Instagram | Sponsor this newsletter

My fellow AI explorers

Picture this. Elon Musk and Jensen Huang are on Air Force One with Donald Trump, flying back from a China trip where Nvidia just unlocked H200 chip sales to Chinese buyers.

And at the exact same moment, the Vatican is quietly preparing the most consequential document on AI ethics the world has ever seen.

This is the week AI stopped being just a tech story.

We've got the Pope, a geopolitical arms race, a Microsoft CEO giving your career an 18-month expiry date, and a Python type checker that just went production-ready at Meta's scale. Wide, weird, and wildly important range.

In today’s edition:

• ⛪ Pope Leo XIV is about to drop the AI ethics encyclical Silicon Valley didn't ask for

• 🌐 Anthropic's 2028 war-game: democracy vs. authoritarianism, and who's winning right now

• ⏱️ 30-Second AI Play: Make Pyrefly the quality layer in your AI coding loop

• 📰 Other Relevant AI News: Malta goes all-in, Microsoft's 18-month warning, and more

Turn AI into Your Income Engine

Ready to transform artificial intelligence from a buzzword into your personal revenue generator?

HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.

Inside you'll discover:

• A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential

• Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background

• Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve

Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.

Get Your Guide

⛪ The Big Story

Pope Leo XIV vs. Silicon Valley: The AI Ethics War Just Got Interesting

This is not a story about an out-of-touch institution chasing a trend. This is a story about the most globally influential moral authority on Earth deciding to get in the ring.

Pope Leo XIV has formally created an in-house AI study group, and in the coming weeks he's releasing his first papal encyclical: a major Church document with binding weight for 1.4 billion Catholics. He’s expected to call for an ethics-based approach to AI centered on human dignity, social relationships, and peace.

Here's what you need to know:

• The Vatican created the study group because of "the acceleration in AI's use, its potential effects on human beings and on humanity as a whole, and the Church's concern for the dignity of every human being."

• The encyclical is expected to become a direct flashpoint between Pope Leo and the Trump administration, which has made rapid AI development a matter of vital national economic and security strategy.

• The document landed right as Trump wrapped up a China visit with Elon Musk (whose platform X hosts Grok) and Nvidia CEO Jensen Huang (who just secured federal approval to sell H200 chips to Chinese buyers) flying alongside him on Air Force One.

Here's what makes this genuinely different from every other AI ethics document you've scrolled past and forgotten.

Every think-tank paper, every EU directive, every congressional hearing on AI regulation has to fight for relevance. It gets buried in technical jargon, captured by industry lobbyists, or quietly watered down before it ever becomes law. The Vatican doesn't have that problem. A papal encyclical lands with moral authority that no policy memo, no matter how well-argued, can manufacture.

And Pope Leo isn't coming in with vague platitudes. The early signals suggest a document that engages seriously with specific harms: automated repression, AI-enabled surveillance, the displacement of human relationships by machines, and the concentration of AI power in the hands of a tiny group of people accountable to nobody but their shareholders.

The timing is also fascinating. This drops into a world where Anthropic is publishing geopolitical war-games, Microsoft is predicting mass white-collar automation, and OpenAI is literally making deals with national governments. The AI debate has left the lab and entered civilization.

The Pope showing up is not a surprise. It was inevitable.

🔮 Prediction: This encyclical becomes the most-cited document in global AI policy debates outside of the US and China. Expect it to land particularly hard in Latin America, sub-Saharan Africa, and Southern Europe, where the Vatican carries deep political weight and where AI governance conversations are still wide open. Silicon Valley can dismiss an EU directive. Dismissing the Pope in those regions is a much harder political play.

What do you think? Does the Pope's voice actually move the needle on AI ethics, or is this just moral theater? Reply and tell me. I read every single one.

Read the full story here →

Future

🌐 Anthropic's 2028 Report Is the AI Geopolitics Briefing Nobody Was Ready For

Two Futures. One Window. And the Clock Is Running.

This week, Anthropic published a detailed policy paper outlining two possible futures for US-China AI competition by 2028. It is the most direct, specific, and unambiguous statement any major AI lab has ever made publicly about what is actually at stake in this race. And it is worth your full attention.

Here's the framing:

• Compute is everything. The most important ingredient in building frontier AI is access to advanced chips. The US controls this supply through export controls on Nvidia and allied semiconductor manufacturing equipment. Those controls have been working… but not completely.

• China is closer than most people admit. Despite chip restrictions, Chinese AI labs have stayed competitive through two workarounds: smuggling banned chips and accessing export-controlled compute in offshore Southeast Asian data centers; and running systematic "distillation attacks," or fraudulently mass-scraping US frontier models through fake accounts to replicate their capabilities at a fraction of the cost.

• The gap is real, but fragile. US frontier systems are estimated to be at least several months ahead on intelligence. But Anthropic argues the window to lock in a decisive 12–24 month lead is right now, in 2026. And it will not stay open indefinitely.

Scenario One — Democracy leads: The US closes the loopholes. This means tighter export controls, aggressive enforcement against distillation attacks, and a coordinated push to make American AI the backbone of the global economy. In this world, democratic values shape how AI is built and deployed everywhere. A commanding lead in capabilities also gives the US leverage to engage China on AI safety, which benefits everyone.

Scenario Two — Neck-and-neck: Policymakers don't act. China catches up on model intelligence. Beijing's whole-of-nation "AI+" push integrates near-frontier AI faster across its economy and military. Huawei data centers win in the Global South on price. AI norms start being written by authoritarian regimes. And the safety practices, which are already dramatically weaker in Chinese labs than American ones, become the global default.

The paper includes a wake-up call moment built around Anthropic's own Mythos Preview model, released to select partners in April. With access to it, Firefox fixed more security bugs last month than in all of 2025 combined: nearly 20 times its monthly average. A Chinese cybersecurity analyst responded publicly that China is "still sharpening our swords while the other side has suddenly mounted a fully automatic Gatling gun."

The policy asks are concrete: close offshore data center loopholes (a House bill to do this passed 369-22 in January), criminalize distillation attacks, and actively promote American AI exports globally. Several of these are already moving through Congress with bipartisan support.

🧠 Prediction: This paper is Anthropic playing the long game on policy, and doing it in public. By publishing it, they're shaping the terms of the export control debate, the distillation attack legislation, and the broader narrative around AI and national security. And this is all before the 2026 midterms. It will be cited in congressional testimony within weeks. The 12–24 month lead framing is already showing up in how US officials are talking about this.

Read Anthropic's full 2028 paper here →

30-Second AI Play

🐍 Add Pyrefly to Your AI Coding Loop and Stop Shipping Type Errors

Meta's open-source Python type checker just hit stable v1.0. And if you're using any AI coding agent (Claude Code, Copilot, Cursor, Codex), this is the quality layer that turns AI-generated code from "probably fine" to "production-ready."

Here's how to set it up in under 5 minutes:

1. Install it. Run pip install pyrefly or uv pip install pyrefly.

2. Initialize your project. Run pyrefly init at your root. If you're migrating from Mypy or Pyright, it automatically converts your config.

3. Add a skill file. Drop pyrefly-cli.md in .agent/skills/ with this instruction: "Run pyrefly check at the root of the project before completing any task that modifies Python files."

4. Back it up with an AGENTS.md directive. Add: "Before completing any task that creates or modifies Python files, run pyrefly check and fix all errors before finishing." The skill alone won't always trigger; the directive makes it mandatory.

5. Hook it to the Stop event (Claude Code users). Add this to .claude/settings.json: "hooks": {"Stop": [{"hooks": [{"type": "command", "command": "pyrefly check >&2 || exit 2"}]}]}. Now type checking runs automatically every time your agent wraps up a task.

🔍 Why it matters: Pyrefly checks over 1.85 million lines of code per second, which is 15x faster than Mypy. It's already the default type checker for Instagram's 20-million-line Python codebase at Meta, and adopted by PyTorch and JAX. As AI agents write more of your code, having a type checker baked into the loop is the difference between shipping a feature and spending three hours debugging a signature mismatch that your coding agent was completely confident about.

Check out Pyrefly v1.0 | Full agentic loop guide here →

Other Relevant AI News!

🇲🇹 OpenAI just made an entire country its AI showcase. Malta became the first nation to give every citizen free ChatGPT Plus for one year, gated behind a free AI literacy course from the University of Malta. It's the cleanest experiment yet in AI as civic infrastructure, and every government in the EU is watching closely.

⏱️ Microsoft's AI chief gave your career an 18-month countdown. Mustafa Suleyman told the Financial Times that AI will hit "human-level performance on most, if not all professional tasks" within a year and a half, naming accounting, legal, marketing, and project management as first in line. The Fortune piece digs into both the warning and the counterdata, and the counterdata is more reassuring than the headline suggests.

🇨🇳 The chip smuggling problem is bigger than anyone admits. Federal prosecutors recently charged a Supermicro co-founder with diverting $2.5 billion worth of servers containing advanced US chips to China. Per Anthropic's new report, DeepSeek's latest model was trained on US chips that are banned from sale to China. This is not a minor loophole. It's a $2.5 billion hole in the export control regime.

🏠 AI is physically reshaping real estate. Data center demand is hitting record highs for the third straight year, with over 80% of new capacity pre-leased before it's even built, as hyperscalers bet everything on AI infrastructure. Power access has officially overtaken connectivity as the number one factor in where data centers get built.

🏛️ Congress is moving on distillation attacks. Legislation from the House Foreign Affairs Committee targeting China's practice of mass-scraping US frontier AI models passed out of committee unanimously this week, marking one of the few moments of genuine bipartisan agreement in the AI policy space.

Golden Nuggets

• ⛪ The Vatican entering the AI debate isn't a punchline. It's the opening of a new front in AI governance that carries moral weight at a scale no policy paper from Washington or Brussels can match.

• 🌐 Anthropic's 2028 paper is required reading for anyone who wants to understand what the US-China AI race actually looks like right now: not the hype version, the real one.

• 🐍 If you're shipping Python with AI coding agents, Pyrefly v1.0 is the missing quality layer. Five minutes of setup, production-grade type checking on autopilot.

Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)

Until our next AI rendezvous,

Anthony | Founder of Uncover AI