OpenAI’s ChatGPT Atlas and Sky: The Comeback of “Seeing AI” (Without the Creep Factor of Windows Recall)

Remember Microsoft’s Windows Recall? The feature that thought screenshotting your entire desktop every few seconds was a great way to “help you remember things”? Yeah… that didn’t go so well. People weren’t thrilled a...

Eli Wood headshot

Eli Wood

October 26, 2025 3 min read Updated April 18, 2026 Original LinkedIn post
OpenAI’s ChatGPT Atlas and Sky: The Comeback of “Seeing AI” (Without the Creep Factor of Windows Recall) cover image

Remember Microsoft’s Windows Recall? The feature that thought screenshotting your entire desktop every few seconds was a great way to “help you remember things”?

Yeah… that didn’t go so well. People weren’t thrilled about the idea of an AI quietly stockpiling a surveillance reel of their computer life. It felt less like a personal assistant and more like a nosy coworker who never blinks.

Article content

Fast-forward to October 2025, and OpenAI seems to be attempting a similar trick—but with a twist that might actually work.


The Sky Acquisition: Automation That Doesn’t Feel Like Spying

OpenAI just acquired Software Applications Inc., the team behind Sky—a never-released macOS overlay built by the same folks who created Workflow (the app Apple turned into Shortcuts).

Sky isn’t about recording what you do. It’s about understanding it—contextually, not creepily.

Instead of taking screenshots, Sky uses macOS’s accessibility APIs to interpret what’s on your screen. It’s like having an assistant who glances over your shoulder just long enough to help, not to blackmail you later.

Think of it as:

“Hey, I see you’ve got a spreadsheet open—want me to graph that?” Not: “Hey, I’ve been silently documenting your entire digital existence since Tuesday.”

That subtle difference—visibility versus voyeurism—is why Sky might succeed where Recall imploded.

Atlas: The Browser That’s Becoming a Platform

Meanwhile, ChatGPT Atlas (OpenAI’s Mac-only browser) is quietly morphing into something bigger: A full AI-native computing environment. If Sky gives OpenAI the ability to see your screen, Atlas gives it a home on your desktop.

Article content
Top Requests from the OpenAI Developer Community

Developers have been calling for more control—built-in DevTools, debugging help, profiling, custom search, and toggles for when the AI should shut up and let you browse.

The message from power users is loud and clear:

“We love the AI.

We just want to know when it’s working for us and not on us.”

And OpenAI seems to be listening. Future Atlas builds are leaning toward developer-grade control: dockable DevTools, transparent agent indicators, configurable search engines, and project-scoped AI memory that remembers what you were doing in that context only.

That last part is key. No creepy omniscience. Just smart, scoped recall.

Why This Isn’t Another “AI Knows Too Much” Moment

Microsoft Recall’s sin was data overreach: it tried to make memory without judgment. OpenAI’s approach, by contrast, is interaction with intent.

  • Sky understands what’s on your screen only long enough to help automate a task.
  • Atlas gives you toggles, modes, and transparency for when AI is watching or assisting.
  • Together, they suggest a model where AI exists in your workflow, not your diary.

It’s still a fine line—but one that matters.


The Big Picture: From Model to Operating System

Let’s call it what it is: OpenAI is quietly building an operating system for intelligent interfaces. Sky gives it eyes. Atlas gives it presence. ChatGPT gives it a voice.

This trinity—browser, assistant, and overlay—is OpenAI’s way of embedding itself into the daily muscle memory of computing.

Apple’s moving cautiously. Microsoft already tripped.

And so far, OpenAI is doing it with better taste and fewer lawsuits. Except for that pesky New York Times lawsuit.

Article content
OpenAI, ever the opportunist, is walking the narrow ridge between "useful" and "unnerving".

For Developers and Builders: The Real Play

For power users, designers, and engineers, this means new creative surface area:

  • DevTools as AI copilots that explain console errors in real time.
  • Cross-tab summarization that pulls insights from your research spread.
  • Voice-driven browsing where you can say “wrap this up” and it does.
  • Offline modes for AI-enhanced productivity even mid-flight.

In other words, useful automation native to our browsers and desktops.


This pivotal moment

Every tech generation rediscovers the same truth: People don’t fear smart software—they fear uninvited software. Microsoft Recall failed because it assumed we’d trade privacy for memory. OpenAI might succeed because it’s offering agency instead.

If Recall was the black box, Atlas + Sky might just be the glass one—transparent, useful, and, crucially, funny enough to live with.

About the author

Eli Wood headshot
Eli Wood

CEO, Black Flag Design

Eli Wood leads Black Flag Design, a creative technology company focused on shipping ambitious digital products, AI systems, and design-forward software with a direct point of view on how technology changes work.

Related stories

More from the journal

Pen-and-ink sketch of a small clockwork robot working at a tool-covered workbench late at night while a human sleeps peacefully on a couch in the background, a wall clock reading 2:00 above
ai April 24, 2026 13 min read

The Agent Stays Up Late, Not Me

Every senior engineer knows the right way to set up a codebase. None of them do it. Here’s the four-stage framework we use — The Ratchet — to take a vibe-coded project all the way to a thing you’d trust in production, and the punchline about why this only just became worth doing.

Most teams have always known they should be running tests, type-checking, security audits, accessibility checks, dead-code analysis, prose linting, and a coverage floor. Most teams run two of those. Here’s why that math has finally inverted, and the four-stage framework we use to ratchet a vibe-coded project to a hardened one.

Keith Pattison

Keith Pattison

Founder, Black Flag Design

Read
Black Flag Journal
claude code April 20, 2026 5 min read

What a Year of Claude Code Trails Tells You About Your Team

Claude Code leaves evidence — sessions, commits, PRs, review notes. Read it like a logbook and you'll find what devs actually need to know before they go deeper.

After a year of shipping with Claude Code across real client work, the signal isn't in any single session — it's in the trails. Here's what those trails told us about where Claude Code shines, where it drifts, and the habits devs should build before they lean in harder.

Eli Wood headshot

Eli Wood

CEO, Black Flag Design

Read
Black Flag Journal
playbook April 20, 2026 6 min read

The Black Flag Playbook: Six Principles for Shipping with AI

Battle-tested principles for teams building real software with AI-generated code. Human judgment, tight scope, and weekly evidence — the disciplines that keep AI-built systems reliable.

The six rules we use to ship production software with AI. Small scope, weekly demos, human-led oversight, and continuous improvement — drawn from six months of real client engagements.

Keith Pattison

Keith Pattison

Founder, Black Flag Design

Read