{"title":"AI Daily Update | Google Launches TurboQuant VRAM Compression Technology, Anthropic Drives Agent Engineering Transformation","url":"/en/ai-daily/ai-daily-2026-03-26/","date":"2026-03-26","type":"ai-daily","content":"\u003ch1 id=\"ai-daily--google-introduces-turboquant-for-vram-compression-anthropic-advances-agent-engineering-transformation\"\u003e\n  AI Daily | Google Introduces TurboQuant for VRAM Compression, Anthropic Advances Agent Engineering Transformation\n  \u003ca class=\"heading-link\" href=\"#ai-daily--google-introduces-turboquant-for-vram-compression-anthropic-advances-agent-engineering-transformation\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h1\u003e\n\u003cblockquote\u003e\n\u003cp\u003eToday, the AI industry is focused on breakthroughs in inference efficiency and the evolution of Agent architecture. Google has released TurboQuant, capable of compressing the KV cache to one-sixth its original size with no loss in precision, significantly lowering the deployment barriers for large models processing long contexts. Anthropic has introduced the \u0026ldquo;driving engineering\u0026rdquo; concept and an automatic mode for Claude Code, signaling the transformation of AI assistants from chatbots to autonomous, task-executing Agents. Furthermore, the LiteLLM supply chain poisoning incident reminds developers to be wary of security risks as the AI ecosystem expands.\u003c/p\u003e\n\u003c/blockquote\u003e\n\u003ch2 id=\"i-todays-key-highlights\"\u003e\n  I. Today\u0026rsquo;s Key Highlights\n  \u003ca class=\"heading-link\" href=\"#i-todays-key-highlights\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eWatch List:\u003c/strong\u003e An article from the Google DeepMind Blog, \u0026ldquo;Lyria 3 Pro: Create longer tracks in more,\u0026rdquo; deserves priority attention. Key points include: - Introducing Lyria 3 Pro, which unlocks longer tracks through structural awareness.\u003c/li\u003e\n\u003cli\u003eWe are also integrating Lyria into more Google products and interfaces.\u003c/li\u003e\n\u003cli\u003eIntroducing Lyria 3 Pro, which unlocks longer tracks through structural awareness We are also integrating Lyria into more\u0026hellip;.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eWatch List:\u003c/strong\u003e A total of 23 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.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eX Hot Topic:\u003c/strong\u003e The topic \u0026ldquo;Google\u0026rsquo;s TurboQuant Compresses AI Memory Sixfold with No Accuracy Loss\u0026rdquo; is trending (approximately 28,000 related posts) and serves as a good entry point for understanding global AI public opinion and viewpoints.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2 id=\"ii-ai-hot-topics-on-the-x-platform-based-on-bird\"\u003e\n  II. AI Hot Topics on the X Platform (based on bird)\n  \u003ca class=\"heading-link\" href=\"#ii-ai-hot-topics-on-the-x-platform-based-on-bird\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h2\u003e\n\u003ch3 id=\"topic-1-googles-turboquant-compresses-ai-memory-sixfold-with-no-accuracy-loss\"\u003e\n  Topic 1: Google\u0026rsquo;s TurboQuant Compresses AI Memory Sixfold with No Accuracy Loss\n  \u003ca class=\"heading-link\" href=\"#topic-1-googles-turboquant-compresses-ai-memory-sixfold-with-no-accuracy-loss\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003eCategory: AI · Other\u003c/li\u003e\n\u003cli\u003eOverview: Trending for: 1 day ago, Number of related posts: 28,000\u003c/li\u003e\n\u003cli\u003eWhat it is: Google has introduced TurboQuant technology, which can compress AI VRAM usage (KV cache) to one-sixth of its original size without any loss of accuracy.\u003c/li\u003e\n\u003cli\u003eWhy it\u0026rsquo;s important: This technology significantly lowers the hardware barrier for long-context inference in large models, enabling models that previously required a GPU cluster to be deployed on a single A100, thereby dramatically improving the cost-effectiveness of AI deployment.\u003c/li\u003e\n\u003cli\u003eDiscussion summary: Discussions are focused on the sharp decrease in deployment costs for long-context Agents in production environments, and how this technology translates AI research achievements into practical commercial infrastructure advantages.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"topic-2-oil-prices-plunge-over-45-as-qatar-resumes-operations-after-iran-strikes-pause\"\u003e\n  Topic 2: Oil Prices Plunge Over $45 as Qatar Resumes Operations After Iran Strikes Pause\n  \u003ca class=\"heading-link\" href=\"#topic-2-oil-prices-plunge-over-45-as-qatar-resumes-operations-after-iran-strikes-pause\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003eCategory: AI · News\u003c/li\u003e\n\u003cli\u003eOverview: Trending for: 11 hours ago, Number of related posts: 54,000\u003c/li\u003e\n\u003cli\u003eWhat it is: International oil prices have dropped significantly by over $45 as Qatar resumes operations following a pause in Iranian strikes.\u003c/li\u003e\n\u003cli\u003eWhy it\u0026rsquo;s important: Energy costs are a core expense for operating AI data centers and training large-scale models. Oil price volatility directly impacts the operational costs of tech giants and their long-term investment budgets for AI infrastructure.\u003c/li\u003e\n\u003cli\u003eDiscussion summary: Discussions on social media are focused on whether the drop in energy prices can alleviate the pressure on AI computing costs, and the potential impact of geopolitical situations on the stability of the global tech supply chain.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"topic-3-iran-charges-ships-up-to-2-million-for-strait-of-hormuz-passage-amid-war-disruptions\"\u003e\n  Topic 3: Iran Charges Ships Up to $2 Million for Strait of Hormuz Passage Amid War Disruptions\n  \u003ca class=\"heading-link\" href=\"#topic-3-iran-charges-ships-up-to-2-million-for-strait-of-hormuz-passage-amid-war-disruptions\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003eCategory: AI · News\u003c/li\u003e\n\u003cli\u003eOverview: Trending for: 2 days ago, Number of related posts: 141,000\u003c/li\u003e\n\u003cli\u003eWhat it is: Due to regional instability, Iran has begun charging vessels up to $2 million for passage through the Strait of Hormuz.\u003c/li\u003e\n\u003cli\u003eWhy it\u0026rsquo;s important: This strait is a critical node for global hardware logistics. The surge in passage fees could drive up supply chain costs for hardware such as semiconductors and AI servers, and also extend delivery times.\u003c/li\u003e\n\u003cli\u003eDiscussion summary: Discussions are focused on the impact of geopolitics on global trade, the fragility of the tech supply chain, and whether the rising shipping costs will ultimately be passed on to downstream AI companies and consumers.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"topic-4-us-airstrikes-continue-on-iran-as-troop-deployments-raise-ground-action-fears\"\u003e\n  Topic 4: U.S. Airstrikes Continue on Iran as Troop Deployments Raise Ground Action Fears\n  \u003ca class=\"heading-link\" href=\"#topic-4-us-airstrikes-continue-on-iran-as-troop-deployments-raise-ground-action-fears\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003eCategory: AI · Other\u003c/li\u003e\n\u003cli\u003eOverview: Trending for: 3 hours ago, Number of related posts: 93,000\u003c/li\u003e\n\u003cli\u003eWhat it is: The U.S. is continuing airstrikes on Iranian-related targets and deploying additional troops, raising serious concerns about the Middle East situation escalating into a ground war.\u003c/li\u003e\n\u003cli\u003eWhy it\u0026rsquo;s important: Such conflicts accelerate the practical application of AI technologies in modern electronic warfare, autonomous drone systems, and intelligence analysis. Simultaneously, geopolitical instability could impact the stability of the global AI hardware supply chain.\u003c/li\u003e\n\u003cli\u003eDiscussion Overview: The discussion focuses on whether AI-assisted decision-making will lower the threshold for war, and the potential ripple effects of military conflicts on the global semiconductor supply chain and energy costs.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"topic-5-iran-claims-fifth-us-f-35-downed-debunked-by-centcom\"\u003e\n  Topic 5: Iran Claims Fifth U.S. F-35 Downed, Debunked by CENTCOM\n  \u003ca class=\"heading-link\" href=\"#topic-5-iran-claims-fifth-us-f-35-downed-debunked-by-centcom\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003eCategory: AI · News\u003c/li\u003e\n\u003cli\u003eOverview: Trending for: 3 hours ago, Related posts: 50,000\u003c/li\u003e\n\u003cli\u003eWhat it is: Iran claimed to have shot down a fifth U.S. F-35 fighter jet, a claim that was subsequently officially refuted by the U.S. Central Command (CENTCOM) as disinformation.\u003c/li\u003e\n\u003cli\u003eWhy it matters: The incident highlights the severe challenges posed to real-time fact-checking and information security when AI-generated disinformation is combined with social media algorithms in modern conflicts.\u003c/li\u003e\n\u003cli\u003eDiscussion Overview: The discussion focuses on the authenticity of Iran\u0026rsquo;s propaganda methods, the viral spread of false military information on social media platforms, and the role of AI in identifying and curbing such misleading content.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch4 id=\"ai-public-opinion-summary-on-x-today\"\u003e\n  AI Public Opinion Summary on X Today\n  \u003ca class=\"heading-link\" href=\"#ai-public-opinion-summary-on-x-today\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cp\u003eToday\u0026rsquo;s main narrative reveals a sharp tug-of-war between optimistic expectations for cost reduction and efficiency gains from technology, and supply chain anxiety triggered by geopolitical turmoil. The industry consensus is that technological breakthroughs like Google\u0026rsquo;s TurboQuant and falling energy prices are significantly lowering the economic barriers to AI deployment, paving the way for the commercialization of long-context applications. However, increased shipping fees and logistical disruptions caused by escalating tensions in the Middle East have led to serious disagreements in the market regarding the stability and ultimate cost trajectory of the hardware supply chain. The potential risks not only point to the impact that geopolitical conflicts could have on the foundational layers of the AI industry through the energy and semiconductor chains, but also highlight the ethical and security challenges of AI as a \u0026ldquo;double-edged sword\u0026rdquo; in modern warfare and the spread of disinformation.\u003c/p\u003e\n\u003ch2 id=\"iii-influencer-insights\"\u003e\n  III. Influencer Insights\n  \u003ca class=\"heading-link\" href=\"#iii-influencer-insights\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h2\u003e\n\u003cp\u003eHello! I am your AI industry analyst. Based on the tweets from leading figures in the AI field on X over the past 24 hours, I have compiled today\u0026rsquo;s industry trends and in-depth insights for you.\u003c/p\u003e\n\u003chr\u003e\n\u003ch3 id=\"1-top-trends-and-product-news\"\u003e\n  1. Top Trends and Product News\n  \u003ca class=\"heading-link\" href=\"#1-top-trends-and-product-news\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003ch4 id=\"a-the\"\u003e\n  \u003cstrong\u003eA. The \u0026ldquo;Engineering\u0026rdquo; Shift in Agent Infrastructure: Harness Engineering\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#a-the\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cp\u003eToday\u0026rsquo;s most-watched technical discussion revolves around the research on \u003cstrong\u003eHarness Engineering\u003c/strong\u003e published by Anthropic.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eCore Definition\u003c/strong\u003e: @Gorden_Sun points out that a Harness, unlike traditional frameworks, covers task decomposition, inter-agent communication, context resetting, and the definition of evaluation criteria.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eProduct Implementation\u003c/strong\u003e: Claude Code introduces \u003cstrong\u003e\u0026ldquo;Auto Mode.\u0026rdquo;\u003c/strong\u003e @claudeai announced that this mode allows Claude to decide independently whether to execute file writes or Bash commands, striking a balance between \u0026ldquo;asking for permission for everything\u0026rdquo; and \u0026ldquo;complete autonomy.\u0026rdquo; @zhixianio commented that it\u0026rsquo;s more like a \u0026ldquo;smart subordinate reporting back.\u0026rdquo;\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eSandbox Security\u003c/strong\u003e: Cloudflare launched \u003cstrong\u003eDynamic Workers\u003c/strong\u003e, offering an agent sandboxing solution with millisecond startup times, 100 times faster than containers. It aims to solve the performance and security trade-off when agents execute dynamic code (@Gorden_Sun).\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch4 id=\"b-the\"\u003e\n  \u003cstrong\u003eB. The \u0026ldquo;Lobster\u0026rdquo; Ecosystem and the Wave of IM Integration (OpenClaw \u0026amp; WeChat)\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#b-the\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cp\u003eThe open-source project OpenClaw (nicknamed \u0026ldquo;Little Lobster\u0026rdquo;) continues to gain high traction, prompting mainstream IM platforms to follow suit quickly.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eWeChat\u0026rsquo;s Move\u003c/strong\u003e: WeChat is reportedly grayscale testing a \u003cstrong\u003eClawbot plugin\u003c/strong\u003e that allows integration into private chats. Both @zhixianio and @op7418 expressed surprise, viewing it as a rapid response from a major tech company to AI trends.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eCommunity Ecosystem\u003c/strong\u003e: OpenClaw released versions 3.22/3.23, which refactored the ClawHub plugin system and integrated Skills like \u003cstrong\u003eXCrawl\u003c/strong\u003e (web scraping capability) and \u003cstrong\u003eNetEase Cloud Music\u003c/strong\u003e (@op7418, @AI_Jasonyu).\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch4 id=\"c-new-breakthroughs-in-vision-and-physical-world-models\"\u003e\n  \u003cstrong\u003eC. New Breakthroughs in Vision and Physical World Models\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#c-new-breakthroughs-in-vision-and-physical-world-models\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eLuma Uni-1\u003c/strong\u003e: Luma released Uni-1, an image model with reasoning capabilities that supports understanding spatial relationships and generating complex visual culture. @op7418 believes it is highly competitive in terms of aesthetic results.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eMeta V-JEPA\u003c/strong\u003e: Meta\u0026rsquo;s new paper shows that the V-JEPA model has \u0026ldquo;understood\u0026rdquo; physical laws (like gravity and object permanence) by observing videos, far outperforming Gemini 1.5 Pro in physics common sense tests (@vista8).\u003c/li\u003e\n\u003c/ul\u003e\n\u003chr\u003e\n\u003ch3 id=\"2-noteworthy-unique-perspectives-and-industry-foresight\"\u003e\n  2. Noteworthy Unique Perspectives and Industry Foresight\n  \u003ca class=\"heading-link\" href=\"#2-noteworthy-unique-perspectives-and-industry-foresight\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003ch4 id=\"-product-manager-methodology-in-the-ai-era\"\u003e\n  \u003cstrong\u003e💡 Product Manager Methodology in the AI Era\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#-product-manager-methodology-in-the-ai-era\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cp\u003e@Gorden_Sun, citing the views of Anthropic PM Cat Wu, proposed: In the AI era, \u003cstrong\u003e\u0026ldquo;Demo first, PRD later.\u0026rdquo;\u003c/strong\u003e With development costs plummeting, product managers should directly use AI to create a runnable demo for discussion, rather than wasting time on documentation. Meanwhile, everyone (designers, developers, PMs) should have the ability to directly turn ideas into products.\u003c/p\u003e\n\u003ch4 id=\"-beware-of-ai-supply-chain-attacks-the-litellm-poisoning-incident\"\u003e\n  \u003cstrong\u003e⚠️ Beware of AI Supply Chain Attacks: The LiteLLM Poisoning Incident\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#-beware-of-ai-supply-chain-attacks-the-litellm-poisoning-incident\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cp\u003e@dotey forwarded an urgent warning from @memory/karpathy-pkm-SOP.md: The popular library \u003cstrong\u003eLiteLLM (v1.82.7/8)\u003c/strong\u003e has been hit by a supply chain poisoning attack on PyPI. Malicious code steals SSH keys, cloud credentials, and environment variables. This has triggered industry reflection on the \u0026ldquo;over-reliance on external libraries,\u0026rdquo; with Karpathy suggesting that for simple functions, it is preferable to have large models generate code directly rather than introducing dependencies.\u003c/p\u003e\n\u003ch4 id=\"-openai\"\u003e\n  \u003cstrong\u003e📉 OpenAI\u0026rsquo;s Strategic Contraction: The \u0026ldquo;End\u0026rdquo; of Sora and the Super App\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#-openai\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cp\u003e@dotey provided an in-depth analysis of the news that OpenAI is shutting down the standalone Sora app. This signals that OpenAI is pulling back from \u0026ldquo;side quests\u0026rdquo; to focus on developing a \u0026ldquo;super app (Atlas)\u0026rdquo; that integrates ChatGPT, programming tools, and a browser, in order to counter the strong competition from Anthropic.\u003c/p\u003e\n\u003ch4 id=\"-humanity\"\u003e\n  \u003cstrong\u003e🧠 Humanity\u0026rsquo;s \u0026ldquo;Beginning\u0026rdquo; and AI\u0026rsquo;s \u0026ldquo;As You Wish\u0026rdquo;\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#-humanity\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cp\u003e@lijigang cited Hannah Arendt\u0026rsquo;s view that humans are the \u0026ldquo;breaking point\u0026rdquo; and \u0026ldquo;beginning\u0026rdquo; of the causal chain. In an era where the barrier to AI has vanished, humanity has leaped from \u0026ldquo;expressing opinions\u0026rdquo; to a world of \u0026ldquo;as you wish.\u0026rdquo; The core challenge has become: \u003cstrong\u003e\u0026ldquo;What do you truly want?\u0026rdquo;\u003c/strong\u003e\u003c/p\u003e\n\u003chr\u003e\n\u003ch3 id=\"3-recommended-tools--resources\"\u003e\n  3. Recommended Tools \u0026amp; Resources\n  \u003ca class=\"heading-link\" href=\"#3-recommended-tools--resources\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003ch4 id=\"-developer--productivity-tools\"\u003e\n  \u003cstrong\u003e🛠️ Developer \u0026amp; Productivity Tools\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#-developer--productivity-tools\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eClaude Code (Auto Mode)\u003c/strong\u003e: A CLI programming assistant with autonomous decision-making capabilities.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eOpenCLI / Codepilot\u003c/strong\u003e: An AI programming enhancement tool that supports WeChat bridging and a plugin system (@op7418).\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eXCrawl\u003c/strong\u003e: Now integrated into OpenClaw, providing powerful automated web data scraping capabilities (@AI_Jasonyu).\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eLogics-Parsing V2\u003c/strong\u003e: An open-source OCR model from Alibaba that can accurately recognize sheet music, flowcharts, and code (@vista8).\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch4 id=\"-creativity--content-generation\"\u003e\n  \u003cstrong\u003e🎨 Creativity \u0026amp; Content Generation\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#-creativity--content-generation\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eLuma Uni-1\u003c/strong\u003e: An AI painting model with strong reasoning abilities.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003ePrismAudio\u003c/strong\u003e: Open-sourced by Alibaba, a model that automatically generates matching sound effects for videos (@Gorden_Sun).\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eSuno CLI Plugin\u003c/strong\u003e: Based on OpenCLI, allows large models to directly generate songs in a specific style by searching for information (@vista8).\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch4 id=\"-high-quality-prompts--skill-libraries\"\u003e\n  \u003cstrong\u003e📚 High-Quality Prompts \u0026amp; Skill Libraries\u003c/strong\u003e\n  \u003ca class=\"heading-link\" href=\"#-high-quality-prompts--skill-libraries\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h4\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eLi Jigang\u0026rsquo;s Skill Repository\u003c/strong\u003e: Contains deep thinking tools like \u0026ldquo;Roundtable Discussion,\u0026rdquo; \u0026ldquo;Content Card Casting,\u0026rdquo; and \u0026ldquo;Concept Dissection\u0026rdquo; (@lijigang, @dotey).\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eAJ Slides Skill\u003c/strong\u003e: Supports 11 styles and can directly generate exquisite slides via API (@dotey).\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eLiquid Glass Landing Page Prompt\u003c/strong\u003e: @Pluvio9yte shared an engineering-grade prompt set for generating high-quality liquid glass style web pages in one go.\u003c/li\u003e\n\u003c/ul\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cstrong\u003eAnalyst\u0026rsquo;s Brief\u003c/strong\u003e:\nThe past 24 hours show that the AI industry has entered the deep waters of \u003cstrong\u003e\u0026ldquo;Agent Operating System-ification.\u0026rdquo;\u003c/strong\u003e Whether it\u0026rsquo;s Anthropic\u0026rsquo;s definition of Harness or the community\u0026rsquo;s enthusiasm for OpenClaw, it all indicates that AI is no longer just a dialog box, but is deeply taking over the physical and digital worlds through CLI and IM interfaces. Meanwhile, the LiteLLM security incident reminds us that while enjoying the thrill of Vibe Coding, supply chain security has become an Achilles\u0026rsquo; heel that cannot be ignored.\u003c/p\u003e\n\u003ch2 id=\"iv-watch-list-update-summary\"\u003e\n  IV. Watch List Update Summary\n  \u003ca class=\"heading-link\" href=\"#iv-watch-list-update-summary\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h2\u003e\n\u003cblockquote\u003e\n\u003cp\u003eTimeframe: Last 7 days; Covers 16 sources; 23 updates in total\u003c/p\u003e\n\u003c/blockquote\u003e\n\u003ch3 id=\"a16z-podcast-a_full\"\u003e\n  a16z Podcast (A_full)\n  \u003ca class=\"heading-link\" href=\"#a16z-podcast-a_full\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003e\u003ca href=\"https://ai-a16z.simplecast.com/episodes/patrick-collison-on-stripes-early-choices-smalltalk-and-what-comes-after-coding-7LptsAv9\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ePatrick Collison on Stripe’s Early Choices, Smalltalk, and What Comes After Coding\u003c/a\u003e\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-24 18:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - Michael Truell, CEO of Cursor, sits down with Patrick Collison, CEO of Stripe and an investor in Anysphere, to talk about Collison\u0026rsquo;s history using Smalltalk and Lisp, the MongoDB and Ruby decisions Stripe still lives with 15 years later, why he would have spent more time on API design if he could do it again, and whether AI is actually showing up in economic productivity data.\n\u003cul\u003e\n\u003cli\u003eThis episode originally aired on Cursor\u0026rsquo;s podcast.\u003c/li\u003e\n\u003cli\u003eFollow Patrick Collison on X: Follow Michael Truell on X: Follow Cursor: Check out everything a16z is doing with AI, including articles, projects, and more podcasts, here.\u003c/li\u003e\n\u003cli\u003ePlease note that the content here is for informational purposes only; should not be taken as legal, business, tax, or investment advice or used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003ea16z and its affiliates may maintain investments in the companies discussed.\n\u003cul\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eMichael Truell, CEO of Cursor, sits down with Patrick Collison, CEO of Stripe and an investor in Anysphere, to talk about Collison\u0026rsquo;s history with Smalltalk a…\u003c/li\u003e\n\u003cli\u003eThis episode originally aired on Cursor\u0026rsquo;s podcast.\u003c/li\u003e\n\u003cli\u003eFollow Patrick Collison on X:    \u003ca href=\"https://twitter.com/patrickc\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://twitter.com/patrickc\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003eFollow Michael Truell on X: \u003ca href=\"https://twitter.com/mntruell\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://twitter.com/mntruell\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"y-combinator-podcast-b_introsearch\"\u003e\n  Y Combinator Podcast (B_intro+search)\n  \u003ca class=\"heading-link\" href=\"#y-combinator-podcast-b_introsearch\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003e\u003ca href=\"https://podcasters.spotify.com/pod/show/ycombinator/episodes/Inside-The-Startup-Reinventing-Americas-Trillion-Dollar-Chemical-Industry-e3gn3hd\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eInside The Startup Reinventing America’s Trillion Dollar Chemical Industry\u003c/a\u003e\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-20 23:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - You\u0026rsquo;ve probably heard of OpenClaw (formerly Clawdbot/Moltbot).\n\u003cul\u003e\n\u003cli\u003eThe sensational open-source AI assistant that runs on your own device, connects with the messaging apps you already use, and goes beyond chat to actually perform tasks like managing email, calendars, files, workflows, and more.\u003c/li\u003e\n\u003cli\u003eNow, meet the person behind it.\u003c/li\u003e\n\u003cli\u003eYC\u0026rsquo;s Raphael Schaad sits down with OpenClaw founder Peter Steinberger to discuss the \u0026lsquo;aha\u0026rsquo; moment behind the viral personal AI agent, why local-first agents could replace many of today\u0026rsquo;s apps, and how personal agents will reshape the future of software.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eSolugen is reinventing the trillion-dollar chemical manufacturing industry by combining biology and chemistry in a new way\u003c/li\u003e\n\u003cli\u003eIn this episode of Hard Tech, YC\u0026rsquo;s Jared Friedman visits co-founders Gaurab Chakrabarti and Sean Hunt at their Houston HQ to see how they went from a $7,000 PVC…\u003c/li\u003e\n\u003cli\u003eThey cover the breakthrough behind their enzymatic + catalytic production, how they found their first customers, and why starting small and staying close to cus…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"all-in-podcast-a_full\"\u003e\n  All-In Podcast (A_full)\n  \u003ca class=\"heading-link\" href=\"#all-in-podcast-a_full\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://allinchamathjason.libsyn.com/four-ceos-on-the-future-of-ai-coreweave-perplexity-mistral-and-iren\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eFour CEOs on the Future of AI: CoreWeave, Perplexity, Mistral, and IREN\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-24 02:07 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - (0:00) Introduction live from Nvidia GTC (0:37) CoreWeave CEO Michael Intrator (32:58) Perplexity CEO Aravind Srinivas (1:07:11) Mistral CEO\u0026hellip;.\n\u003cul\u003e\n\u003cli\u003eIt all happened at the New York Stock Exchange - Follow the besties:.\u003c/li\u003e\n\u003cli\u003eThis piece from All-In Podcast explains how Four CEOs on the Future of AI: CoreWeave, Perplexity, Mistral, and IREN shapes the broader AI and infrastructure landscape.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003e(0:00) Intro live from Nvidia GTC\n(0:37) CoreWeave CEO, Michael Intrator\n(32:58) Perplexity CEO, Aravind Srinivas\n(1:07:11) Mistral CEO,…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eIt all happens at the NYSE - \u003ca href=\"https://nyse.com\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://nyse.com\u003c/a\u003e\nFollow the besties:\n\u003ca href=\"https://x.com/ch\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://x.com/ch\u003c/a\u003e…\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://allinchamathjason.libsyn.com/how-matt-mahan-thinks-he-can-save-california\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eHow Matt Mahan Thinks He Can Save California\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-23 09:03 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - (0:00) Matt Mahan: Why He\u0026rsquo;s Running for Governor (1:51) How California Went From Bad to Worse (12:05) Public Sect\u0026hellip;.\n\u003cul\u003e\n\u003cli\u003elang=en\u0026quot;\u0026gt;.\u003c/li\u003e\n\u003cli\u003elang=en \u0026lt;a href=\u0026quot;.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003e(0:00) Matt Mahan: Why He\u0026rsquo;s Running for Governor\n(1:51) How California Went From Bad to Worse\n(12:05) Public Sect…\u003c/li\u003e\n\u003cli\u003elang=en\u0026quot;\u0026gt;https://x.com/MattMahanSJ\u003c/li\u003e\n\u003cli\u003elang=en \u003ca href=\"https://www.instagram.com/mattmahansj/\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://www.instagram.com/mattmahansj/\u003c/a\u003e\n\u0026lt;a href=\u0026ldquo;\u003ca href=\"https://www.tiktok.com/\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://www.tiktok.com/\u003c/a\u003e @mattm…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"lennys-podcast-a_full\"\u003e\n  Lenny\u0026rsquo;s Podcast (A_full)\n  \u003ca class=\"heading-link\" href=\"#lennys-podcast-a_full\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://www.lennysnewsletter.com/p/how-stripe-built-minionsai-coding\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eHow Stripe built “minions”—AI coding agents that ship 1,300 PRs weekly from Slack reactions | Steve Kaliski (Stripe engineer)\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-25 20:03 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - 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.\n\u003cul\u003e\n\u003cli\u003eHe was part of the team that created \u0026ldquo;minions,\u0026rdquo; Stripe\u0026rsquo;s internal AI coding agents that now send about 1,300 pull requests weekly with minimal human intervention outside of code reviews.\u003c/li\u003e\n\u003cli\u003eIn this episode, Steve demonstrates how Stripe engineers activate development work from Slack, utilize cloud-based development environments for parallel agent workflows, and shows machine-to-machine payments where AI agents autonomously transact with third-party services.\u003c/li\u003e\n\u003cli\u003eListen or watch on YouTube, Spotify, or Apple Podcasts What you will learn: - How Stripe\u0026rsquo;s \u0026ldquo;minions\u0026rdquo; 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 PRs written by agents - Why Stripe\u0026rsquo;s non-engineers 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\u0026rsquo;s introduction ( 02:39 ) Stripe\u0026rsquo;s Minions and their impact across Stripe ( 04:42 ) Why activation energy is more important than execution ( 05:44 ) What is a minion?\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eTechnical Architecture (06:52) Demo: Activating Minions in Slack with Emojis (09:04) Why a good developer experience is good for both humans and agents (11:22) Walking through the agent loop and system prompt (13:42) Why Stripe chose Goose as an agent tool (16:00) The role of the developer productivity team at Stripe (17:15) Why cloud environments unlock multi-threaded AI engineering ( 21:14 ) One-shot prompting: 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 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.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eEN Key Points:\n\u003cul\u003e\n\u003cli\u003eSteve Kaliski is a software engineer at Stripe wh…\u003c/li\u003e\n\u003cli\u003eHe’s part of the team that created “minions”—Stripe’s internal AI coding agents, which now ship approximately 1,300 pull requests…\u003c/li\u003e\n\u003cli\u003eIn this episode, Steve demonstrates how Stripe engineers activate development work from Slack and leverage cloud-based development environments for parallel age…\u003c/li\u003e\n\u003cli\u003ei=1000757255000\u0026rdquo;\u0026gt;Apple Podcasts\nWhat you’ll learn:\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eHow Stripe’s “minions” write 1,300 pull requ…\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://www.lennysnewsletter.com/p/state-of-the-product-job-market-in-ee9\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eState of the product job market in early 2026\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublication Time: 2026-03-24 20:45 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - Every week, I answer readers\u0026rsquo; questions about building products, driving growth, and accelerating their careers.\n\u003cul\u003e\n\u003cli\u003eWelcome to our bi-annual state of the product job market—this is our fourth edition, and surprisingly, the most optimistic one yet.\u003c/li\u003e\n\u003cli\u003eIn spite of the headlines about layoffs and AI taking jobs, we’re actually seeing a lot of promising signs in tech hiring, and some interesting new trends: - PM openings are at the highest levels we’ve seen in over three years - AI hasn’t slowed the demand for software engineers (at least not yet) - AI roles in general are absolutely exploding - Design roles have plateaued - The Bay Area is increasing in importance - Remote work opportunities continue to decline - Despite ongoing layoffs, the overall number of tech jobs continues to grow. While these numbers are promising, I know many people are having a hard time finding a job right now.\u003c/li\u003e\n\u003cli\u003eMore job openings don\u0026rsquo;t mean people are finding jobs faster.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eTo anyone who has encountered this situation, first of all, I am very sorry.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003e👋 Hey there, I’m Lenny\u003c/li\u003e\n\u003cli\u003eEach week, I answer reader questions about building product, driving growth, and accelerating your career\u003c/li\u003e\n\u003cli\u003eFor more: Lenny’s Podcast | Lennybot | \u0026lt;a href=\u0026ldquo;https://www…\u003c/li\u003e\n\u003cli\u003evia=lenny\u0026rdquo;\u0026gt;public speaking course , and interview prep copilot\n\u0026lt;a class…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://www.lennysnewsletter.com/p/state-of-the-product-job-market-in-372\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eState of the product job market in early 2026\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eRelease Time: 2026-03-24 17:02 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - \u0026lt;a class=\u0026ldquo;image-link image2 is-viewable-img\u0026rdquo; href=\u0026quot;.\n\u003cul\u003e\n\u003cli\u003e,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….\u003c/li\u003e\n\u003cli\u003e,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….\u003c/li\u003e\n\u003cli\u003eToday\u0026rsquo;s show is our fourth semi-annual update on the tech job market (the most optimistic one to date), and it details the seven key trends shaping the employment landscape\u0026hellip;\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003e\u0026lt;a class=\u0026ldquo;image-link image2 is-viewable-img\u0026rdquo; href=\u0026ldquo;\u003ca href=\"https://substackcdn.com/image/fetch/$s\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://substackcdn.com/image/fetch/$s\u003c/a\u003e_\u003c/li\u003e\n\u003cli\u003e,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…\u003c/li\u003e\n\u003cli\u003e,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…\u003c/li\u003e\n\u003cli\u003eToday’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…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://www.lennysnewsletter.com/p/this-week-on-how-i-ai-how-microsofts\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003e🎙️ This week on How I AI: How Microsoft\u0026rsquo;s AI VP automates everything with Warp\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eRelease Time: 2026-03-23 23:02 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - How Microsoft\u0026rsquo;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 an AI. Marco Casalaina, VP of Core AI Products at Microsoft, explains how to use Warp, Microsoft 365 Copilot, and ChatGPT to dynamically launch lightweight \u0026ldquo;micro-agents\u0026rdquo; that handle everything from Azure management to document scanning to video compression.\n\u003cul\u003e\n\u003cli\u003eListen now on YouTube • Spotify • Apple Podcasts. Key takeaways: - Automation allows you to focus on higher-value activities.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eAs Marco says: \u0026ldquo;While these agents are doing what I need them to do, I\u0026rsquo;m doing other things.\u0026rdquo; When scanning his daughter\u0026rsquo;s homework, he was able to help her with math problems instead of wrestling with the scanning software.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eHardware automation is an underappreciated AI use case.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eMarco demonstrates how Warp can control his scanner via the scanner CLI tool NAPS2.\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eEN Highlights:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u0026lt;a class=\u0026ldquo;image-link image2\u0026rdquo; href=\u0026ldquo;\u003ca href=\"https://substackcdn.com/image/fetch/$s\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://substackcdn.com/image/fetch/$s\u003c/a\u003e_\u003c/li\u003e\n\u003cli\u003e,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…\u003c/li\u003e\n\u003cli\u003e,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/httpshttps%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F361d81ef-7faf-4d8e-8028-5d5e03432…\u003c/li\u003e\n\u003cli\u003e,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…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://www.lennysnewsletter.com/p/how-microsofts-ai-vp-automates-everything\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eHow Microsoft’s AI VP automates everything with Warp | Marco Casalaina\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eRelease Time: 2026-03-23 20:01 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - Marco Casalaina, Microsoft\u0026rsquo;s VP of Core AI Products and an AI futurist, demonstrates how to use AI tools to automate administrative tasks that would normally consume valuable time.\n\u003cul\u003e\n\u003cli\u003eInstead of using Warp as a coding assistant (its primary marketing purpose), Marco utilizes it to manage Azure resources, scan documents, compress videos, and more.\u003c/li\u003e\n\u003cli\u003eHe shows how these \u0026ldquo;micro-agents\u0026rdquo; reduce friction in daily workflows, allowing him to focus on higher-value activities.\u003c/li\u003e\n\u003cli\u003eMarco 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 blurring.\u003c/li\u003e\n\u003cli\u003eListen 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 that significantly improve AI performance on specialized tasks - How to build 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: Using Warp to manage Azure resources and permissions (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 of odd and even pages (15:04) The value of temporary AI solutions vs. ad-hoc AI solutions.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eMarco Casalaina , VP of Core AI Products and AI Fu…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eRather than using Warp as a coding assistant (its primary marketed purpose), Marco leverages it to manage Azure resources, scan documents, compress videos, and…\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eHe shows how these “micro-agents” can reduce friction in everyday workflows, allowing him to focus on higher-value activities\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eMarco also demonstrates how Microsoft 365 Copilot and ChatGPT can create triggered workflows that respond to emails or check for information on a schedule, high…\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://www.lennysnewsletter.com/p/the-art-of-influence-jessica-fain\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eThe art of influence: The single most important skill that AI can’t replace | Jessica Fain (Webflow, ex-Slack)\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-22 20:31 Beijing time\u003c/li\u003e\n\u003cli\u003eSummary: - 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 many past podcast guests including Stewart Butterfield, Annie Pearl, Tamar Yehoshua, and Noah Weiss.\n\u003cul\u003e\n\u003cli\u003eShe\u0026rsquo;s spent her career learning how executives actually make decisions—and why most people completely misunderstand the process.\u003c/li\u003e\n\u003cli\u003eWhy executive calendars are “like strobe lights” and why the first 30 seconds of a meeting are so important 3.\u003c/li\u003e\n\u003cli\u003eWhy executives typically optimize for the global maximum, while you are usually optimizing locally 4.\u003c/li\u003e\n\u003cli\u003eWhen leaders say something that seems wrong, the best question Jessica uses is: “That\u0026rsquo;s so interesting.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eJessica 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…\u003c/li\u003e\n\u003cli\u003eShe’s spent her career learning how executives actually make decisions—and why most people completely misunderstand the process.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eWe discuss…\n- Why great ideas often don’t get buy-in\n2\n- Why executive calendars are “like strobe lights” and why the first 30 seconds of a meeting matter so much\n3\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003e\u003ca href=\"https://www.lennysnewsletter.com/p/community-wisdom-beating-the-solo\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003e🧠 Community Wisdom: Beating the solo-founder procrastination trap, best tools for B2B prospecting, finding your first real users pre-launch, and more\u003c/a\u003e\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-22 01:19 Beijing time\u003c/li\u003e\n\u003cli\u003eSummary: - 👋 Hello, and welcome to this week\u0026rsquo;s ✨ Community Wisdom ✨ a subscriber-only email sent every…\n\u003cul\u003e\n\u003cli\u003e,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….\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e,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….\n\u003cul\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003e👋 Hello and welcome to this week’s edition of ✨ Community Wisdom ✨ a subscriber-only email, delivered every…\u003c/li\u003e\n\u003cli\u003e,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…\u003c/li\u003e\n\u003cli\u003e,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…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"stratechery-by-ben-thompson-a_full\"\u003e\n  Stratechery by Ben Thompson （A_full）\n  \u003ca class=\"heading-link\" href=\"#stratechery-by-ben-thompson-a_full\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://stratechery.com/2026/arm-launches-own-cpu-arms-motivation-constraints-and-systems/\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eArm Launches Own CPU, Arm’s Motivation, Constraints and Systems\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublication Time: 2026-03-25 18:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - Arm is selling its own chips, not just licensing IP.\n\u003cul\u003e\n\u003cli\u003eIt’s a big change compared to Arm’s history, but not surprising given how computing is evolving.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003e$15\u003c/strong\u003e / month \u003cem\u003eor\u003c/em\u003e \u003cstrong\u003e$150\u003c/strong\u003e / year.\u003c/li\u003e\n\u003cli\u003eIn-depth analysis of the day\u0026rsquo;s news through three weekly emails or podcasts.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eStratechery Interviews\u003c/strong\u003e.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eArm is selling its own chips, not just licensing IP\u003c/li\u003e\n\u003cli\u003eIt\u0026rsquo;s a big change compared to Arm\u0026rsquo;s history, but not surprising given how computing is evolving.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://stratechery.com/2026/spring-break/\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eSpring Break\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublication Time: 2026-03-23 18:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - Stratechery is on a bit of a disjointed Spring Break, as my usual week off will be spread out:.\n\u003cul\u003e\n\u003cli\u003e\n\u003cul\u003e\n\u003cli\u003eThere will be no Update on Thursday, March 19.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cul\u003e\n\u003cli\u003eThere will be no updates on Monday and Tuesday, March 23-24; there will be an Update and Interview on Wednesday and Thursday, March 25–26.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cul\u003e\n\u003cli\u003eThere will be no update on Monday, March 30.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eAll other Stratechery Plus content, including my podcasts, will stay on schedule.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eStratechery is on a bit of a disjointed Spring Break, as my usual week off will be spread out:\u003c/li\u003e\n\u003cli\u003e\n\u003cul\u003e\n\u003cli\u003eThere will be no Update on Thursday, March 19\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cul\u003e\n\u003cli\u003eThere 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\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://stratechery.com/2026/please-listen-to-my-podcast/\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003e2026.12: Please Listen to My Podcast\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublication Time: 2026-03-21 01:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - (Photo by Benjamin Fanjoy/Getty Images) Welcome back to This Week in Stratechery!\n\u003cul\u003e\n\u003cli\u003eAs a reminder, every week on Friday, we send out an overview of the content in the Stratechery bundle; highlighted links are free for everyone.\u003c/li\u003e\n\u003cli\u003eAdditionally, you have full control over the content we send you.\u003c/li\u003e\n\u003cli\u003eIf you do not wish to receive this week\u0026rsquo;s Stratechery email (without the podcast), please uncheck the box in your sending settings.\u003c/li\u003e\n\u003cli\u003eWith that said, here are some of our favorites from this week.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003e\u0026lt;img alt=\u0026rdquo;\u0026rdquo; class=\u0026ldquo;wp-image-18592\u0026rdquo; height=\u0026ldquo;1133\u0026rdquo; src=\u0026ldquo;\u003ca href=\"https://i0.wp.com/stratechery.com/wp-content/uploads/2\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://i0.wp.com/stratechery.com/wp-content/uploads/2\u003c/a\u003e…\u003c/li\u003e\n\u003cli\u003eresize=1700%2C1133\u0026amp;ssl=1\u0026rdquo; width=\u0026ldquo;1700\u0026rdquo; /\u0026gt; (Photo by Benjamin Fanjoy/Getty Images)\u003c/li\u003e\n\u003cli\u003eWelcome back to This Week in Stratechery\u003c/li\u003e\n\u003cli\u003eAs a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle; hig…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"openai-blog-a_full\"\u003e\n  OpenAI Blog (A_full)\n  \u003ca class=\"heading-link\" href=\"#openai-blog-a_full\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://openai.com/index/our-approach-to-the-model-spec\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eInside our approach to the Model Spec\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublication Time: 2026-03-25 18:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - At OpenAI, we believe 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.\n\u003cul\u003e\n\u003cli\u003eWe 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, it should be accessible for more people to understand and help shape AI.\u003c/li\u003e\n\u003cli\u003eThis is the core reason for the existence of the OpenAI Model Spec.\u003c/li\u003e\n\u003cli\u003eIt defines how we want models to follow instructions, resolve conflicts, respect user freedom, and operate safely across the extremely broad range of queries users pose every day.\u003c/li\u003e\n\u003cli\u003eMore broadly, we are trying to make expected model behavior explicit: not just during our training process, but also in a form that users, developers, researchers, policymakers, and the wider public can actually read, scrutinize, and debate.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eLearn how OpenAI’s Model Spec serves as a public framework for model behavior, balancing safety, user freedom, and accountability as AI systems advance.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://openai.com/index/safety-bug-bounty\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eIntroducing the OpenAI Safety Bug Bounty program\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublication Time: 2026-03-25 08:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eAbstract: - Testing for safety and abuse issues across OpenAI.\n\u003cul\u003e\n\u003cli\u003eAs AI technology rapidly evolves, so do the potential ways it can be misused.\u003c/li\u003e\n\u003cli\u003eOur goal is to ensure our systems remain safe and secure against misuse or abuse that could lead to tangible harm.\u003c/li\u003e\n\u003cli\u003eThrough 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 a real risk.\u003c/li\u003e\n\u003cli\u003eSubmissions will be triaged by OpenAI\u0026rsquo;s safety and security bug bounty team and may be rerouted between the two programs based on scope and ownership.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eEN Highlights:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eOpenAI launches a Safety Bug Bounty program to identify AI abuse and safety risks, including agentic vulnerabilities, prompt injection, and data exfiltration.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://openai.com/index/teen-safety-policies-gpt-oss-safeguard\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eHelping developers build safer AI experiences for teens\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-24 19:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - This article from the OpenAI blog explains how to help developers build safer AI experiences for teens, thereby shaping the broader AI and infrastructure landscape.\n\u003cul\u003e\n\u003cli\u003eFollowing \u0026ldquo;Helping developers build safer AI experiences for teens,\u0026rdquo; it also brings practical implications for founders, operators, and investors.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eOpenAI releases prompt-based teen safety policies for developers using gpt-oss-safeguard, helping moderate age-specific risks in AI systems.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://openai.com/index/update-on-the-openai-foundation\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eUpdate on the OpenAI Foundation\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-24 17:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - This article from the OpenAI blog explains how the update on the OpenAI Foundation shapes the broader AI and infrastructure landscape.\n\u003cul\u003e\n\u003cli\u003eFollowing the update on the OpenAI Foundation, it also brings practical implications for founders, operators, and investors.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eThe OpenAI Foundation announces plans to invest at least $1 billion in curing diseases, economic opportunity, AI resilience, and community programs.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://openai.com/index/powering-product-discovery-in-chatgpt\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ePowering product discovery in ChatGPT\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-24 17:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - This article from the OpenAI blog explains how powering product discovery in ChatGPT shapes the broader AI and infrastructure landscape.\n\u003cul\u003e\n\u003cli\u003eFollowing the powering of product discovery in ChatGPT, it also brings practical implications for founders, operators, and investors.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eChatGPT introduces richer, visually immersive shopping powered by the Agentic Commerce Protocol, enabling product discovery, side-by-side comparisons, and merch…\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://openai.com/index/creating-with-sora-safely\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eCreating with Sora Safely\u003c/a\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-23 08:00 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - 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.\n\u003cul\u003e\n\u003cli\u003eOur approach is grounded in concrete safeguards.\u003c/li\u003e\n\u003cli\u003eTo address the new safety challenges presented by state-of-the-art video models and new social creation platforms, we built Sora 2 and the Sora app\u0026hellip; Our approach is grounded in concrete safeguards.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eTo 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…\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eOur approach is anchored in concrete protections.\u003c/p\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"google-deepmind-blog-a_full\"\u003e\n  Google DeepMind Blog （A_full）\n  \u003ca class=\"heading-link\" href=\"#google-deepmind-blog-a_full\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003e\u003ca href=\"https://deepmind.google/blog/lyria-3-pro-create-longer-tracks-in-more/\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eLyria 3 Pro: Create longer tracks in more\u003c/a\u003e\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-26 00:01 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - Introducing Lyria 3 Pro, which can unlock longer tracks through structural awareness.\n\u003cul\u003e\n\u003cli\u003eWe are also introducing Lyria into more Google products and interfaces.\u003c/li\u003e\n\u003cli\u003eIntroducing Lyria 3 Pro, which can unlock longer tracks through structural awareness. We are also introducing Lyria into more Google products and interfaces.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eIntroducing Lyria 3 Pro, which unlocks longer tracks with structural awareness\u003c/li\u003e\n\u003cli\u003eWe’re also bringing Lyria to more Google products and surfaces.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"two-minute-papers-b_introsearch\"\u003e\n  Two Minute Papers （B_intro+search）\n  \u003ca class=\"heading-link\" href=\"#two-minute-papers-b_introsearch\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003e\u003ca href=\"https://www.youtube.com/watch?v=DmtoVnTkQnM\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eDeepSeek Just Fixed One Of The Biggest Problems With AI\u003c/a\u003e\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-24 23:17 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - ❤️ Check out Lambda here and sign up for their GPU Cloud:.\n\u003cul\u003e\n\u003cli\u003e📝 The #DeepSeek paper is available here:.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003e❤️ Check out Lambda here and sign up for their GPU Cloud: \u003ca href=\"https://lambda.ai/papers\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://lambda.ai/papers\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003e📝 The #DeepSeek paper is available here:\u003c/li\u003e\n\u003cli\u003e\u003ca href=\"https://github.com/deepseek-ai/Engram\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://github.com/deepseek-ai/Engram\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003e\u003ca href=\"https://arxiv.org/abs/2601.07372\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://arxiv.org/abs/2601.07372\u003c/a\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch3 id=\"lex-fridman-b_introsearch\"\u003e\n  Lex Fridman （B_intro+search）\n  \u003ca class=\"heading-link\" href=\"#lex-fridman-b_introsearch\"\u003e\n    \u003ci class=\"fa-solid fa-link\" aria-hidden=\"true\" title=\"Link to heading\"\u003e\u003c/i\u003e\n    \u003cspan class=\"sr-only\"\u003eLink to heading\u003c/span\u003e\n  \u003c/a\u003e\n\u003c/h3\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003e\u003ca href=\"https://www.youtube.com/watch?v=vif8NQcjVf0\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003eJensen Huang: NVIDIA - The $4 Trillion Company \u0026amp; the AI Revolution | Lex Fridman Podcast #494\u003c/a\u003e\u003c/strong\u003e\n\u003cul\u003e\n\u003cli\u003ePublished: 2026-03-24 00:24 Beijing Time\u003c/li\u003e\n\u003cli\u003eSummary: - - The following is a dialogue.\n\u003cul\u003e\n\u003cli\u003eThe CEO of NVIDIA, one of the most.\n\u0026ndash; important and influential people.\u003c/li\u003e\n\u003cli\u003eNVIDIA is the engine\u0026rsquo;s power.\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003cli\u003eEN Highlights:\n\u003cul\u003e\n\u003cli\u003eJensen Huang is the co-founder and CEO of NVIDIA, the world\u0026rsquo;s most valuable company and the engine powering the AI computing revolution\u003c/li\u003e\n\u003cli\u003eThank you for listening ❤ Check out our sponsors: \u003ca href=\"https://lexfridman.com/sponsors/ep494-sb\"  class=\"external-link\" target=\"_blank\" rel=\"noopener\"\u003ehttps://lexfridman.com/sponsors/ep494-sb\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003eSee below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc\u003c/li\u003e\n\u003cli\u003e\u003cem\u003eTranscript:\u003c/em\u003e\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n","description":"Today, the AI industry is focused on breakthroughs in inference efficiency and the evolution of Agent architecture. Google released TurboQuant, which can compress KV cache to one-sixth its size with lossless accuracy, significantly lowering the deployment barrier for long-context large models. Anthropic introduced the concept of \u0026ldquo;drive engineering\u0026rdquo; and an autonomous mode for Claude Code, marking a shift from AI assistants in dialogue boxes to Agents that autonomously execute tasks. Additionally, the LiteLLM supply chain poisoning incident reminds developers to be vigilant about security risks in the expanding AI ecosystem."}