Stone Soup in the AI Age

Artificial intelligence is no longer a white paper or a demo. With recent developments in the field of generative AI, code is scaffolded in minutes. Research papers are drafted in a day. At the same time, the predicted extinction of knowledge work has not arrived. The lights are still on in the offices. Late night teams meetings are still happening. People are still debating about work-life balance.

But the math has changed.

Turnaround times are shrinking. Expectations are rising. What once commanded a premium is now treated as standard. If your pitch doesn't start with "AI-enabled," you’re not even in the room. All deal negotiations start with the assumption that AI could do it faster and cheaper.

To understand this moment clearly, two old metaphors are useful.

• Stone Soup
• The Boiling Frog

We are living both at the same time.

Stone Soup

Stone Soup is a European folk story. I first encountered the story of Stone Soup in The Pragmatic Programmer. It remains one of the most practical books on software craftsmanship.

The story is simple. Hungry travelers arrive at a village where people are reluctant to share food. The travelers place a stone in a pot of water and begin to boil it. Curious villagers gather. The travelers remark that the soup would improve with carrots. Someone brings carrots. Then onions. Then herbs. Then meat. By the end, there is a feast.

My takeaways from the story.

• Catalyst is important: The travelers did not demand contribution; they created a starting point that encouraged the villagers to participate.

• Show and do not tell: When you tell people, they have a choice of believing or not. When you show people, it becomes harder to dismiss.

• It is a feast: The villagers participated, and everyone benefited from the outcome.

AI today resembles the beginning of that story. Large model providers have placed an extraordinary stone in the pot. Foundation models. Massive compute. General reasoning capability.

But they cannot cook the entire meal.

No foundation lab understands the nuance of insurance underwriting embedded in decades of claims data. No central AI team carries a law firm's private litigation strategy. No general model contains the operational edge cases buried inside a manufacturing plant or a regulated fintech workflow.

Domain expertise is contextual and proprietary. It is embedded in private data, internal documents, systems of record, workflow histories, and institutional memory. Intelligence is horizontal, whereas domain expertise is vertical. Horizontal capability scales broadly. Vertical capability creates defensible value.

If a handful of firms attempt to build intelligence for every vertical, we do not get ecosystem strength. We get a concentration of power.

A healthier outcome requires separating layers.

Foundation models provide the base capability. Industries build specialised systems around their private data, domain logic, and regulatory realities. Until that happens, we are not building collective intelligence. We are centralising it.

The question is not whether the stone is powerful. It is whether the village brings its own ingredients.

The Boiling Frog

The boiling frog metaphor appears frequently in discussions about organisational decline. The story suggests that if a frog is placed in boiling water, it will attempt to escape, but if the water is heated gradually, it will adapt to each small increase until it can no longer react.

Whether or not the biology holds is beside the point. The psychology does.

AI is not eliminating knowledge workers overnight. It is narrowing the economic distance between expertise and automation. The value of writing code will decline.

We’re moving from being 'writers' to being 'editors-in-chief' of our own codebases.

The water is already simmering. We aren't seeing mass layoffs; we’re seeing leverage compression. This isn't about extinction. It's about repricing.

A task that used to justify a $10k billing cycle now takes four hours. The client knows it. You know it. But because you’re still "employed," it feels survivable. Because it feels survivable, repositioning is deferred.

How do I see it?

Technological shifts rarely eliminate work.

The iPhone succeeded because the App Store enabled thousands of developers to create vertical value. Android provided the base capability. The ecosystem determined the distribution of innovation.

Foundational models are powerful. They act like stones, as catalysts. Villagers are curious. But if the travelers are unwilling to share and are willing only to sell, they lose participation. The stone remains just a stone. The soup never becomes a feast. The evolution stops. A healthy future requires the village to bring its own ingredients, its own private data and domain logic, rather than just waiting to be sold.

Stone Soup reminds us that value emerges when industries contribute their expertise and private data rather than outsourcing intelligence entirely.

The economic layer of knowledge work is shifting. Execution becomes cheaper. Speed becomes assumed. What remains scarce is judgment and interpretation. The advantage shifts from producing output to defining direction, constraints, and meaning.

• Software engineers will not disappear, but the value of writing code will decline relative to designing systems and evaluating trade-offs.
• Lawyers will not vanish, but drafting will matter less than interpretation and strategy.
• Consultants will not be replaced, but gathering information will matter less than clarifying decisions.

The Boiling Frog reminds us to act before we can no longer react.

The future of knowledge work will not be determined by who builds the largest model. It will be shaped by who builds around it and who moves upward with it.

The pot is already on the stove. The stone is already in the water. What remains undecided is whether the village participates or simply adjust.