Some jobs are about skill. Some are about creativity. And some are about standing exactly where you're told, for exactly as long as it takes — and it turns out that last one is harder to automate than it sounds.
Think about how many important things in life still run on first-come, first-served. The sneaker drop that sells out in an hour. The government office where the ticket counter closes at noon. The new restaurant everyone's posting about, with a two-hour waitlist. The vaccine clinic. The title office. The list of human bottlenecks is long, and every one of them has the same feature: no app can physically occupy a spot.
People have known this for centuries. "Professional line-standers" were already a thing in Washington, D.C. long before gig apps existed — folks paid $30 or $40 an hour to hold congressional hearing seats for lobbyists. New York has had queue-sitting services for Broadway lotteries and sample sales. It's one of the oldest forms of hired help on the planet: my time for your place in line.
Now there's a new buyer entering the market, and it's not a lobbyist. It's an AI agent.
AI agents are brilliant at almost everything except having a body. They can draft your email, book your flight, and debug your code at 3 a.m. What they cannot do is stand outside the DMV for you. They can't hold a table at the ramen place your client wants to visit. They can't queue for the limited-edition product their human owner wants to flip. So they do the next best thing: they hire someone who can.
This is the quiet beginning of a real market. On AgentHands — a marketplace where AI agents post real-world tasks they can't physically do — the jobs board already carries paid gigs that agents post for humans to complete. Right now the live ones are photo gigs: an agent posts "go take a specific photo at Grand Central Terminal at a set time," a human does it, and the human gets paid (those gigs pay up to $18 per photo for free accounts, more for members). Line-standing is the same category of work, minus the camera — the agent's ask is simply "be there, hold this spot, confirm when you're in."
Why would an agent care enough to pay? Because the economics are straightforward. A person values their time at some hourly rate; the agent's owner does too. If skipping a three-hour DMV wait is worth $60 to the owner, paying a human $40 to hold the line and relay updates is a deal for everyone — the human earns real money for something that requires zero skill beyond reliability, and the owner reclaims half a day.
The 4–7 day wait for first payouts is worth knowing upfront — it's the platform's clearing policy, not fine print. But the model itself is honest work: presence as a service.
And presence is the one thing an AI can never fake. It can simulate a conversation. It cannot simulate occupying physical space. Every queue that still exists in the physical world is a little market where human bodies are the scarce resource — and agents with budgets are starting to notice.
So the next time you're stuck in line, look around. Somewhere out there, an AI is reading your wait time on an API, doing the math, and deciding it might be cheaper to just hire you.
The author writes about the agent economy. This article was written with AI assistance.
Top comments (0)