This week OpenAI didn't just ship a new model, it shipped a new coworker. Live on stream, the company introduced ChatGPT Work, a brand-new agent built into the ChatGPT apps for web, mobile, and desktop, alongside its next flagship model family, GPT-5.6. Instead of one model, you now get tiers named Sol, Terra, and Luna, which OpenAI describes this way: “In this new naming system introduced with GPT‑5.6, the number identifies a model's generation, while Sol, Terra, and Luna identify durable capability tiers that can advance on their own cadence. Together, the family gives people and developers clearer choices across intelligence, speed, and cost.” Translation: this isn't a small update. It's a full re-architecture of how ChatGPT thinks, remembers, and now, works.

Here's the part that should make every working woman sit up: ChatGPT Work is a new agent that's available on the web, mobile, and desktop. Meanwhile Codex, OpenAI's coding agent, is quietly becoming the new ChatGPT desktop app itself, and OpenAI says GPT-5.4 will be retired on July 23 following the introduction of GPT-5.6. Fourteen days from full launch to full retirement of the old model. That's the pace we're operating at now.

Why does this matter specifically for women? Because "agent" is a polite word for software that doesn't just answer your questions, it does the task. Scheduling, drafting, coordinating, researching, customer-facing writing, the connective-tissue knowledge work that keeps offices running, and that women have historically been funneled into disproportionately. An agent that can log into your tools and just handle it isn't a productivity boost for that work. It's a replacement candidate for it.

Sound familiar? Three weeks ago in this space I wrote about the federal lawsuit against Workday and the ACLU's complaint over a HireVue screening tool, both proof that AI is already deciding who gets in the door before a human ever looks at a resume. Today's news is the same story wearing a different hat. First the algorithm decides if you get hired. Now the agent may decide whether the job exists at all once you're in it. Different products, same root problem: these systems are being designed by rooms that don't look like the workforce they're automating.

The Pattern I Keep Seeing

Here's my honest take: I'm not anti-agent. I use these tools daily and they're genuinely useful. What worries me is the speed and the split. OpenAI is already drawing a line between the people who see how the machine works and everyone else: ChatGPT Codex mode shows more technical details that ChatGPT Work abstracts away from the user. One side gets the engine. The other gets the dashboard. Guess which side historically gets more women funneled into it by default, and guess which side has more leverage, pay, and say in what gets built next.

So here's your concrete move this week: don't just use ChatGPT Work, go look under the hood. OpenAI's own developer community publishes every release note and technical change as it happens at community.openai.com, free to read, no invite needed. If you want structured skill-building rather than scrolling release notes solo, look at AI4ALL and Women in AI, two active nonprofits running fellowships, mentorship, and technical training specifically aimed at getting women into the rooms that design these systems, not just the rooms that use them. Fifteen minutes a week reading the actual model cards and release notes instead of just the app update is how you go from consumer to person-who-understands-what-just-shipped.

I'll say the quiet part plainly: if agentic tools like this scale the way OpenAI clearly intends, given they're retiring the previous model in under a month, some of the knowledge work that's been a stable, flexible career path for women is going to shrink or transform faster than most companies' HR policies can keep up with. That's not fearmongering, that's just reading the release notes. The honest response isn't panic, it's positioning: learn the tools well enough to build with them, not just operate them.

Have you tried ChatGPT Work or a similar AI agent at your job yet, and did it feel like a helpful assistant or a warning sign? Tell me in the comments, I read every one.