sound quality and creative control.

Summary of AI Public Opinion on X Today Link to heading

The core narrative in the current AI discourse is the technology’s comprehensive leap from simple interactive tools to highly autonomous “Agents.” In the fields of audio-visual creation and programming, the sense of artificiality has been largely eliminated, leading to a consensus that a qualitative change in “personification” has been achieved. Although the industry is optimistic about the prospects of deeply embedding AI into social ecosystems and developer workflows, there remains a significant divergence within the developer community on the technical implementation path: whether to adopt standardized complex protocols or return to a simple and efficient command-line model. Meanwhile, public opinion is also highly vigilant about the proliferation of false information caused by real-time AI hallucinations, potential threats to the hardware supply chain from geopolitical instability, and the security and controllability risks of Agents after gaining higher system permissions.

III. Influencer Insights Link to heading

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

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

The most central discussion today revolves around the next stage of large model capabilities. @dotey provided an in-depth analysis of the viewpoint from former Qwen technical lead Lin Junyang: AI is shifting from “thinking longer (Reasoning)” to “thinking for action (Agentic).”

  • Core Logic: Reasoning models (like o1/R1) focus on the correctness of the answer, whereas agentic thinking focuses on closed-loop interaction with the environment.
  • Product Manifestation: Claude Code has launched the new auto mode, allowing the AI to autonomously decide on file writing and command execution permissions without needing user approval for each action. Both @op7418 and @zhixianio believe this marks a shift for AI from a “must-ask-for-everything” assistant to an “intelligent subordinate.”

The Explosion of the “Little Lobster (OpenClaw)” Ecosystem and WeChat’s Entry Link to heading

OpenClaw (affectionately called “Little Lobster” by the community) continues to dominate the charts.

  • WeChat’s Move: WeChat has officially started grayscale testing an OpenClaw plugin that supports connections in private chats. @zhixianio points out that this entry point will crush other domestic IMs.
  • Ecosystem Enhancement: OpenClaw released v3.24, launching the ClawHub plugin marketplace and ensuring compatibility with the plugin systems of Codex and Cursor (@op7418).
  • Remote Control: The feature for remotely controlling Claude Code via Telegram/Discord (Claude Code channels) has been officially released, enabling “remote programming from a phone” (@zhixianio).

A New Round of “Muscle Flexing” from Multimodal Models Link to heading

  • Luma Uni-1: Luma released Uni-1, a drawing model based on an autoregressive Transformer architecture. It possesses extremely strong reasoning abilities and Chinese character writing skills, with performance considered to be on par with top-tier models (@Gorden_Sun).
  • Google Lyria 3 Pro: Google officially released its music model, which supports generating 3-minute songs with vocals and has an open API (@Gorden_Sun).
  • Open-Source Audio/Video: GAIR Labs has open-sourced the 15B-parameter daVinci-MagiHuman, which supports synchronized audio and video generation in multiple languages including Chinese, English, and Japanese (@vista8).

2. Unique Perspectives and Industry Foresight Link to heading

Observations on the “Asymmetry” of China’s AI Ecosystem Link to heading

@dotey translated a Western VC’s in-depth assessment of China’s AI ecosystem:

  • Hardware Barriers: Shenzhen’s hardware reverse engineering and supply chain advantages are difficult for the West to match.
  • Founder Traits: Chinese founders are often “top-tier executors who solve known problems,” lacking a “maverick” spirit of originality. However, Liang Wenfeng of DeepSeek is seen as an exception who breaks the mold.
  • Software Bubble: Despite top-tier talent, there is a clear bubble in domestic software valuations, and they face the threat of major corporations making their features native.

Test Cases: The New Moat in the AI Era Link to heading

@ruanyf issued a warning: When AI can replicate large software like Next.js at an extremely low cost (e.g., $1100), the code itself no longer serves as a moat. “Test cases” will become the key barrier to prevent replication by AI.

Foundation Models’ “Self-Awareness” and the Side Effects of RLHF Link to heading

@vista8 shared a paper from Apple: Foundation models naturally possess the ability to assess the confidence of their answers (semantic calibration), but Reinforcement Learning from Human Feedback (RLHF) and Chain of Thought (CoT) actually undermine this accuracy, causing the model to become “blindly confident” to please humans.

The Reversal of “Data Sovereignty” in the Age of Agents Link to heading

@lijigang predicts that the age of AI may bring about a reversal of data sovereignty. Users will achieve an equal footing with service companies through local, private data isolation, rather than having all data uploaded to the cloud as in the internet era.

Developer Tools Link to heading

  • Claude Code Stable Version: @Pluvio9yte noted that if you notice abnormal Token consumption in v2.1.8x, it’s recommended to roll back to v2.1.77.
  • Dynamic Workers: An Agent sandbox solution from Cloudflare that starts 100 times faster than containers (@Gorden_Sun).
  • Type4Me: A macOS voice input tool with dual local/cloud engines and text optimization via large models (@dotey).

Prompts & Skills Library Link to heading

  • Official OpenAI Prompt Pack: An official template library for professions like HR, engineers, and executives (@Pluvio9yte).
  • GEO Prompts: 27 sets of Generative Engine Optimization (GEO) prompts open-sourced by Teacher Yao (@vista8).
  • Suno CLI Plugin: An open-source plugin by @vista8 that generates Suno songs in specific styles by searching for information.

Learning Resources Link to heading

  • Claude Code Cheat Sheet: A cheat sheet for quickly mastering common commands (@Gorden_Sun).
  • 8 Levels of AI Programming: Steve Yegge’s proposed proficiency levels for AI programming, useful for self-assessment (@ruanyf).

Analyst’s Brief: Today’s developments show that the AI industry is transitioning from a “dialog box” to an “operating system.” Whether it’s the permission automation in Claude Code or OpenClaw’s penetration into the IM space, it’s clear that Agents are beginning to take over real-world workflows. For practitioners, the focus should shift from “how to write a good prompt” to “how to build environments that support continuous Agent action (Harness Engineering).”

IV. Watch List Update Summary Link to heading

Timeframe: Last 7 days; Covers 16 sources; 26 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 an investor in Anysphere, to discuss Collison’s history with Smalltalk and Lisp, the MongoDB and Ruby decisions Stripe still stands by 15 years later, why he would have spent more time on API design if he could do it over, and whether AI is actually showing up in economic productivity data.
      • This episode originally aired on Cursor’s podcast.
      • Follow Patrick Collison on X: Follow Michael Truell on X: Follow Cursor: See everything a16z is doing with AI here, including articles, projects, and more podcasts.
      • Please note that the content here is for informational purposes only; it should not be taken as legal, business

  • Summary: - The Pod by 8 Sleep can cool your bed to 55°F and uses Autopilot to optimize your sleep all night.

    • Follow Bryan: Follow the besties: 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 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-1, abundance, and human optimization (35:35) The longevity drug no one is talking about?
    • Follow Bryan: Follow the besties: 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 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: <a href=“https…

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

    • Published: 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 from CoreWeave, Perplexity, Mistral, and IREN are shaping the broader AI 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 sector….
      • 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)

    • Release Time: 2026-03-25 20:03 Beijing Time
    • Abstract: - 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 send approximately 1,300 pull requests per week with minimal human intervention, aside from code review.
      • In this episode, Steve demonstrates how Stripe engineers activate development work from Slack, utilize cloud-based development environments to enable 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 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) Introduction to Steve (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 beneficial for both humans and agents (11:22) Exploring the agent loop and system prompts (13:42) Why Stripe chose Goose as its agent tool (16:00) The role of the Developer Productivity team at Stripe (17:15) Why cloud environments unlock multithreaded AI engineering (21:14) One-shot prompt: From Slack to a 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: 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 Key Points:
      • 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

    • Publication Time: 2026-03-24 20:45 Beijing Time
    • Summary: - Every week, I answer questions from readers about building products, driving growth, and accelerating their careers.
      • Welcome to our biennial update on the state of the product job market—this is our fourth, and surprisingly, the most optimistic one yet.
      • Despite headlines about layoffs and AI taking jobs, we’re actually seeing many promising signs in tech hiring, along with some interesting new trends: - Product manager job openings are at their highest level in over three years - AI has not slowed down the demand for software engineers (at least not yet) - AI jobs in general are exploding - Design jobs have stabilized - The Bay Area’s importance is growing - Remote work opportunities continue to decrease - The total number of tech jobs continues to grow despite the constant layoffs. While these numbers are promising, I know many people are having a tough time finding a job right now.
      • More job openings don’t necessarily mean people are finding jobs faster.
      • To anyone in this situation, first of all, I’m sorry.
    • EN Highlights:
      • 👋 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

    • Publication Time: 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 show is our fourth biannual update on the tech job market (the most optimistic one to date), with a detailed analysis of the seven key trends shaping the employment landscape…
    • EN Highlights:

  • ,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 applications by simply chatting with AI. Marco Casalaina, Vice President of Microsoft’s Core AI Products, explains how to use Warp, Microsoft 365 Copilot, and ChatGPT to dynamically launch lightweight “micro-agents” that handle everything from Azure management to document scanning and video compression.
      • Listen now on YouTube • Spotify • Apple Podcasts. Key takeaways: - Automation allows you to focus on higher-value activities.
      • As Marco says: “While these agents are doing what I need them to do, I’m doing other things.” When scanning his daughter’s homework, he was able to help her with her math problems instead of struggling with the scanning software.
        • Hardware automation is an underappreciated AI use case.
      • Marco demonstrates how Warp can control his scanner via the scanner CLI tool NAPS2.
    • EN Key Points:
      • <a class=“image-link image2” 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%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

    • Published: 2026-03-23 20:01 Beijing Time
    • Summary: - Marco Casalaina, Vice President of Microsoft’s Core AI Products and an AI futurist, demonstrates how to use AI tools to automate administrative tasks that would typically consume valuable time.
      • Instead of using Warp as a coding assistant (its primary marketed 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, highlighting 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 using AI-generated FFmpeg commands to analyze and compress video files - How to create simple rules that 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. Brought to you 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 chose 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) Using Python to automate merging odd and even pages (15:04) The value of ad-hoc AI solutions vs. purpose-built ones.

    • EN 要点：
      • 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)

    • Published: 2026-03-22 20:31 Beijing Time
    • Summary: - Jessica Fain is a product lead at Webflow and former Chief of Staff to the Chief Product Officer 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 a flashbulb” and why the first 30 seconds of a meeting are so critical.
      • Why executives typically optimize for the global maximum, while you are likely optimizing for a local one.
      • When a leader says something that seems wrong, the best question Jessica uses is: “That’s so interesting.
    • EN 要点：

  • 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

  • 🧠 Community Wisdom: Beating the solo-founder procrastination trap, best tools for B2B prospecting, finding your first real users pre-launch, and more

    • Publication Time: 2026-03-22 01:19 Beijing Time
    • Summary: - 👋 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 Key Points:
      • 👋 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…

Stratechery by Ben Thompson （A_full） Link to heading

  • An Interview with Arm CEO Rene Haas About Selling Chips

    • Publication Time: 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 get into why CPUs matter for AI, and how Arm’s CPU compares to Nvidia’s, x86, and other custom Arm silicon.
      • 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 in podcast form; 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 Key Points:
      • Listen to this post:

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

    • Published: 2026-03-25 18:00 Beijing Time
    • Summary: - 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.
      • $15/month* or *$150/year.
      • Get substantive analysis of the day’s news via three weekly emails or podcasts.
      • Strategy Interviews.
    • EN Key Points:
      • 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

    • Published: 2026-03-23 18:00 Beijing Time
    • Summary: [To be translated] - 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 - There will be no Update on Monday, March 30 I will return to my usual posting schedule on Tuesday, March 31.
      • All other Stratechery Plus content, including my podcasts, will stay on schedule.
      • Stratechery is on a bit of a disjointed Spring Break, as my usual week off will be spread out: I will return to my usual posting schedule on Tuesday, March 31.
      • 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 Spring Break.
    • EN Key Points:
      • 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

  • Inside our approach to the Model Spec

    • Published: 2026-03-25 18:00 Beijing time
    • Summary: - At OpenAI, we believe that AI should be fair, safe, and accessible, so that more people can use it to solve tough 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: not concentrating the benefits or control of AI in the hands of a few, but making it accessible so that more people can access, understand, and help shape AI.
      • This is the core reason for the OpenAI Model Spec’s existence.
      • It defines how we want our models to follow instructions, resolve conflicts, respect user freedom, and operate safely amidst the extremely broad range of queries users pose every day.
      • More broadly, we are trying to make expected model behavior explicit: not just in our training process, but also in a form that users, developers, researchers, policymakers, and the wider public can actually read, scrutinize, and debate.
    • EN Key Points:
      • 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

    • Published: 2026-03-25 08:00 Beijing time
    • Summary: - Testing OpenAI for security and abuse issues.
      • As AI technology rapidly advances, its potential for misuse also evolves.
      • Our goal is to ensure our systems remain secure and prevent misuse or abuse that could lead to tangible harm.
      • Through this program, we look forward to continuing our collaboration with security and safety researchers to help us identify and address issues that do not fall under traditional security vulnerabilities but still pose real risks.
      • Submissions will be triaged by OpenAI’s safety and security bug bounty team and may be rerouted between the two programs based on scope and ownership.
    • EN Key Points:
      • 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

    • Published: 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 from 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 has practical implications for founders, operators, and investors.
    • EN Key Points:
      • OpenAI releases prompt-based teen safety policies for developers using gpt-oss-safeguard, helping moderate age-specific risks in AI systems.

  • Update on the OpenAI Foundation

    • Published: 2026-03-24 17:00 Beijing time
    • Summary: - The OpenAI Foundation has announced plans to invest at least $1 billion in disease treatment, economic opportunity, AI resilience, and community projects.
      • This post from the OpenAI blog explains how the update on the OpenAI Foundation is shaping the broader AI and infrastructure landscape.

  • Following the OpenAI Foundation update, it also brought practical implications for founders, operators, and investors.

    • EN Highlights:
      • The OpenAI Foundation announces plans to invest at least $1 billion in curing diseases, economic opportunity, AI resilience, and community programs.

  • Powering product discovery in ChatGPT

    • Publication Time: 2026-03-24 17:00 Beijing Time
    • Summary: - ChatGPT introduces 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 powering product discovery in ChatGPT shapes the broader AI and infrastructure landscape.
      • Following the support for product discovery in ChatGPT, it also brought practical implications for founders, operators, and investors.
    • EN Highlights:
      • ChatGPT introduces richer, visually immersive shopping powered by the Agentic Commerce Protocol, enabling product discovery, side-by-side comparisons, and merch…

  • Creating with Sora Safely

    • Publication Time: 2026-03-23 08:00 Beijing Time
    • Summary: - To address the new safety challenges posed by state-of-the-art video models and new social creation platforms, we have built Sora 2 and the Sora application with a foundation of safety.
      • Our approach is grounded in concrete protective measures.
      • To address the new safety challenges posed by state-of-the-art video models and new social creation platforms, we have built Sora 2 and the Sora application… Our approach is grounded in concrete protective measures.
    • EN Highlights:
      • 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

    • Publication Time: 2026-03-26 23:23 Beijing Time
    • Summary: - Our latest voice model has improved precision and lower latency to make voice interactions more fluid, natural and 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 founders, operators, and investors regarding Gemini 3.1 Flash Live: Making audio AI more natural and reliable.
    • EN Highlights:
      • 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 AI in fields like finance and health to develop new safety measures.
      • This article from the Google DeepMind blog explains how protecting people from harmful manipulation shapes the broader AI and infrastructure landscape.

  • Beyond protecting people from harmful manipulation, it also brings tangible impact to founders, operators, and investors.

    • EN Key Takeaways:
      • 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

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