⚙️ Microsoft Restructures for AI, Apple Pay Goes Dark, and China's AI Shocks USA

In partnership with

Get the tools, gain a teammate

• Impress clients with online proposals, contracts, and payments.

• Simplify your workload, workflow, and workweek with AI.

• Get the behind-the-scenes business partner you deserve.

Try HoneyBook for free

1. ⚖️ U.S. House Proposes 10-Year Ban on State-Level AI Regulation

In a surprising twist buried inside a broader tax bill, U.S. House Republicans have introduced a clause that would ban state governments from regulating artificial intelligence for the next decade. The goal? To create unified, federal oversight that avoids a patchwork of state-level rules which could hinder innovation and create compliance headaches for tech companies. However, the move is already drawing sharp criticism — not just from Democrats, but also from some Republican governors.

Critics argue this is an example of federal overreach and an attempt to tie the hands of states just as AI regulation becomes more urgent, especially in areas like education, labor, and surveillance. Legal experts point out that bundling the provision into a budget-related bill could violate Senate rules, making its passage uncertain. Still, it’s a clear sign that AI policy is moving from theory to political battleground — and fast.

2. 🤖 AI Agents Are Smarter Than Ever — But Still Locked Out of the Real World

AI agents — autonomous bots powered by large language models like GPT-4 or Claude, have made massive leaps in reasoning, memory, and planning. They can now write emails, schedule meetings, and even simulate entire workflows. But there’s one big catch: they’re terrible at actually doing things in the real world.

Why? Because most digital platforms still require human authentication, multi-factor security, or deep API access. For example, if you wanted an AI agent to book your flight, it would still need your login, payment method, and two-factor code — steps it can’t reliably handle. That’s where companies like Arcade.dev (Not Sponsored) come in. They’re building secure “middleware” infrastructure that gives AI agents access to apps like Stripe, Google Calendar, and airline booking systems, but with guardrails and human oversight.

Meanwhile, Anthropic is developing something called the Model Context Protocol (MCP) — a way to standardize how AI models plug into external services. Think of it like a translator between AI and software APIs. Once these hurdles are cleared, we could see a real wave of autonomous digital assistants handling your email, errands, and admin without touching a keyboard.

3. 💼 Microsoft Lays Off 7,000 as It Doubles Down on AI

In a move that shocked many inside and outside the company, Microsoft announced it is laying off roughly 7,000 employees, about 3% of its global workforce. The layoffs span departments from sales to engineering and are part of a major reallocation effort to support Microsoft’s massive investment into AI.

CEO Satya Nadella said the company will spend up to $80 billion on AI development in fiscal year 2025, much of it directed toward Azure-based infrastructure, partnerships (like OpenAI), and internal product transformations. Microsoft has already integrated AI into Office, Teams, and Windows, and now it’s turning its focus to building a true “AI-first” operating model across all divisions.

The layoff announcement has sparked debate: Is this a responsible strategic shift — or a sign that AI is replacing human workers faster than we’re prepared for? Either way, Microsoft is signaling that AI isn’t a feature — it’s the core business model going forward.

4.📱 Apple Pay Outage Exposes the Fragility of Cashless Systems

On May 16, users across the U.S. were caught off guard when Apple Pay suddenly went down, disrupting transactions in stores, apps, and online platforms. For millions who rely on it daily, it was a wake-up call. The outage, which also affected Apple Cash and Apple Card, lasted several hours and left users unable to make purchases at registers or send money through iMessage.

Apple hasn’t disclosed the root cause, but the incident highlights a growing vulnerability in our increasingly cashless economy: when digital payment rails fail, there’s often no fallback. Financial experts have long warned that while fintech offers convenience, it also creates central points of failure — especially when linked to a few dominant providers like Apple, Google, or PayPal.

This may be a turning point in how governments and businesses think about redundancy in financial infrastructure. For everyday users, it’s a reminder that “Apple Wallet ≠ a wallet.” Keep a real one, and maybe even some actual cash — just in case.

5. 🧠 OpenAI Launches Codex: A New AI Coding Assistant

OpenAI has unveiled Codex, a new AI agent designed to assist with software engineering tasks. Built on the codex-1 model, Codex can write code, fix bugs, run tests, and explain codebases. It’s available to ChatGPT Pro, Team, and Enterprise users. OpenAI aims to strengthen its position in the AI coding space, competing with companies like Microsoft, Google, and Amazon. Codex is cloud-based and can execute tasks independently, with delivery times ranging from one to thirty minutes. It also features advanced capabilities such as matching an organization’s coding style and assisting in code reviews.

6. 🇨🇳 DeepSeek’s R1 Model Stuns Microsoft — “Closest to OpenAI Yet”

In a surprising public statement, Microsoft CEO Satya Nadella praised Chinese AI company DeepSeek, calling its flagship model R1 the first non-OpenAI language model that genuinely rivals GPT-4. This wasn’t just a passing compliment — Nadella explicitly said he believes R1 outperforms Google’s Gemini and Elon Musk’s xAI model, Grok.

So what is DeepSeek R1? It’s a multimodal large language model developed by DeepSeek, a relatively young but fast-rising Chinese AI lab backed by Alibaba’s cloud unit and former researchers from Tsinghua University. R1 supports coding, multilingual dialogue, document analysis, and even vision-based reasoning, making it a true generalist like GPT-4 or Claude 3.

Why does this matter? First, it signals that China is rapidly closing the gap in frontier AI. For months, most global discussions have focused on U.S. players — OpenAI, Anthropic, Google DeepMind. Nadella’s recognition of R1 as a viable competitor legitimizes China’s role in the global LLM race.

Second, this raises geopolitical and regulatory questions. As U.S. and EU governments debate AI regulation and export restrictions, having a Chinese model that performs on par with GPT-4 may challenge the idea of Western AI dominance. It may also increase pressure on governments to support homegrown models more aggressively.

For developers and founders, the takeaway is clear: Open-source and globally competitive LLMs are multiplying fast. And DeepSeek R1 is now one to watch closely.

What type of Content do you want to see?

I’d love to hear what kind of content you actually want more of.

Click 👉 Here to fill out the 5 second survey. That would be much appreciated!