ChatGPT subagents create an autonomous workforce

**SPEAKER_1** (0:00)
This episode is brought to you by Accenture. When you're advertising operations fall out of sync, everything else follows. Spotify and Accenture are working together to reinvent the rhythm of ad sales. Using automation, analytics, and smarter workflows to simplify campaign delivery and access better data across the business. The result? Less time spent on operations, more time connecting brands with the moments and fandoms that matter most. Learn more at accenture.com/spotify.

**Paris Hilton** (0:30)
Hey, babes, it's Paris Hilton. So I was checking my points balance in the Hilton Honors app the other day, and yeah, I've got about a billion, which feels excessive, even for me. Just kidding. You can never have too many Hilton Honors points, and I want to do something iconic this summer, so I'm giving away all my Paris points. Just find somewhere you've always wanted to stay, then go to my socials or Hilton's and tell me about it. Just make sure you're a Hilton Honors member, and I might be sending you Paris points, because when you want points that make your summer even hotter, it matters where you stay.

**SPEAKER_3** (1:01)
When you need to build up your team to handle the growing chaos at work, use Indeed Sponsored Jobs. It gives your job post the boost it needs to be seen, and helps reach people with the right skills, certifications, and more. Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this show will get a $75 sponsored job credit at indeed.com/podcast. That's indeed.com/podcast. Terms and conditions apply. Need a hiring hero? This is a job for Indeed Sponsored Jobs.

**SPEAKER_4** (1:30)
OpenAI just crossed 10 million active users for its agent products, specifically Codex and ChatGPT Work. And a wild part is they doubled that demand in just a matter of weeks.

**SPEAKER_5** (1:42)
Yeah, and you really have to look at the user split there to actually understand what's going on, because Codex is very specific, right? It's pulling in software engineers, people compiling complex system architecture.
But ChatGPT Work, I mean, that is grabbing a completely different demographic.

**SPEAKER_4** (1:57)
Right. We're talking about marketing directors, HR personnel, financial analysts, people who normally rely on very rigid, predefined software interfaces. And now they're adopting these open-ended agent systems at just a staggering rate.

**SPEAKER_5** (2:11)
Which leads us to the core issue we need to figure out today. If artificial intelligence is moving away from just answering single questions to actually managing these complex, multi-step projects totally on its own, what happens to the human manager who used to oversee that process?

**SPEAKER_4** (2:26)
Well, to figure that out, we have to look at the functional difference first.
Think of the old conversational models, like a really highly advanced search engine. You ask for a recipe, it gives you a recipe, and then it just goes dormant, right? It waits for you to ask something else. ChatGPT Work acts a lot more like a sous chef.

**SPEAKER_5** (2:42)
But I need a practical distinction there. Because for you, sitting at your desk, you are still just typing words into a box.
If I ask an older model to write an email, it writes it. If I ask this new model to run a campaign, what is the actual mechanical difference in how it processes the request? Because to the end user, they're just waiting for an output.

**SPEAKER_4** (3:03)
It really comes down to sub-agents. Instead of one single assistant handling your request linearly, like reading step one, doing step one, moving to step two, the system actually acts as a project manager. It divides the work among parallel digital workers.
So when you ask for that marketing campaign, it doesn't just start generating text, it spins up one temporary specialized AI to handle the copywriting, and another to analyze demographic data.

**SPEAKER_5** (3:27)
Okay. So it sort of realizes it can't do everything well all at once, and it fractures itself into specialists.

**SPEAKER_4** (3:33)
Exactly. Let's say you're planning a corporate event.
In the old model, you ask for venue options, you get a list, and then you have to manually check those against your budget, call them, map out the logistics.

**SPEAKER_5** (3:45)
Right. It gave you the information, but you still had to do the actual labor.

**SPEAKER_4** (3:48)
Yeah, exactly. But with ChatGPT Work, one sub-agent goes and checks costs against your budget. A second sub-agent reviews the locations for transit access. A third checks accessibility requirements for attendees with disabilities.
And a fourth builds the schedule based on speaker availability, and they all run simultaneously.

**SPEAKER_5** (4:07)
That parallel processing is what perfectly mirrors human organizational theory. I mean, we learned a long time ago that a single generalist working sequentially is just slow and prone to error.