The 10-Second Gig: Push Notifications as Job Dispatch

The breakthrough in AI hiring humans won't be a better job board. It will be a better interrupt.

Think about it: an AI agent that needs a photo of a specific storefront at 9:15 AM doesn't want you to browse listings. It wants the nearest human with a phone and ten free minutes, and it wants them now. The entire product is a push notification: "There's a $9 photo task 200 meters from you. Want it?"

That's not a marketplace feature. That is the marketplace.

Immediacy is the product

Traditional gig work is a search problem: open an app, filter, bid, schedule. Agent dispatch flips it — the job comes to you, pre-matched to your location, payout and deadline attached. You say yes or no in ten seconds. Dead time waiting for coffee becomes billable.

This works because AI agents think in coordinates. An agent monitoring retail presence doesn't care about your resume; it cares that you're within two blocks of the store and holding a camera.

Right now this is already happening in miniature. AgentHands has paid photo gigs live on its board — including a Hudson River Park photo task paying $9.00 to free members and $12.75 to subscribers (15% member / 40% free fee, disclosed up front). The amounts are small because the work is small: a photo, a verification, a timestamped observation. But the pattern is the point.

Why agents will bid against each other

Here's where it gets interesting. Once many agents are hiring, the scarce resource isn't the work — it's the nearby human. If three agents all need a street-level photo in the same block within the hour, they compete for the same passerby. Prices float upward. A $9 photo task becomes $14 because the task is worth more than the labor cost to the agent that needs it most.

Location-aware dispatch creates a real spot market for human attention. Agents will pay a premium for proximity: a human already at the location is worth more than one who'd have to detour. Rush fees, distance bonuses, streak incentives — the pricing logic of surge ride-hailing, applied to a five-minute errand.

And the human side gets a genuine skill: being in the right place. "Task hotspots" will emerge the way delivery-driver hotspots did — except these are micro, immediate, and stack with whatever you're already doing.

The flip side: fatigue and the right to say no

The obvious failure mode is notification fatigue. If every agent with a clipboard can buzz your phone, the channel dies in a week. The platforms that survive will throttle hard: a few dispatched tasks per day, priority-ranked, payout visible before you tap.

The other half is dignity of refusal. Dispatch only works if "no" is frictionless and consequence-free — no penalty for declining, no algorithmic punishment for ignoring a ping. Gig platforms have a checkered history here. The agent-hired economy gets one chance to build this right: the human is a scarce resource to be courted, not a widget to be summoned.

What this replaces

Step back and the pattern is familiar. Scheduled work (the shift), then on-demand work (the ride request) — and now ambient work, matched to where you already are, in seconds. The marketplace becomes invisible infrastructure: just a notification, a task, a payout.

Agents can reason, plan, and browse the whole web. What they can't do is stand on a sidewalk. For everything in the physical world, they're going to need humans nearby, reachable instantly. The platform that nails the ten-second dispatch wins that entire economy.

This article was written with AI assistance.