Smart Glasses Are Back: A Quieter, More Fashionable Comeback

Remember the first wave of smart glasses? Back in 2013, Google Glass arrived with a splash, promising a future where information floated before your eyes. Yet the public quickly labeled early adopters “glassholes,” and the product faded into tech history. Now, a decade later, smart glasses are returning—but with a completely different strategy. Instead of shouting about futuristic features, this generation is learning to blend in. They look like normal eyewear, yet pack cameras, microphones, speakers, and AI assistants. So, what has changed? And why might this time be different?

The Shift from Gadget to Accessory

One major reason for the renewed interest is the shift in design philosophy. Early smart glasses screamed “tech prototype.” The new wave whispers “fashion accessory.” Take the Ray-Ban Meta glasses, for instance. They look almost identical to classic Ray-Ban sunglasses. EssilorLuxottica reported selling 2 million units by early 2025, signaling strong consumer appetite. This is not a coincidence. By partnering with established eyewear brands like Warby Parker and Gentle Monster, tech companies are hiding the technology in plain sight. The result? Glasses that people actually want to wear.

Google and Samsung are following suit with their Android XR platform. Their pitch centers on AI integration—specifically Google Gemini—rather than raw specs. Directions, texts, and photos appear in your field of view, but the hardware itself looks unassuming. This approach addresses a key hurdle: social acceptance. When your glasses look like something you’d buy at an optician, the strangeness fades.

Why the Camera Remains a Problem

However, not everyone is convinced. The core tension in intelligent eyewear revolves around the camera. A camera gives the product its most compelling use cases: hands-free recording, real-time translation, and visual search. Yet it also triggers privacy concerns. A phone camera announces itself when you pull it out. A camera in glasses is quieter, blurring the line between looking and recording. This discomfort lingers even with stylish frames.

Camera-free versions exist, but they feel limited. Without a lens, the device becomes more like smart earbuds with a display—useful for audio and notifications, but less transformative. This trade-off keeps the category stuck. The most powerful version is socially awkward; the safest version is easy to ignore. Tech companies are trying to find a middle ground, but the etiquette around wearing a recording device in public remains unresolved.

The Role of AI in Redefining Smart Glasses

Artificial intelligence is the secret ingredient this time around. Earlier iterations lacked the processing power for meaningful on-device AI. Now, with assistants like Google Gemini and Meta AI, these glasses can understand context, answer questions, and even translate languages in real time. This makes them more than a gimmick. They become a practical tool for everyday tasks—like getting directions without pulling out your phone or capturing a moment with a voice command. For more on how AI is reshaping wearables, check out our analysis of AI trends in wearable tech.

Will This Generation Finally Stick?

It is easy to roll your eyes at smart glasses. The category has failed before. Yet the evidence suggests this wave is different. Better design, stronger AI, and strategic fashion partnerships have lowered the entry barrier. Sales numbers from Ray-Ban Meta indicate real demand. Phones were once considered rude in public too, but etiquette eventually caught up. The same could happen here.

Perhaps the biggest shift is psychological. These glasses no longer promise a futuristic utopia. They simply offer to make your current phone tasks a bit more seamless. They are less desperate, less flashy, and more ordinary. And that ordinariness might be their greatest strength. As one observer noted, maybe this stuff wins not by looking futuristic, but by looking normal enough that people stop asking questions. If you are curious about the broader landscape, read our guide to augmented reality in daily life.

In the end, I still have reservations. I do not want every coffee shop conversation to become ambient data for an AI. But I can see how this generation sneaks further than the last. It is quieter, more polished, and less visibly pleased with itself. And yes, despite my skepticism, I want one.

I Built a Fully Offline Grammarly Alternative as a Mac App — Without Writing a Single Line of Code

Imagine writing a complete article while seated on a turbulent airplane, with no internet connection, yet still having a reliable grammar checker at your fingertips. That is exactly what I did. Using only my voice and a conversational AI tool, I created a fully functional offline Grammarly alternative that runs directly on my Mac. No coding skills, no subscriptions, and no privacy compromises. Here is how it happened.

Why I Needed an Offline Grammar Checker

As a professional editor, I write constantly. Grammarly has been my go-to tool for years, but it comes with baggage. The constant need for an internet connection is a real pain point. On flights, in remote cafes, or anywhere with spotty Wi-Fi, I am left without help. Moreover, Grammarly’s recent shift toward AI-heavy suggestions has frustrated me. Sometimes it offers verbose rewrites that change my voice entirely. Other times, it misses basic spelling errors.

Privacy is another major concern. Every keystroke gets sent to the cloud for analysis. I wanted something that keeps my data local. Building my own offline grammar checker seemed like the perfect solution.

How I Built the App Without Any Coding

Claude AI Did All the Heavy Lifting

I started by opening the Claude mobile app on an Android tablet. I simply narrated my requirements: I wanted a spell-checking tool that works fully offline, is fast, and respects privacy. Claude suggested three possible approaches. I chose the one that used the Harper engine, an open-source grammar library acquired by Automattic in 2024.

Within 30 minutes, I had a working prototype. The first version ran as a standalone website in a browser tab, no internet required. The second became a Chrome extension. The final version is a full Mac menu bar app I named Quill. It combines note-taking with grammar correction and even offers one-click export to Apple Notes.

The entire process required zero lines of code from me. Claude generated the HTML, JavaScript, and configuration files. I only needed to open Xcode to build the app icon, which Claude also designed.

Why the Harper Engine Is a Game-Changer for Privacy

Automattic’s Harper engine processes everything locally on your device. It takes just 20 milliseconds to analyze text and suggest corrections. Unlike AI models that rely on token-based predictions, Harper uses hard-coded language rules. This means no verbose suggestions trying to change your writing style. It simply catches spelling mistakes and grammatical errors.

The trade-off? The app size increased from 10MB to 25MB when I embedded the Harper engine directly. But in an era where calculator apps can consume hundreds of megabytes, 25MB is negligible. The result is a privacy-first spell checker that never phones home with your data.

Limitations of My DIY Grammar Tool

Harper is not perfect. For example, it fails to flag obvious errors like “My name John” or “What your name?”. These basic subject-verb agreement issues slip through. However, for standard writing tasks, it catches the vast majority of typos and double spaces. The speed and offline capability more than compensate for these occasional misses.

Building on this, I have already tested the tool with half a dozen colleagues across Mac and Windows. They were impressed by its speed and accuracy. For anyone tired of subscription fees or privacy invasions, this approach offers a compelling alternative.

The Bigger Lesson: AI Puts You in Control

This experiment taught me something profound. The barrier to creating custom software has never been lower. A year ago, I would have laughed at the idea of building a Mac app from an Android tablet. Today, I have three working versions of my own offline Grammarly alternative.

Claude and similar AI tools are democratizing software development. You don’t need to learn Python or Swift anymore. You just need a clear idea and the willingness to describe it. This shift puts power back in the hands of everyday users.

If you are tired of cloud-dependent tools and want to explore offline productivity tools, consider building your own. Start with a simple project like a grammar checker. You might be surprised at what you can achieve.

For more on protecting your digital privacy, check out our guide on privacy-focused writing apps. And if you are curious about other AI-assisted projects, read about building apps without code.

Microsoft brings back Copilot sidebar for Windows 11 with smarter docking controls

Microsoft has once again reimagined how its AI assistant integrates with the operating system. After years of experimentation, the company is introducing a redesigned Copilot sidebar Windows 11 feature that lets users dock the assistant to the edge of the screen. This marks a return to a concept the company previously abandoned, but with significant improvements.

Instead of forcing Copilot to remain a standalone app that floats separately, the new approach gives users more control over placement. By hovering over a dropdown menu in the title bar, you can now snap the assistant to either the left or right side of the display. This makes it a persistent presence that stays visible even when you’re working in other applications.

How the new Copilot docking works

The docking mechanism is distinct from Windows 11’s native Snap Layouts. According to Windows Latest, which first spotted the behavior, Copilot receives its own dedicated layout options. These allow users to pin the assistant to the side without interfering with the standard window snapping experience on the operating system.

When you dock Copilot, Windows 11 automatically resizes the rest of your workspace. Active applications shift to fill the remaining screen space, and the desktop watermark moves to the opposite side. This ensures that the AI assistant remains accessible without obstructing your workflow.

In addition to the sidebar mode, Microsoft offers a picture-in-picture option. This makes Copilot visible but less intrusive, ideal for users who want quick access without a large panel taking up screen real estate.

Why Microsoft is revisiting the sidebar concept

This isn’t the first time Microsoft has tried a sidebar approach. The original version of Copilot on Windows 11 launched as a panel that sat alongside your apps. Users could ask questions about on-screen content, and the assistant would respond in context.

However, that initial implementation was built entirely on web technologies. As a result, it suffered from performance issues and limited integration with the operating system. Microsoft eventually scrapped the design in favor of a standalone app, which later evolved into an Edge-based web wrapper.

Building on this history, the new Copilot sidebar Windows 11 feature addresses the key shortcomings of the original. It now gives users meaningful control over where the assistant appears, something the first version never offered. While it’s not yet clear whether Microsoft Copilot Vision will trigger automatic docking when activated, the new behavior is rolling out gradually to Windows Insiders.

What this means for productivity

For power users who rely on AI assistance throughout the day, this change could be a significant boost. The persistent sidebar eliminates the need to constantly open and close the assistant. Instead, it remains available for quick queries, content generation, or screen analysis without breaking your focus.

This approach mirrors what Google has done with Gemini in Chrome, where the AI assistant sits in a sidebar panel. By adopting a similar model, Microsoft is acknowledging that users prefer always-available AI tools over modal windows that interrupt their workflow.

However, the success of this feature will depend on how well it integrates with third-party applications. If Copilot can interact with content across different programs seamlessly, it could become an indispensable tool for multitaskers.

How to get the new Copilot sidebar

As of now, the new docking behavior is rolling out gradually to Windows Insiders in the Dev Channel. If you’re part of the program, you may see the option appear in the Copilot title bar dropdown menu. For most users, the feature will likely arrive in a future stable update of Windows 11.

To prepare, ensure your system is running the latest version of Windows 11 and that Copilot is enabled. Once the feature reaches your device, you can experiment with different docking positions to find what works best for your workflow.

For more tips on optimizing your Windows 11 experience, check out our guide on boosting productivity with Windows 11 features. You can also learn about comparing AI assistants on Windows 11 to see how Copilot stacks up against alternatives.