Is It Time for a Divorce? Why Microsoft Needs to Split Windows in Two

3 min read Original article ↗

Costin Raiu

Windows, AI, and the Backlash: Does Microsoft Need a New Direction?

Microsoft is pushing hard to make AI the centerpiece of Windows. It started with the infamous Recall project — a feature that silently screenshotted your desktop every few seconds. After widespread outrage, it was pulled back and redesigned. But the story didn’t end there. Copilot arrived next, popping up with intrusive suggestions and “help” that most people never asked for.

Why this relentless push? Microsoft has several motivations:

  • they need Windows to look fresh and modern
  • they want to ride the AI hype wave
  • they want users to upgrade hardware
  • most importantly — they want Windows to feed their growing cloud and AI subscription ecosystem.

The deeper the integration, the better they can predict user behavior and personalize (or monetize) experiences.

But the reaction from users? Overwhelmingly negative. Microsoft’s AI president even admitted he was shocked people didn’t see the value. So what’s going on?

Press enter or click to view image in full size

Readers added context

Interestingly, we accept Siri as a natural part of iOS and macOS, and when Apple announces a partnership with Google to improve its AI, people celebrate. Yet when Microsoft does something similar, the pitchforks come out.

The difference comes down to approach, intrusiveness, and privacy.

Siri feels like a companion you can call on. Windows AI feels like Big Brother checking in every five minutes. Combine that with Windows becoming more bloated — more telemetry, more preinstalled apps, more features nobody asked for — and the frustration becomes understandable.

Speaking of frustration: if you work in security, you know the pain. Trying to run malware scanners on Windows is almost impossible. Even if you disable Defender — because you have to — it eventually re-enables itself, starts blocking samples, and quietly sends data back to Microsoft. That’s why many of us now do this work exclusively on Linux.

So, what’s the solution? How does Microsoft fix its AI future?

My proposal: Split Windows into two versions.

Here’s my suggestion to Microsoft. It’s free, feel free to take or it skip it, Microsoft. Split Windows into two versions:

Windows Classic
A lightweight, fast, reliable, privacy-first version. No cloud. No forced online accounts. No bundled Teams, Skype, Defender, Office, or other services. No flashy animations — just solid performance. Perfect for enthusiasts, gamers, developers, researchers, and anyone who likes a tidy system. I would use Windows Classic.

Windows AI
A bold, futuristic Windows built from the ground up for the AI era. Deep AI integration into browsing, writing, shopping, payments, security, and everyday workflows. A fluid glass UI, voice commands, face ID — the works. It leans heavily on the cloud and runs best on modern hardware.

Microsoft can focus its innovation on Windows AI while keeping Windows Classic updated with security patches and new hardware support. This isn’t Windows Pro vs. Enterprise; it’s a fundamentally different philosophy for the AI age.

In the long run, AI might evolve into something trustworthy — privacy-aware, encrypted end-to-end, respectful of user control, and genuinely helpful. When that happens, people might willingly switch!

But today? People want choice.

So, what do you think? Should Microsoft build a sleek “Windows AI” while offering a minimal, efficient “Windows Classic”? Or should there be only one Windows to rule them all?

Press enter or click to view image in full size

Sauron prefers Windows AI