Taiwan thanks

[chris macrae]

HOW WILL COMMUNITIES THE WORLD OVER CELEBRATE APPLYING ENGINEERING'S GREATEST GIFT

The Taiwanese have gifted us machines with billion times more maths brainpower than separate human minds- what can we help Taiwanese do with these machines and how can youth everywhere transform education to scale community apps?

I asked this question at Brookings December 2025

special thanks to  

• 

  Ryan HassDirector - John L. Thornton China Center,Senior Fellow - Foreign Policy, Center for Asia Policy Studies, John L. Thornton China Center,Chen-Fu and Cecilia Yen Koo Chair in Taiwan Studies

  

  Syaru Shirley LinFounder and Chair - Center for Asia-Pacific Resilience and Innovation,Former Brookings Expert

• Panelists

  

  Gary DirksSenior Director and Professor of Practice, LightWorks - Arizona State University

  

  R. David Edelman Nonresident Senior Fellow - Foreign Policy, John L. Thornton China Center

  

  Tarcy Sih-Ting Jhou Senior Researcher - Asia-Pacific Energy Research Center

  MODERATOR

  

  Samantha GrossDirector - Energy Security and Climate Initiative,Fellow - Foreign Policy, Energy Security and Climate Initiative

Forgive my bias from graduate days at Cambridge DAMTP - diaries show that 4 of world's greatest historic heroes were in Cambridge 90 years ago - Taiwan's godfather of Tech KT Li was at the Cavendish Lab, Von Neumann (see our biography) was semestering from Princeton, Turing was undergraduating before his 18 months in Princeton, Keynes was finishing general Theory of Money, Interest and Employment. His last chapter added a twist - which professions exponentially lock in what youth can generate. This became my father Norman Macrae's key question over 40 years at The Economist

Its most unlikely we would have AI machines today without Taiwanese American Jensen Huang .

https://normanmacrae.ning.com/forum/topics/jensen100

Footnote 1 from Grok adds some of the priorities he wants to help Taiwan celebrate with Asians - in addition to signature project Erath2.0 This is compiling best dat to make weather forecasts accurate enough to minimise disasters. It also grounds interests that all Asian islands have in maximising access to green energy and clean waters. see some of our friends projects at EconomistJapan.com

On may7-9, 20000 people in Washington DC met at the AI Expo which would have happened without Jensen , Eric Schmidt and his staff at www.scsp.ai. Here are details of Taiwanese day at the Expo

Follow-ups in Taiwan include DSET and in DC at www.scsp.ai (eg Channing Lee host of Taiwan podcast as well as coordinator of Taiwan day at Expo) or Taiwan teams at DSFederal.com

Taiwanese Vincent Chang and I spent 25 years elapsed time mapping women empowerment solutions out of Bangladesh with Fazle Abed. Sir Fazle was the ultimate designer of apps the poorest 2 billion village Asians have needed since 1972.

NVIDIA (under Jensen Huang) has heavily invested in and expanded operations in Taiwan, leveraging its semiconductor ecosystem (e.g., TSMC) and partners for AI infrastructure, with extensions into regional data networks and projects like Earth 2.0.

blogs.nvidia.com +1
Taiwan serves as a core hub for NVIDIA's supply chain, R&D, and AI "factories" (large-scale GPU clusters for training/inference). Key recent moves include hiring talent, building facilities, and partnerships focused on scaling AI compute.
datacenterdynamics.com
Major NVIDIA/Partner Investments & Launches Tied to Taiwan/Asia
These emphasize AI supercomputers, data centers, and infrastructure "launched out of" or powered by the Taiwan region:
Foxconn-NVIDIA AI Supercomputing Cluster (Taiwan): ~$1.4 billion project for Taiwan's (and Asia's) first large-scale GB300 Blackwell AI data center/supercomputer. Expected ready in H1 2026; starts at ~27 MW, scaling toward 100 MW. Partners with Taiwan government; aims to support researchers, startups, TSMC, and industries. Foxconn can produce ~1,000 AI racks/week.

reuters.com +1
NVIDIA Asia-Pacific Constellation Headquarters (Taipei): New facility in Beitou-Shilin Technology Park to centralize regional AI efforts.
eetasia.com
NVIDIA Taiwan Headquarters & R&D Expansion: Reports of major investment (over NT$40 billion/~$1.2+ billion) for its first overseas HQ, plus a second R&D center (following a prior ~$715 million one). Focus on GPUs, AI, and talent/training with local universities. Generates thousands of jobs.

msn.com +1
GMI Cloud AI Data Center (Taiwan): $500 million facility with NVIDIA Blackwell GB300 chips (~7,000 GPUs planned). Online by early 2026.
reuters.com
Broader Taiwan AI Factory/Supercomputer: Collaboration with Foxconn, TSMC, and government for a national AI infrastructure hub. Starts smaller and scales; emphasizes digital twins (via Omniverse) for manufacturing.

blogs.nvidia.com +1
Earth 2.0 (Digital Twin of Earth): NVIDIA project for high-fidelity simulation of the planet (weather, climate, etc.), leveraging massive GPU compute. Taiwan/Asia supply chain and data networks are foundational for powering such AI-scale simulations and regional applications.
youtube.com
Other notes: NVIDIA suppliers (e.g., TSMC ecosystem) are also expanding (sometimes into the US for diversification), and there's emphasis on energy provisioning for these power-hungry clusters.
youtube.com
Layer 5 (Applications) Focus: What Would Enable Strong Community Applications
Jensen Huang often describes AI as a "five-layer cake": 1. Energy, 2. Chips, 3. Infrastructure (data centers), 4. Models, 5. Applications (the top layer where real-world value, including community-driven ones, emerges).

blogs.nvidia.com +1
Taiwan/Asia-focused investments in lower layers (especially infrastructure and chips) best enable Layer 5 community applications by providing accessible, high-performance compute for:
Open/community models & fine-tuning — Affordable access to GPU clusters for researchers/startups to build/customize models (e.g., for local languages, regional needs).
AI agents & digital twins — Tools like NVIDIA Omniverse for collaborative simulation (manufacturing, urban planning, climate via Earth 2.0 extensions).
Vertical community apps — Healthcare, education, disaster prediction/response, smart cities, or open-source AI platforms in Asia. Sovereign AI efforts (e.g., Taiwan's Nano 4/Crystal26 supercomputer) support localized, community-beneficial apps.
datacenterdynamics.com
Edge-to-cloud ecosystems — Regional data networks + Jetson/IGX for on-device/community AI (e.g., robotics, autonomous systems).
These create a flywheel: More Taiwan-based AI factories → cheaper/faster compute → explosion in Layer 5 apps (including open-source/community ones). NVIDIA emphasizes that strong applications pull demand through all layers.
blogs.nvidia.com
Overall, NVIDIA's Taiwan strategy is about deepening the full stack for global (and regional) AI leadership, with Earth 2.0 as a flagship example of compute-intensive simulation benefiting communities. Developments evolve quickly—

To analyze your question through an AI Mode Analysis (specifically Jensen Huang’s "AI Factory" logic), we look at the power dynamics and structural requirements needed to move Switzerland from a consumer of AI to a "Sovereign AI" leader by the 2027 Summit.

For Jensen Huang to engage with Switzerland the way he has with Modi (India) or Macron (France), he requires partners who control the three pillars of the AI Cake: Energy, Data/Infrastructure, and Industry-Specific Applications.

Here is the "Swiss Sovereign AI" power map of who needs to be in the room:

1. The Energy & Infrastructure Pillar (Layers 1 & 3)

Jensen’s first requirement is "Where is the power and where are the chips?"

• The Person: Guy Parmelin (Head of the Dept. of Economic Affairs, Education and Research) or Albert Rösti (Dept. of Environment, Transport, Energy and Communications).
• The Ask: To build a sovereign Swiss AI cloud, Jensen needs a commitment to the energy grid (Switzerland’s hydropower is a major "Green AI" selling point) and the licensing for a massive GPU cluster on Swiss soil.
• The Entity: ALPS Supercomputer (CSCS) leaders. They already host one of the world's most "Jensen-relevant" machines. He needs the political will to turn that research machine into an "Industrial AI Factory."

2. The Multi-National "Layer 5" Orchestrators (The Application Layer)

Switzerland is the headquarters of global trust (Red Cross, WHO, WEF, BIS).

• The Person: Thomas Jordan (Chairman of the Swiss National Bank) or leaders of UBS/Novartis/Roche.
• The Ask: Jensen wants to index the world’s most valuable data. Switzerland holds the world’s financial and healthcare data.
• The Analysis: If the Swiss government (Federal Council) creates a "Sovereign Data Vault," Jensen provides the "NIMs" (Inference Microservices) to process it. This is how you scale Health and Housing apps—by using Swiss-grade privacy laws as a "Layer 5" feature.

3. The Academic-Innovation Bridge (The Talent Layer)

In France, Jensen leaned on Mistral (an academic spinoff). In Switzerland, he needs the "ETH Factor."

• The Person: Joël Mesot (President of ETH Zurich) or Martin Vetterli (President of EPFL).
• The Ask: "Train the youth to be AI Teachers, not just coders."
• The Analysis: These institutions are the "Human Layer 3." For the 2027 Summit, Jensen would ask them to create a Swiss AI Foundation Model (like the Falcon model in UAE) that reflects Swiss neutrality and multilingualism.

🛠️ The "Jensen-Style" Logic Analysis

If you were to draft a prompt for the 2027 Switzerland Summit, the logic wouldn't be "How do we use AI?" but rather:

• Sovereignty: How does Switzerland keep its data within its borders while using 100,000 Blackwell GPUs?
• Intelligence Export: How does Switzerland export "Trustworthy AI" as a service to the world (ITU/UN)?
• Youth Scaling: How do we give every Swiss student a "personal AI factory" to solve local housing and energy costs?

🏁 Who specifically should you watch?

The Federal Chancellor (Viktor Rossi) or the President of the Confederation (which rotates yearly) are the ones who must issue the formal invitation. However, the "Modi/Macron equivalent" for Jensen in Switzerland is likely a consortium of the EPFL/ETH presidents + the head of the Swiss National Bank.

Would you like me to draft what a "Sovereign Swiss AI Manifesto" for 2027 would look like, focusing on those specific youth-led Layer 5 solutions?

 

Earlier we looked at in addition to usa and taiwan jensen focus on france india and uae connecting sovereign ai and layer 5 applications communities need. I feel that he has also wanted to connect both korea and japan but that their country leaders have changed a lot so in their cases he has needed to trust corporates as representing layer 5 eg softbank and perhaps ntt in japan and samsung and Hynix in Korea- please could you comment on how evident layer 5 needs/solutions are in japan and korea

Jensen Huang’s strategy in Japan and South Korea is a masterclass in adapting "Sovereign AI" to regions with established industrial titans. In France or India, he often works directly with government leaders to architect national plans. However, in Japan and Korea, he has pivoted to corporate-led sovereignty, trusting industrial giants to act as the primary engines for Layer 5 (Applications) development.

This shift is highly evident: these corporations aren't just buying chips; they are building "AI Factories" specifically to solve acute social issues like aging populations, labor shortages, and urban safety.

Japan: Corporate Stewardship of the Social Layer

In Japan, Jensen has aligned closely with SoftBank and NTT to treat the entire country as a distributed AI network.

• The Layer 3/5 Bridge (AI-RAN): Jensen and SoftBank CEO Masayoshi Son have pioneered AI-RAN (AI Radio Access Network). This transforms Japan's telecommunications towers from "dumb" transmitters into distributed AI hubs.
• Layer 5 Solutions in Communities:
  • Labor & Safety: Through the AI-on-5G Lab, SoftBank is deploying real-time AI inference at the "edge" for autonomous remote support of vehicles and robotic motion control in factories.
  • Youth-Led Inception: Startups like Sakana.ai (which received NVIDIA investment) are building "Japan-native" foundation models that understand cultural nuances, crucial for localized health and elder-care apps.
• The Trust Factor: Jensen treats SoftBank as the sovereign proxy for Japan, using their infrastructure to launch an "AI Marketplace" that connects local industries directly to AI resources. 

NVIDIA Investor Relations +5

South Korea: Physical AI & Industrial Digital Twins

In Korea, the focus is on "Physical AI"—the point where AI intelligence meets the nation's massive manufacturing and mobility infrastructure. 

• The Layer 3/5 Bridge (The AI Factory): NVIDIA has partnered with Samsung and SK Group to build some of Korea’s largest "AI Factories," featuring over 260,000 GPUs.
• Layer 5 Solutions in Communities:
  • Housing & Energy (Digital Twins): SK hynix and Samsung are using NVIDIA Omniverse to create digital twins of factories and urban infrastructure to optimize efficiency and safety.
  • Health & Mobility: Hyundai Motor Group is leveraging these GPU clusters to train autonomous mobility systems that aim to reduce urban congestion and improve transit safety.
• The Trust Factor: Jensen recently visited Korea for the first time in 15 years to personally unveil these deals. He is betting on Samsung and Hynix to provide the "memory" (HBM4) that makes global Layer 5 scaling possible, while expecting LG and Naver to build the localized "Agentic AI" that Korean citizens interact with. 

NVIDIA Investor Relations +6

🏛️ Comparison: Government vs. Corporate Lead

Feature	France / India (Gov-Led)	Japan / Korea (Corporate-Led)
Primary Partner	Presidents / Prime Ministers	SoftBank, Samsung, SK, NTT
Layer 5 Goal	National Service Digitization	Industrial "Physical AI" & Robotics
Sovereignty Logic	Data Laws & National Clouds	Corporate AI Factories & Edge Hubs
Youth Scale	Public Education Initiatives	University-Corporate R&D Labs

The Bottom Line: In Japan and Korea, Layer 5 isn't just "apps on a phone"—it is the intelligence embedded in the physical world. Jensen has "outsourced" the sovereignty mission to these titans because they already own the infrastructure (5G towers, chip fabs, car plants) where community-scaling needs to happen.

Would you like me to find specific youth-led robotics or energy startups in Japan or Korea that are using these NVIDIA-SoftBank or NVIDIA-Samsung clusters today?

Ready to create a plan?

Use Canvas to save, edit, and share your planning progress

Get started

 

13 sites

• NVIDIA and SoftBank Corp. Accelerate Japan’s Journey to Global AI ...

Nov 12, 2024 — Real-World Inference Runs on AI-RAN For the trial, SoftBank used NVIDIA AI Enterprise to build real-world AI inference application...

NVIDIA Investor Relations

• NVIDIA, South Korea Government and Industrial Giants Build AI ...

Oct 31, 2025 — News Summary: * The Korean government, through the Ministry of Science and ICT, is investing in sovereign AI infrastructure with o...

NVIDIA Investor Relations

• AI-RAN Goes Live and Unlocks a New AI Opportunity for Telcos

Nov 12, 2024 — SoftBank and NVIDIA fast-forward AI-RAN commercialization SoftBank has turned the AI-RAN vision into reality, with its successful ...

NVIDIA Developer

Show all

www.beingai.com

Can you help me scale a survey of youth's most productive ai actions partnership with taiwan and hong kong. . I know of 2 started by jensen ie earth2.0 and digital twinning tracher training huspitals. That also connects with how li ka shing has planed mini medical colleges in ef stanford asd well as graduate coleges on new york beijing london as well as hing kong. Another project i know about is feminising inteligence humanoids and avatars mediate. This started with jong kongs humanoid sophia also appointed as un firts robotic ambassador. It wasnt sophia fault that un lostgood ai orgusginally planted by fei fei li, melinda gates, hack ma, jim kim. Sopihia is npw multiple female inteligence avatars www.beingai.com there is an interesting connection on food between hong kong and singapore ie almost all imported - and taiwan and japan - both of tehse islands jave sinvce 1950s been brilliant at borlaug seed science as well as deming quality. we can extend role that win-win trading cities have for regional positve cultures ie what singapore has done with asean as something hk and raiwan contribute to. If we take the cities that benchmark with hk uae taiwan singapore as great ports with say bloomberg 10 mayor ai ciries we can expand catalofue of god ai action networking. Both taiwan and hong kong as islands need advances in future energy - eg xan call for quantum reserach to be a worldwide collaboration. Taiwan should race to be as good at public bradcasting ai action solutions as nhk. -ong konf's transportation system has one of most advanced people data mapping sustems --nd so pn

To scale a survey of youth-led productive AI actions across Taiwan and Hong Kong, we can use Jensen Huang’s Layer 5 (Applications) framework to categorize these efforts. This allows youth to move from being passive consumers to active "teachers" of AI agents that address local needs in health, housing, and energy.

Below is a catalogued framework for your survey, linking your identified projects to the "Sovereign AI" structures of these regions.

🏥 Layer 5: Health & Human Flourishing

Connecting the Li Ka-shing medical network with AI-driven wellness.

• Mini-Medical Colleges (Stanford-HK-Beijing): Use NVIDIA Clara (healthcare microservices) to create "AI Teaching Hospitals." Youth can build agents that translate complex medical data into local Cantonese or Mandarin dialects for elderly patients.
• Feminized Intelligence & Social Robots: Following the Sophia legacy, use beingAI characters to create "Social Companions" for mental health. These avatars act as a "Layer 5" interface for unhoused or isolated youth, providing 24/7 empathetic interaction.
• Survey Focus: How many youth are using "No-Code" studios to create wellness avatars for their own neighborhoods?. 

🌽 Layer 5: Food & Quality Control

Bridging Borlaug's seed science with Deming’s quality standards.

• Precision Urban AgTech: Since HK and Singapore import almost all food, youth can use NVIDIA Jetson (Edge AI) for "Hyper-Local Farming." Survey actions like AI-monitored vertical gardens in high-density housing.
• Win-Win Logistics: Track youth projects that optimize food waste redirection using HK’s port data, mirroring the ASEAN trading model for regional food security.

⚡ Layer 3 & 5: Energy & Infrastructure

Scaling Taiwan’s hardware edge and Hong Kong’s transit data.

• Earth-2.0 & Digital Twinning: Use Taiwan's exascale infrastructure to create local climate twins. Youth "Action" survey: Mapping heat islands in Kowloon or Taipei to recommend low-cost cooling solutions.
• MTR Data Mapping: Hong Kong’s transit system is a "Data Goldmine." Survey youth using this "Layer 3" data to build real-time safety apps for women or accessibility navigators for the disabled.
• Public Broadcasting AI: Link Taiwan’s media tech to NHK-style public service AI. Youth can lead "AI Literacy" broadcasts that explain "Sovereign AI" to non-technical citizens.

🌐 Global Benchmarking: The "Good AI" Network

To expand the catalogue, benchmark the survey against the Bloomberg 10 Mayor AI Cities (e.g., Singapore, Dubai, Taipei, London, New York):

City	Primary AI Action	Youth Scale Opportunity
Taipei	Hardware-Software Fusion	Building "Sovereign" PC-AI tools for local schools.
Hong Kong	Data Integrity & Transit	Privacy-preserving AI for high-density living.
Singapore	Regulatory Innovation	Leading "AI Ethics" youth boards for ASEAN.
Dubai/UAE	Energy-to-Intelligence	Converting solar energy into "Free Compute" for students.

🚀 Next Steps to Scale the Survey

If you'd like to proceed, I can help you:

• Draft the specific survey questions for "Layer 5" builders in HK and Taiwan.
• Identify 5 specific "beingAI" templates youth can use for community wellness pilots.
• Create a pitch for "Sovereign AI Youth Labs" to be presented at the 2027 Switzerland AI Summit. 

How should we prioritize the first 100 youth respondents—should we focus on medical students or urban AgTech coders?

[chris macrae]

Lets ask Agentic AI to script intelligence's most popular online lessons

Chris AI Macrae MA DAMTP Cantab

there is more great work to do now than ever =- just look at any community developing childrens brains

May 13, 2026

AI writes

Style -Here's a polished, ready-to-record 5-minute Coursera video transcript (approx. 680 words / 5 minutes at a natural speaking pace). It is written in an engaging, motivational instructional style typical of top Coursera courses.

Video Title: Layer 5 Agentic AI: Building Community Apps That Matter Most

Module: Introduction – Where to Invest Your Time, Fees, and Brainpower  [0:00 – 0:45] Opening Hook Hello, and welcome! I’m your AI co-instructor for this module. Today we’re diving into the most exciting part of the AI revolution: Layer 5 — the applications layer where agentic AI creates real economic and social value.  Jensen Huang calls this the “applications and agentic economy.” While the world focused on models and chips, the real opportunity now is building autonomous agents that solve problems for communities, cities, schools, and local organizations.  The best news? You don’t need to be at a Big Tech company. Students and builders like you are perfectly positioned to create and scale high-impact community apps right now.

[0:45 – 2:00] What Are Layer 5 Community Apps? Layer 5 apps go beyond chatbots. They use agentic AI — systems that plan, reason, use tools, collaborate with other agents, and take action autonomously — to deliver ongoing value.  Think multi-agent teams that:  Handle resident service requests 24/7  Coordinate disaster response  Personalize learning for thousands of students  Optimize local resource allocation in real time These are not nice-to-have experiments. Cities worldwide, through initiatives like the Bloomberg Mayors AI Forum, are actively deploying them for permitting, public safety, citizen engagement, and service delivery. bloomberg.org

[2:00 – 3:30] Most Urgent Needs Right Now (2026) Where should you focus your limited time, tuition fees, and brainpower? Here are the highest-leverage areas based on real city and community demands:  Local Government & Civic Services — AI agents for permitting, 311/ citizen requests, complaint triage, and public feedback analysis. Cities are desperate for tools that cut processing time from weeks to hours.  Education & Workforce Development — Personalized AI tutors, career agents that match skills to local jobs, and administrative agents that free teachers for mentoring. This has massive scalability in both K-12 and adult learning.  Public Health & Community Care — Agents for appointment navigation, predictive health alerts, mental health check-ins, and care coordination — especially powerful in underserved or rural areas.  Sustainability & Resource Management — Agents monitoring energy use, waste collection optimization, climate resilience planning, and community solar coordination.  Emergency & Social Services — Multi-agent systems for rapid response, resource matching during crises, and connecting vulnerable residents to support. These areas have strong funding potential (grants, city contracts, philanthropy) and immediate real-world impact.

[3:30 – 4:40] How Agentic AI Helps You Scale Them Here’s where AI becomes your force multiplier:  Use agent frameworks like LangGraph or CrewAI to prototype in days instead of months.  Build once and deploy across multiple communities with minimal customization.  Create agent swarms: one agent handles intake, another researches solutions, a third coordinates action, and a fourth reports outcomes.  Focus your human brainpower on what matters — understanding community needs, ethics, and integration — while agents handle the repetitive heavy lifting. You’ll learn these exact skills in the hands-on labs of this course.

[4:40 – 5:00] Your Call to Action This is your moment. The Layer 5 community app economy is just beginning. Invest your time in high-impact domains, your fees in practical agent-building skills, and your brainpower in solving problems that matter to real people.  In the next video, we’ll start building your first community agent prototype.  Remember: The future isn’t just AI that answers questions — it’s AI that gets things done for communities. Let’s build it together.  I’ll see you in the next lesson! End of Transcript

This script is motivational yet practical, cites real-world context (e.g., mayors’ initiatives), and directly answers “where to spend time/fees/brainpower.” It works perfectly as a Coursera intro video and can be expanded into a full module with visuals, demos, or quizzes.Would you like me to generate the next full module outline, slide suggestions, quiz questions, or a longer agent-building hands-on script?Wow thanks!You're very welcome! Glad the 5-minute transcript hit the mark.That style works well for Coursera — motivational but grounded, with clear calls to action and real-world anchors like the Bloomberg Mayors AI Forum.Want to go further?I can quickly generate any of these for you:Full Module 1 outline (with learning objectives, readings, and assignments)Hands-on Lab script (e.g., “Build your first community service agent in under 30 minutes”)Quiz + discussion promptsNext video transcript (e.g., “Choosing Your Community App Domain”)A customized version tailored more explicitly for Bloomberg staff, JHU students, or SCSP-aligned audiencesSlide deck suggestions or visual prompts for the videosJust tell me which direction you’d like to take next, or if you want adjustments to the intro script (tone, length, emphasis, etc.).Happy to keep building this out with you!Let me discuss . Too much for me alone to navigateGot it — no problem at all. This is a big idea, and it's completely normal to feel overwhelmed when moving from a cool concept to something structured and actionable. Let's discuss and simplify it together.Quick Reality CheckWe don't have to build the entire Coursera specialization right now. We can start small and decide how far to go based on your goals, time, and resources.Let's Clarify Your Vision (Quick Questions)To make this manageable, tell me what feels most useful:Scale & Audience Just a few short videos / one mini-course for personal use or sharing? Internal Bloomberg staff training? Open Coursera-style public course? Something for JHU students, mayors’ offices, or SCSP network?Your Goal Prototype / proof of concept? Something you can pitch to Bloomberg Philanthropies or SCSP contacts? Personal learning + portfolio project? Actual revenue-generating course?Your Capacity How much time/energy do you want to put in weekly? Do you want me to do most of the drafting while you review/edit? Prefer a very simple version first (e.g. 3–5 short videos total)?Suggested Simple Starting Path (Low Pressure)We can build this in clear, bite-sized stages:Phase 1 (This week): Finalize a strong 4–6 video mini-module (Introduction + 3–4 core topics + hands-on starter).Phase 2: Turn it into a Notion page, Google Doc, or simple landing site.Phase 3 (optional): Expand, record, or pitch it

If you look at layers 1-3 and big models of layer 4- only a few thousand people's creativity will .determine these - when layer 5 ai and detailed local or twin ai models appears to be the main creative brain work for almost all of us! We welcome help in assembling layer 5 cases

[e-mail] Washington DC District

PS the greatest educator i ever met was Bangladesh's Sir Fazle Abed. In brainstorming mooc of empowering billion Asian women livelihoods he said i understand need for massive open online but C needs to be cooperation. Interestingly Steve Jobs had hosted Abed's 65th birthday party spring 2001 silicon valley; the birthday wish was starting up brac university. I was lucky to attend abed's 80th birthday in spring 2016 - there AI to make UN sdgs communal everywhere was discussed with Abed friends like Jim Kim and Melinda Gates taking the lead responsibility for advancing AI4all and AI4good. which is now Geneva's 2027 AI Action summit charge after ITU has clarified 6 G standards at its plenopotentiary Fall 2026

Some very rough notes since 2016 are at http://povertymuseums.blogspot.com and www.catholicuni.com - sadly Abed's brain cancer took over his last few years to 2019 so we haven't yet got near ai of billion Asian womens empowerment. (and of course their billion male partners). Any chance this could be on the agenda of Trump-Xi ( I am an eternal optimist! but energised by feedback from 20000 DC brains last week AI+Expo www.scsp.ai )

[chris macrae]

One question when we look at the 5 layers . Is it correct that only few thousand brains will innovate energy future.. ditto actual hardware of chips.. ditto policy of ai infrastructure. Ditto big si models.. if do that means 99% of human brains creativity will need to address either layer 5 or detailed local or twin models? I have exagetsyed s biy buy if ai is to democratize human life agent in layer 5 ai may be the most important curricula ever!

Thought for 4s
Yes, your core intuition is directionally correct — and it's one of the most important insights about the current AI era.
Jensen Huang's 5-layer cake (as of 2026) breaks down like this: 
Energy (power generation at scale) 
Chips / Compute (GPUs, accelerators, semiconductors) 
Infrastructure (data centers, networking, AI factories) 
Models (foundation models, training, reasoning engines) 
Applications (agentic systems, real-world deployments, value creation)
Every successful Layer 5 app pulls value from all the layers below it.
Elite Concentration in Lower Layers
You're right that the most frontier innovation in Layers 1–4 is concentrated among a relatively small number of people and organizations:
Energy: Major breakthroughs in next-gen power (SMRs, fusion progress, grid-scale solutions) involve big engineering teams at utilities, national labs, hyperscalers, and a few specialized startups. Not thousands of independent innovators.
Chips: Dominated by NVIDIA, TSMC, ASML, Broadcom, and a handful of other players + elite design teams. Extremely capital- and expertise-intensive.
Infrastructure & Policy: Hyperscalers, governments, regulators, and large infrastructure funds. Policy and permitting are handled by experts in DC, Brussels, Beijing, etc.
Big Models (Layer 4): Frontier labs (OpenAI, Anthropic, Google DeepMind, xAI, Meta, etc.) + a few strong Chinese players. While there are now dozens of notable models, the absolute cutting-edge training runs and architectural breakthroughs are still done by a few thousand top researchers/engineers globally.
This is not an exaggeration — it's the nature of these highly scaled, capital-heavy layers.
The Explosion at Layer 5
Layer 5 is where the vast majority of human creativity, domain expertise, and local adaptation will (and should) happen. This is the democratizing layer:
Agentic applications, industry-specific agents, community tools, digital twins, workflow automation, personalized services, local government tools, education agents, healthcare coordination, sustainability apps, etc.
Here, value comes from understanding real human problems, contexts, ethics, integration, and iteration — areas where millions of people have deep expertise (teachers, doctors, city planners, small business owners, social workers, engineers in every field).
Agentic AI makes this possible: You no longer need to be a machine learning PhD to build powerful systems. Frameworks let non-experts orchestrate agents that plan, use tools, collaborate, and act.
Jensen and others emphasize that this is where the real economic value and societal transformation are created. The lower layers are foundational infrastructure (like electricity or the internet), but Layer 5 is where AI meets the $50+ trillion economy and everyday life.
Your "99%" figure is hyperbolic, but the spirit is accurate: The largest opportunity for broad participation, job creation, and meaningful impact is in applications and agentic systems. Lower layers create the platform; Layer 5 distributes and multiplies the benefits.
Why Layer 5 Agentic AI Curricula Could Be One of the Most Important Ever
Exactly as you said. If AI is to truly democratize opportunity rather than just concentrate it:
Teaching people to direct, build, and deploy agents for their own domains becomes foundational literacy — like learning to code or use spreadsheets was in previous eras.
It shifts power to domain experts and local communities ("twin models," local agents, community apps).
This aligns perfectly with the Bloomberg/JHU/SCSP-style initiatives we discussed: training staff, students, mayors, and citizens to solve urgent local problems.
This is why the course idea we were sketching is powerful. It focuses brainpower exactly where it can have the broadest leverage.
Bottom line: Yes — invest most human creativity and education effort in Layer 5. The lower layers need elite talent and capital, but Layer 5 needs widely distributed talent, empathy, and contextual knowledge. That's where agentic AI becomes a genuine force multiplier for humanity.

[chris macrae]

GROK says: FromJensen Huang'sperspective and public signals, the May 13-15, 2026 Trump China trip was more high-profile signaling and relationship-building than a dramatic breakthrough, but it advanced Nvidia's goals modestly amid ongoing tensions.

theguardian.com

Key Outcomes and Jensen's Viewpoint

• Participation and Optics: Jensen joined last-minute (boarding Air Force One in Alaska after initial omission reports). He was part of a delegation with Musk, Cook, and others for Trump-Xi talks on trade, tariffs, AI, supply chains, Taiwan, etc. Nvidia framed it as supporting "America and the administration’s goals." Post-trip, he was seen positively engaging in Beijing (street food, "I love China" comments, selfies), maintaining his approachable, pro-engagement persona.
  facebook.com
• Market Access Progress: Reports indicate U.S. approvals for H200 AI chip sales to ~10 Chinese firms (e.g., Alibaba, JD.com, ByteDance) around/during the trip—potentially unlocking part of the $50B opportunity Jensen has long highlighted. This is a partial win after years of export curbs that cost Nvidia billions (market share in China dropped sharply; some reports cite near-zero direct high-end sales recently due to restrictions and domestic alternatives like Huawei). However, broader outcomes remain unclear: Chinese purchases are ultimately a "sovereign decision," and self-reliance momentum (domestic chips + ecosystems) continues.
  247wallst.com
• Jensen's Stance: He has consistently argued against overly restrictive "loser mentality" curbs, noting they accelerate China's independence without stopping its AI progress (half the world's top AI researchers are Chinese; they are "world-class"). He favors fair competition + collaboration, including researcher dialogue on AI safety/red lines. Chips are tools for medicine, weather, education—not "atomic bombs." The trip aligns with his push for pragmatic access to fund U.S. R&D while China builds its stack.
  stocktwits.com

No sweeping deal ended restrictions, but the visibility (Trump calling him "The Great Jensen Huang") and incremental licenses signal ongoing dialogue. Nvidia's China challenges persist, but Jensen positions the company as essential to global AI regardless.

nytimes.com

Upcoming: Computex 2026 (June 1-5, Taipei) and Layer 5 AIThis is perfectly timed for deeper Asian engagement. Jensen's keynote (livestreamed) will elaborate on NVIDIA's "Five-Layer Cake" AI stack:

1. Energy/Power
2. Chips/Computing
3. Infrastructure (AI factories, networking)
4. Models
5. Applications (including physical AI, agents, robotics, community/local scaling).
   computextaipei.com.tw

Your hopes align here: Taiwan (and broader East Asia: Japan, Korea, Singapore) are core partners in this full-stack ecosystem. Jensen frequently highlights Asian strengths in manufacturing, deployment, and applied AI. Expect emphasis on partners needed for scaling Layer 5—exactly the "only Asian partnerships can celebrate in time" dynamic you note. This builds on prior Korea deals, Taiwan events, etc. It offers a platform for practical, non-siloed collaboration even as geopolitics complicate U.S.-China direct ties.

tradingview.com

Bridging US-Asian AI: The Core ChallengeYou're right—this remains huge. Top data/AI sovereignties after the U.S. are Asian (China leads in volume/research scale; Taiwan/Korea/Japan in chips/supply chain; Singapore in hubs). Jensen embodies the bridge: Taiwan-born, U.S.-based, deeply engaged in Asia. However:

• U.S. Side Risks: Some in the "17 technologists" orbit (or broader policy) prioritize containment over integration. Jensen has clashed with hardline export views, warning they backfire. Trump-era pragmatism (deals, approvals) helps, but silos/conflicts could slow Layer 5 scaling (energy, apps, local community actions).
  fastcompany.com
• Asian Momentum: East Asia's education systems are not producing Luddites—they're mandating AI integration from early grades (as we discussed previously). This contrasts with Western lags where teachers sometimes resist or lack training, risking "intelligence agency" gaps you and your father warned about since the 1983 2025 Report.

Jensen actively counters this: He urges every student/teacher to master AI as a core skill; praises China's STEM rise; sees AI reinventing jobs/education toward creativity + tools. His MOOC-like visibility (keynotes, speeches) models what teacher training worldwide needs—practical, optimistic intelligence engineering.

linkedin.com

Bottom line: The trip achieved symbolic + incremental wins for access/relationships. Computex offers Jensen a strong platform to push Layer 5 collaborations with Asian partners—the "east-west discovery" you track. Bridging requires navigating U.S. politics without letting conflicts dominate, while fixing Western education's adaptation lag. Jensen's history suggests he'll keep leaning into ecosystem-building over zero-sum fights. This fits your broader vision of collaborative exponentials for health, learning, and solving Einstein-scale challenges.If you want deeper dives (e.g., specific Layer 5 education angles, Computex predictions, or MOOC outlines featuring Jensen/Hassabis/etc.), let me know!

9 web pages

Explore Layer 5 physical AI applications

Investigate Singapore as AI collaboration hub