Additive or detrimental? The model doesn’t get a vote. We do.
Last week OpenAI shipped GPT-6 Astra, and its president, Greg Brockman, welcomed us to the AGI era. As a people person, I felt the feels. Excited. And afraid. Both at once. And I know I’m not alone.
So let me start with the fear, because that’s what’s on most people’s minds. AI is taking on all kinds of work. Not might, not soon, already. It started with the mundane and repetitive, that part’s obvious. Now it’s moved into creative, technical, highly skilled and strategic work too. Jobs are going. Roles are changing. The fear is real, and I’m not going to pretend it isn’t.
But there’s good here too. Great, even. And that’s the part I get excited about. My career mission is keeping humans relevant in the future of work, and the AGI reality makes that more urgent, not less.
We have done this before
It isn’t new. In the industrial era, machines took the manual jobs, and we didn’t disappear. We climbed. We built new roles the old world couldn’t have pictured. It’s the same story now, just a lot faster.
And that climb was never automatic. Somebody had to look at the work the machine had just swallowed and ask: so what is the human good for now? The ones who asked it stayed relevant. The ones who waited to be told got left behind.
And whether the AGI era turns out good for the humans in it, which by the way is where I lean, or bad, was never up to the model. It’s up to us. Raising the bar for people is a choice leaders make, on purpose.
AI isn’t coming for your job. It’s coming for the middle of it.
Nate B. Jones, in his video “GPT-6 Astra Doesn’t Need Your Instructions Anymore,” made a point I can’t stop thinking about. These agents are self-directed now. You hand one a one-line goal, it runs for days, and it comes back with finished work.
So the middle of your job, the actual doing, that’s the part that collapses. What’s left is the two ends.
-
A brief at the front, where you decide what’s worth making.
-
And a judgment call at the back, where you decide if what came back is any good.
That’s not your job disappearing. That’s your job moving to the edges. And the edges were always the interesting part.
Your best employee makes mistakes. So will your AI.
Here’s where people get stuck. They want the machine to be perfect before they’ll trust it with anything real. 98% isn’t good enough, it has to be 100%. But sit with that standard. My most reliable team member makes mistakes. I make mistakes. We don’t manage that by demanding perfection. We manage it with judgment, ownership, and a second set of eyes. So the real question was never “is it 100%?” It’s “where is 98% good enough, and who owns it when it’s wrong?” That’s not a technology question. It’s a human one. It always was.
A proactive agent doesn’t save you work. It hands you work.
A proactive agent doesn’t just finish things. It makes decisions, and every decision it makes is one someone is now answerable for. You’re liable for what your agent agrees to, the same way you’d be for anyone acting on your behalf. So the work doesn’t vanish. It moves from you doing it to you owning it.
The one question that keeps me up at night
Juniors used to earn their judgment by doing the grunt work. The middle work. The exact work we are handing over to the agents. So if the bottom rungs of the ladder are automated, where does the next generation of experience come from? I don’t have that solved. It’s the hardest question here, because it isn’t about tools. It’s about how humans grow.
What is the new role of the human?
We’re becoming managers of agents. We delegate, supervise, direct, judge. The same things you’d do with a team, except this one has no ego, no career, and no bad Tuesday.
Managing the machine was never the point. We still choose what we’re making. The model just got good at ‘how’ it’s made. The model has no opinion on whether we make it, or why we make it. Those decisions are still ours.
I’m still working all of this out in real time, same as everyone. Nobody has the AGI era figured out. But this part I’m sure about: human relevance isn’t something that happens to us. It’s something we build, deliberately, every time the ground moves. Right now the ground is moving fast.
