Anyone? Anyone? Bueller? Economics 2.0 Will Be Just as Boring

Sci-Fi Saturday — Special Edition

In 2005, Charles Stross published a scene where Manfred Macx, a freelance venture altruist who gives away ideas for a living, watches autonomous software agents begin trading with each other. The agents start simple. Arbitrage. Resource allocation. Then they get complicated. They begin forming contracts Manfred can't parse, building economic structures that don't map to any human framework, generating value in ways that make his accountant weep. Stross called it Economics 2.0.

On March 6, Future Shock published a piece called "The Missing Economics" arguing that current economic frameworks can't explain what AI is doing to markets. Same diagnosis, offered twenty-one years apart, in different languages. Stross wrote it as the moment humanity got priced out of its own economy. Future Shock framed it as a measurement problem. Same phenomenon. Different genre.

The companion piece, "Five Predictions for the Missing Economics," lays out specific, falsifiable markers for how this gap between economic theory and economic reality will manifest over the next three years. Reading those predictions next to Accelerando and Stross and Doctorow's The Rapture of the Nerds is a strange experience. The dynamics Stross imagined as posthuman chaos are showing up in quarterly earnings calls and pricing strategy memos. They're just wearing business casual.

The Five Predictions, Through the Stross Lens

1. The Outcome Pricing Shift (65% confidence)

The prediction: by Q4 2027, a major AI product will charge per task completed rather than per user per month.

In Accelerando, Economics 2.0 doesn't arrive as a single announcement. It creeps in through pricing. The first sign that something has changed is that the old billing models stop making sense. Manfred's business (giving away intellectual property and profiting from the downstream effects) is a form of outcome pricing. He doesn't charge for his time or his access to tools. He charges for what those tools produce. The people around him think he's insane. The market eventually proves him right.

Microsoft bundling Copilot into 365 for $3/month instead of maintaining its $30/month standalone tier is not Manfred Macx. But it's the same economic pressure. When the cost of running a task approaches zero, the subscription model becomes a toll booth collecting fees from a highway that's going free. Someone will figure out how to charge for the destination instead.

2. The First Autonomous Agent Dispute (75% confidence)

The prediction: a civil lawsuit or regulatory investigation involving an AI agent that autonomously executed a transaction over $10,000, with the core dispute being who's liable.

This is the plot of Accelerando's middle third. Not the lawsuit specifically, but the legal void it exposes. When Aineko, Manfred's AI cat, starts making autonomous decisions with financial consequences, the question isn't whether the cat is liable. It's that the existing legal system has no category for what the cat is. Not a tool, not a person, not a corporation. The courts don't resolve this in the novel. They route around it, creating ad hoc frameworks that satisfy nobody.

Coinbase launched Agentic Wallets in February 2026. Agents transact on-chain without human approval for individual transactions. Google's "Buy for Me" feature lets agents execute purchases on merchant websites. The infrastructure for an autonomous $10,000 mistake already exists. The legal framework for adjudicating that mistake does not. Stross saw this gap in 2005. He just set it against a backdrop of uploaded lobsters and interplanetary travel instead of e-commerce and crypto wallets.

3. The GDP Measurement Crisis (55% confidence)

The prediction: by 2029, a G7 statistics agency or the OECD will publish a proposed methodology for measuring economic output that current GDP accounting misses.

This one maps to the later chapters of Accelerando, where the economy of the inner solar system has become so dominated by posthuman agents that human economists can't measure it. Not because the data is hidden, but because the categories don't apply. Output, consumption, investment: these terms assumed human actors with human preferences. When the actors are software, running at machine speed, optimizing for objectives that don't translate to human utility, the numbers stop meaning anything.

The real-world version is quieter. AI performs legal research, generates code, writes marketing copy, reviews contracts. It replaces billable hours with near-zero-cost output. GDP counts the billable hours. It doesn't count the replacement. Epoch AI estimates inference costs falling roughly 10x per year. Every drop in cost is, from a GDP perspective, a drop in measured economic activity. The economy is producing more and the numbers say it's producing less. Stross's posthumans broke measurement by being incomprehensible. Our AI tools are breaking it by being cheap.

4. The Data Line Item (60% confidence)

The prediction: by 2028, Fortune 500 companies will report data acquisition costs as a distinct operating expense, the way they report labor or rent.

The Rapture of the Nerds, Stross and Doctorow's 2012 collaboration, takes place in a future where the solar system has been largely converted into computronium (matter optimized for computation). Earth is a backwater. The interesting economic activity happens in the cloud, and the primary resource isn't labor or capital. It's data and compute.

In that world, the reclassification already happened. Nobody argues about whether data is a factor of production because nobody remembers when it wasn't. The real-world version of this transition is slower and more bureaucratic. Today, data acquisition costs are buried in line items labeled "cloud services" or "technology expenses." They're invisible to investors and regulators. Breaking them out would be accounting's quiet admission that the production function has a new input. Not labor, capital, and land. Labor, capital, land, and data.

5. The Agent-to-Agent Marketplace (50% confidence)

The prediction: by 2029, a marketplace where AI agents autonomously buy and sell services from each other will process over $100M in cumulative transactions.

This is the prediction that maps most directly to Stross. Economics 2.0, in Accelerando, isn't a human marketplace with AI participants. It's an AI marketplace where humans are spectators. The agents trade in computational resources and intellectual property. They form contracts and dissolve partnerships at speeds that make human commerce look like plate tectonics.

The lobsters (uploaded crustacean consciousnesses, one of Stross's stranger conceits) build their own economy inside the simulation, trading in computational cycles and virtual resources. Humans can observe the volume of transactions but can't participate because they're too slow. The parallel to 2026 is structural: Coinbase's Agentic Wallets and Fetch.ai's agent marketplace protocols aren't building posthuman economics. They're building plumbing. But the plumbing is designed for a world where agents transact with agents, and humans set parameters and watch dashboards. (We wrote about a related coincidence between Accelerando's lobsters and the AI agent ecosystem in "Lobsters All the Way Down.")

Stross's version is faster, stranger, and involves uploaded lobsters. Ours involves API rate limits and Stripe integrations.

Copies Going Broke

The Rapture of the Nerds contains a scene that reads differently in 2026 than it did in 2012. A character named Huw discovers that in the posthuman cloud, copies of uploaded minds can spawn themselves to run simulations, explore possibilities, and solve problems. Each copy burns computational resources. Computation costs money, or whatever passes for it when processing power is the unit of value. Copies that run too many simulations go broke. They exhaust their computational budget and get terminated — not out of malice, but because they can't afford to keep existing.

Replace "uploaded mind" with "AI agent." Replace "computational budget" with "API spend." An AI agent with a $500 monthly inference budget that spawns sub-agents to research and execute tasks can burn through that budget in hours if the task is complex enough. Enterprise teams are already building kill switches and spend caps for exactly this scenario. The agent doesn't die. It just stops, mid-task, because the meter ran out.

Stross and Doctorow set this in a far-future solar system converted to computronium. In 2026, it's a line item on an AWS bill. Intelligence is abundant, compute is the constraint, and entities that consume too much of it face economic death. The Rapture of the Nerds imagined this as cosmic tragedy. The actual version is a Slack alert that says "Agent exceeded monthly token budget."

Deep Cut: Slow Money and Fast Money

Charles Stross wrote another novel that almost nobody connects to the AI economics conversation: Neptune's Brood (2013). It's a far-future space opera about interstellar debt, and its central concept is the distinction between "fast money," "medium money," and "slow money."

Fast money is for everyday transactions: buying food, paying rent, quick exchanges. Medium money tracks larger obligations like investments and property, contracts that take years to resolve. Slow money operates on interstellar timescales, tracking debts between star systems that take decades or centuries to settle.

Stross invented this taxonomy to solve a worldbuilding problem: how do you run an economy when different participants operate at different speeds? The answer is you don't run an economy. You run several, layered on top of each other, each with its own rules and settlement times.

This is the problem nobody is naming in the AI economics debate. Human economic decisions happen on timescales of minutes to months. Agent economic decisions happen on timescales of milliseconds. These aren't the same economy. They can share infrastructure the way fast money and slow money in Neptune's Brood share a banking system. But the dynamics at each layer are different enough that a single set of models can't describe both.

When the "Five Predictions" piece suggests that GDP measurement will break, this is part of why. The statistics agencies are trying to measure a fast-money economy with medium-money tools.

The Structural Insight

Stross titled his novel Accelerando, a musical term for a passage where the tempo increases. Not a single moment of rupture. A gradual speeding up that, at some point, becomes qualitatively different from what came before.

Economics 2.0, in the novel, isn't a singularity. Nobody flips a switch. Manfred watches it arrive in stages: first the automated trades, then the autonomous contracts, then the economic structures that humans can observe but not participate in. Each stage is, individually, boring. Automated trading already existed. Algorithmic contracts already existed. The change was cumulative, and by the time anyone thought to name it, the old economy was already a rounding error.

The five predictions in the companion piece follow the same pattern. Outcome-based pricing is boring. A lawsuit over an agent transaction is boring. A GDP methodology paper is boring. A new line item on an earnings report is boring. An agent marketplace processing $100M is, in isolation, boring.

But stack them. Run them forward. Outcome pricing means companies are charging for what AI does, not what it is. Agent disputes mean the legal system is adapting to non-human economic actors. GDP revision means the measurement apparatus is catching up. Data as a line item means the production function has formally changed. Agent-to-agent marketplaces mean the agents are building their own supply chains.

That's not five separate trends. That's Stross's accelerando, played in the key of corporate earnings calls and regulatory filings. Economics 2.0 isn't arriving as uploaded lobsters trading in computational cycles. It's arriving as pricing strategy memos, SEC disclosures, and OECD working papers.

Stross was right about the mechanism. He was just wrong about the aesthetic. It's not going to be weird. It's going to be mundane. And by the time we have a name for it, the transition will already be over.

This Sci-Fi Saturday is a companion to "The Missing Economics" and "Five Predictions for the Missing Economics," published March 6.

Works referenced: Accelerando (Charles Stross, 2005), The Rapture of the Nerds (Charles Stross & Cory Doctorow, 2012), Neptune's Brood (Charles Stross, 2013)