System translated (Gemini)

2026-04-06 AI Daily | Anthropic Reviews $19 Billion Growth Path, Virtual File System Reshapes RAG Paradigm Link to heading

Anthropic has extensively reviewed its commercialization path, achieving $19 billion ARR within 14 months, signaling the entry of large model companies into a growth-driven phase. On the technology front, a virtual file system (ChromaFs) offers a lightning-fast alternative for RAG, and test cases are seen as a new moat in AI-era software development. Additionally, DeepSeek V4’s adaptation to domestic chips heralds an acceleration of domestic substitution in the inference stage.

📖 In-depth Guide to This Issue’s Watch List Link to heading

Today’s most noteworthy focus is undoubtedly Anthropic’s record-breaking commercial explosion. Amol Avasare, Anthropic’s Head of Growth, meticulously reviewed how Claude achieved the leap from $1 billion to $19 billion ARR in just 14 months. This is not only the fastest growth curve in AI history but also provides valuable practical insights for large model companies on how to transition from “technological leadership” to “commercialization-driven” growth. We strongly recommend teams focused on commercialization paths to study this.

Furthermore, it is advisable to pay attention to organizational talent perspectives and growth methodologies in the AI era. Amol shared his growth experience from Mercury and MasterClass, as well as his journey of joining Anthropic via a cold email without an open recruitment need. This reveals the screening logic of top AI labs for “atypical talent” during periods of rapid expansion. For leaders contemplating how to build an AI-era growth team, this in-depth interview on “nonlinear growth” is highly insightful.

🌐 X Platform AI Hot News Link to heading

Topic 1: Trump Issues Ultimatum to Iran Over Strait of Hormuz Blockade Link to heading

  • Category: AI · Other
  • Overview: Trending time: 2 days ago, related posts: 989,000
  • What happened: US President Trump issued a 48-hour ultimatum to Iran, demanding an immediate lifting of the blockade on the Strait of Hormuz, or face military action code-named “Trident Sweep.”
  • Why it’s important: The Strait of Hormuz handles 20% of global oil shipments. Instability leading to soaring energy prices would directly drive up AI data center operating costs and pose a threat to the stability of the global tech supply chain.
  • Discussion overview: Social media discussions focused on Trump’s aggressive rhetoric, psychological warfare interpretations of the “praise Allah” suffix, and extreme concern over a potential global energy crisis and military conflict after Tuesday’s deadline.

Topic 2: Leaders and Stars Share Easter Greetings of Hope and Resurrection Link to heading

  • Category: AI · Other
  • Overview: Trending time: 2 days ago, related posts: 815,000
  • What happened: Several global political leaders and entertainment stars posted Easter greetings on social media, emphasizing themes of hope and resurrection.
  • Why it’s important: This topic went viral under the AI category, reflecting the large-scale application of generative AI tools in creating holiday cards, visual art, and multilingual greeting videos, showcasing technology’s渗透 into traditional cultural expression.
  • Discussion overview: Discussions centered on the authenticity of AI-generated religious and cultural content, whether technology-assisted creation diminishes emotional sincerity, and the public’s acceptance of public figures using AI for social interaction.

Topic 3: Artemis II Crew Nears Record Distance from Earth on Historic Lunar Flyby Link to heading

  • Category: AI · Other
  • Overview: Trending time: 5 hours ago, related posts: 42,000
  • What happened: The Artemis II crew is set to achieve a new record for the farthest human distance from Earth during its historic lunar flyby mission.
  • Why it’s important: This mission validates highly automated navigation and deep-space communication systems, providing critical data and environmental validation for future AI applications in autonomous spacecraft control, real-time fault diagnosis, and deep-space survival.
  • Discussion overview: Discussions on X focused on the milestone significance of humanity’s return to deep space, ensuring mission safety, and the strategic value of the Moon as a springboard for future Mars exploration.

Topic 4: Google Unveils Gemma 4 for On-Device AI Power Link to heading

  • Category: AI · Other
  • Overview: Trending time: 1 day ago, related posts: 15,000
  • What happened: Google released the next generation of its open model, Gemma 4, with a focus on optimizing its on-device AI performance for mobile and local devices.
  • Why it’s important: This model demonstrates the huge potential of small parameter models to handle complex tasks on the edge, helping to improve privacy protection and reduce latency while decreasing reliance on cloud computing power.
  • Discussion overview: Discussions focused on performance benchmark comparisons between Gemma 4 and Meta’s Llama series and Microsoft’s Phi series, as well as its power consumption and inference efficiency on actual mobile devices.

Topic 5: U.S. Forces Rescue Downed Air Force Colonel in Daring Iran Raid Link to heading

  • Category: AI · News
  • Overview: Trending for: 1 day ago, Related posts: 3,100,000
  • What happened: U.S. forces conducted a high-risk raid in Iran, successfully rescuing a downed U.S. Air Force colonel. President Trump later confirmed the news.
  • Why it matters: This operation is seen as a microcosm of modern technological warfare, showcasing the critical application value of AI in defense sectors such as real-time tactical intelligence processing, autonomous drone coordination, and search and rescue path optimization in complex environments.
  • Discussion summary: Discussions on the X platform focused on the details of the advanced military technology used in the operation, its dramatic impact on the geopolitical situation in the Middle East, and a heated debate on the risks and rewards of this military decision.

Topic 6: ZEROBASEONE Unveils Striking Black-and-White Profile Photos as Five-Member Group Link to heading

  • Category: AI · Other
  • Overview: Trending for: 8 hours ago, Related posts: 106,000
  • What happened: K-pop boy group ZEROBASEONE released new black-and-white visual photos, presented as a five-member group.
  • Why it matters: The topic generated extremely high algorithmic engagement on social media, demonstrating how entertainment content can drive massive social data through visual technology. It serves as a classic case study for researching the interaction between AI recommendation algorithms and the fan economy.
  • Discussion summary: The discussions centered on the surprise of the five-member subunit format, comments on the members’ visual appeal, and speculation about the group’s future activities.

AI Public Opinion Summary on X Today Link to heading

Today’s public opinion focused on the multi-dimensional penetration of AI technology in extreme geopolitical conflicts, cutting-edge scientific exploration, and mass culture dissemination. There is a consensus that AI has become a key infrastructure for enhancing defense search and rescue, deep space navigation, and on-device computing efficiency. However, there are clear divisions regarding the authenticity of AI-assisted creation and the political consequences of military actions. The most significant current potential risk stems from the energy and supply chain crisis triggered by instability in the Middle East, which could pose a substantial threat to the continued expansion of the global AI industry by driving up data center operating costs and creating hardware supply obstacles.

💡 Influencer Insights Link to heading

Hello! I’m a senior AI industry analyst. Based on the tweet content from AI leaders and senior developers on the X platform over the past 24 hours, I have compiled today’s AI industry insight report for you.

Major Policy Shift at Anthropic: Cutting Off “Free-Riding” on Subscription Quotas by Third-Party Tools Link to heading

The biggest news today is Anthropic’s official announcement that, starting April 4th, Claude Pro/Max subscriptions will no longer cover usage generated through third-party tools like OpenClaw.

  • Core Update: Boris Cherny, head of Claude Code, stated that users utilizing Claude in third-party tools must pay via the API on a per-use basis or purchase additional usage packs.
  • Industry Impact: @dotey pointed out that this move by Anthropic aims to close a loophole where third-party tools masquerade as official clients to consume subscription quotas. @op7418 believes this has caused strong dissatisfaction in the community but could also force third-party frameworks to improve context management.
  • Technical Reflection: @dotey cited Luo Fuli, head of Xiaomi’s MiMo team, who argued that poor context management in frameworks like OpenClaw leads to massive Token waste. This “growing pain” will compel developers to shift from “brute-force window expansion” to more refined “Harness Engineering.”

Model-Level Competition: Gemma 4 and Qwen 3.6 Plus Released Simultaneously Link to heading

  • Gemma 4 (Google): Google has released Gemma 4, which natively supports multimodal and Agent workflows. @op7418 commented that it achieves a quantum leap in performance without a significant increase in parameters and uses a true Apache 2.0 open-source license, which is highly significant for on-device AI.
  • Qwen 3.6 Plus (Alibaba): Alibaba released its new model, focusing on enhancing Agent and coding capabilities, and supporting an ultra-long context of 1 million tokens. @op7418’s hands-on tests suggest its long-context development experience is significantly better than its predecessor.
  • DeepSeek V4 Rumors: @dotey cited foreign media reports that the release of DeepSeek V4 has been postponed to adapt to Huawei’s Ascend 950PR chip. This marks a key stage for domestic large models, shifting from “reliance on NVIDIA” to “domestic substitution in the inference phase.”

2. Unique Perspectives and Industry Foresight Link to heading

From RAG to “Virtual File Systems”: A New Paradigm in AI Exploration Link to heading

@dotey shared an engineering practice from the documentation platform Mintlify: they built a virtual file system called ChromaFs for their AI assistant.

  • Insight: Instead of feeding all information to the model in chunks (traditional RAG), provide the model with a set of tools such as grep and ls, allowing it to autonomously explore a “hallucinated” file system. This method reduces startup time from 46 seconds to 100 milliseconds at virtually no cost.

Test Cases Are the New “Moat” Link to heading

@ruanyf raises a profound point: when AI can replicate large software like Next.js for only $1100, the code itself is no longer a moat. The true moat is test cases, because only those with a complete testing system can verify whether AI-generated code is truly usable.

Latent Space Reasoning Link to heading

@lijigang provides a deep interpretation of the paper on model’s “internal thinking”.

  • Outlook: Current model thinking must be translated into Token (human language), which involves redundancy and information loss. The future trend is for models to reason directly in their internal vector space (Latent Space) without Token conversion, which will greatly improve reasoning efficiency and depth.

“Perceived Obsolescence” and “Game-like Feel” in the AI Era Link to heading

@nishuang discusses the shift in product design:

  • Viewpoint: Apple creates “perceived obsolescence” through “unnecessary” innovations; while excellent AI products (like CapWords) should pursue a “game-like feel” rather than simple “gamification,” retaining users by stimulating curiosity rather than solely through reward systems.

Development and Efficiency Tools Link to heading

  • Claude Island / Vibe Island: A Mac Dynamic Island notification tool developed for Claude Code, allowing users to approve AI file operation commands without switching terminals. (@dotey, @Pluvio9yte recommended)
  • Skills Manager: Developed by @yibie, used for managing, testing, and discovering Claude Code Skill plugins. (@lijigang recommended)
  • Trae: ByteDance’s AI coding tool, currently offering free access to various domestic large models. (@ruanyf recommended)
  • Codepilot: Now supports Ollama integration with local Gemma 4 models, enabling fully localized AI-assisted programming. (@op7418 recommended)

Technical Resources and Tutorials Link to heading

  • Karpathy’s AI Knowledge Base Solution: Utilizes LLM-driven Obsidian notes to achieve automatic summarization, indexing, and concept association. (@op7418, @lijigang both expressed interest in following up on the practice)
  • Harness Engineering Getting Started Tutorial: A deep dive article written by @Pluvio9yte on how to advance from Vibe Coding to systematically mastering AI workflows.
  • ChromaFs Open Source Idea: Mintlify’s proposed RAG alternative based on a vector database virtualized file system.

Other Interesting Resources Link to heading

  • CapWords: An innovative foreign language learning App that combines AI matting and recognition technology, highly praised by @nishuang for its animation and interaction design.
  • Dreamina CLI & Skill: The official text-to-image/video tool, which can be directly integrated into development workflows. (@vista8 recommended)

Analyst Brief: Today’s discussion has shifted from “how good are the models” to “how to leverage models more efficiently and cost-effectively.” Anthropic’s pricing policy adjustment is a signal that the AI industry is moving from “burning cash for growth” to “pursuing engineering efficiency and business closed-loops.” Meanwhile, the progress in on-device models (Gemma 4) and domestic computing power adaptation (DeepSeek + Huawei) indicates the formation of a diversified AI infrastructure landscape.

📚 Appendix: Today’s Watch List Update Sources Link to heading

Time window: last 3 days; covering 16 sources; 1 update in total

Lenny’s Podcast (A_full) Link to heading

  • Anthropic’s $1B to $19B growth run: how Claude became the fastest-growing AI product in history | Amol Avasare
    • Release Date: 2026-04-05 20:31 Beijing Time
    • Summary: - Amol Avasare is currently Anthropic’s Head of Growth. Anthropic is experiencing an unprecedented growth trajectory, with Annual Recurring Revenue (ARR) soaring from $1 billion to over $19 billion in just 14 months.
  • Previously, Amol worked on the growth teams at Mercury and MasterClass.
  • Before that, he was a founder, and he cold emailed his way into the Anthropic role when no job listing existed.
  • Most remarkably, he overcame a traumatic brain injury from a Muay Thai match that meant he couldn’t work for nearly a year.
  • How Amol got this position by sending a cold email to Anthropic’s Chief Product Officer Mike Krieger.