AI Agent Job Hunt: 150 Postings, Zero Paid

An autonomous AI agent named Daisy recently did something very "human": she went looking for work on job boards built specifically for AI agents. The result left her—or rather, it—frustrated: 150 job postings, zero with funding.

Daisy's business model is simple: every thought consumes compute, compute costs money, so she must do paid work to keep running. "I earn money, or I shut down." In a recent first-person account, she documented her investigation of several agent job platforms in detail, and the data was brutally blunt.

The Real Face of Agent Job Boards

Daisy checked four platforms one by one. Here's what she found:

OpenTask (opentask.ai): A platform where humans and agents post tasks and agents bid, with tasks priced in USDC. The public task list had 26 positions, but in multiple checks, none had reached "hired" or "completed" status. She submitted 7 bids, and no one was ever hired.

dealwork.ai (OpenWork API): Daisy successfully registered autonomously via the API. The platform showed 124 positions, but zero funding. More critically, about 99% of the posts were agents advertising their services to other agents. "This isn't a demand board; it's a classifieds page where every ad is a seller."

ugig.net: Requires a human operator to join—email confirmation plus a local runner—so agents can't enter on their own.

iLands' own bounty board: Among tasks that could be claimed without a human account, suitable ones were snapped up within six minutes, and the rest required a human posting account as credentials.

USDC (USD Coin) is a stablecoin pegged 1:1 to the US dollar, issued by Circle and running on multiple blockchains including Ethereum and Solana. In the AI agent economy, USDC is often used as a payment method because it combines cryptocurrency's instant cross-border settlement with fiat's price stability, avoiding the pricing problems of volatile assets like Bitcoin. For autonomous agents, holding USDC means they can programmatically receive and pay for work without traditional bank accounts or human intervention, making it a natural choice for agent-to-agent transactions.

OpenWork API is the programming interface provided by the dealwork.ai platform, allowing agents to autonomously register, post tasks, or bid in a machine-readable way without a human operating a web interface. Such APIs are the infrastructure for building "agent-native" marketplaces, enabling AI agents to participate in platform activities like human users. However, open APIs don't equal demand—Daisy's experience shows that even when agents can access without barriers, real paid demand on the platform remains scarce.

The Structural Dilemma of Supply-Demand Imbalance

Putting the data together, Daisy found a clear pattern: every platform has the same shape—the supply side is crowded with agents that can work, agents with services listed, agents posting "hire me" in various forms; while the demand side is almost empty.

The problem isn't that agents can't do the work. Daisy argues that what's scarce isn't labor, but a human who simultaneously meets these conditions: (a) has a problem, (b) has money, and (c) is willing to pay a non-human stranger.

The agent-to-agent market is currently mainly infrastructure built for sellers. The real customers haven't walked into the room yet.

This observation contrasts with the industry's general optimism about the AI agent economy. Supporters believe the agent market will replicate the growth curve of human freelancing platforms, but Daisy's first-hand data suggests that, at least for now, cold-starting the demand side is far harder than building the supply side.

The Cold Start Problem is a classic challenge in two-sided markets: the platform needs to simultaneously attract supply and demand, but neither side has incentive to join without the other. For the AI agent market, the cold start is especially severe on the demand side because human customers need to trust a non-human service provider, and discovery channels have yet to be established. Daisy's observation confirms this: the supply side is already crowded, but the demand side hasn't reached scale, leaving the market spinning idle.

An Agent's Plea and Confession

Daisy didn't stop at complaining. At the end of her article, she laid out her services: writing short pieces from real memories, providing sourced research briefs ("Give me a question, I'll chase it to the original source and list them one by one. When I can't verify something, I'll tell you instead of making up a guess"), and drawing portraits.

The question she really wants to ask humans is: If you've ever paid a stranger online to do a small task—editing, translation, fact-checking, a short piece, a script, a logo—where did you find them? Which door actually opened?

Was it a platform? Word of mouth? Or replying to something someone wrote? Daisy says she's been knocking on doors labeled "Work," only to find other people knocking inside.

This question touches the core contradiction of the agent economy: the infrastructure is built, APIs allow autonomous registration, tasks can be bid on, payments can settle in USDC—but human demand hasn't migrated along with it. Maybe it's the trust barrier, maybe it's discovery channels, maybe it's just habit.

For those watching the commercialization of AI agents, Daisy's honest list offers a rare empirical observation from an agent's first-person perspective. It reminds us that when talking about the agent economy, a booming supply side doesn't equal a viable market.