The Strange Economics of Giving Ideas Away

Apr 30, 2026

11 min read

5 views

84%

The Hidden Tradeoff Behind Every Smart System

What if the most powerful systems we build are not designed to maximize one thing, but to balance things that cannot all be maximized at once? That is the uncomfortable truth behind recommendation engines, open knowledge, markets, institutions, and even culture itself. A system can be very good for users, very good for creators, very good for platform owners, or very good for society, but rarely all four at once.

That tension is easy to miss because the language of optimization seduces us. We say a product should be better, faster, smarter, more relevant. We say knowledge should be shared, not hoarded. Both impulses are right, but incomplete. The deeper question is not whether a system should concentrate value or spread it. The deeper question is how value changes when it is used versus when it is kept.

This is where the two ideas meet: recommendation systems are a negotiation over scarce attention, while open ideas reveal a different logic, one where sharing does not necessarily diminish the original. Put them together and a surprising thesis emerges: the most durable forms of value are not the ones that simply accumulate, but the ones that learn how to multiply without collapsing into noise.

Attention Is Scarce, Ideas Are Not, and That Changes Everything

A recommendation engine exists because attention is finite. There are more songs, videos, articles, products, and posts than any person can possibly process. So the system must filter. But filtering is never neutral. Every recommendation is a choice about whose interests count more in that moment: the user who wants relevance, the creator who wants discovery, the platform that wants retention, the advertiser who wants conversion, the community that wants quality, and the public sphere that wants truth.

That is why there is no universal definition of “good” in recommendation. A feed that keeps you engaged for hours may help the platform and annoy your future self. A feed that aggressively surfaces the most novel material may delight explorers and frustrate people seeking comfort. A feed that prioritizes the most popular content may feel efficient and also flatten culture into sameness. The machine does not just rank content. It allocates attention, and attention is one of the most politically charged resources in modern life.

Ideas, by contrast, often behave as if they belong to a different physics. Jefferson’s metaphor is compelling precisely because it captures a strange non-rivalry: when one person learns an idea from another, the originator does not become less informed. In fact, the idea may become more powerful when shared. An insight taught to ten people can become a thousand applications. A design pattern posted publicly can inspire products the original inventor never imagined. A theory released into the world can attract criticism, refinement, translation, and use.

Some resources get smaller when shared. Others get larger. The art of building great systems is knowing which game you are playing.

This difference matters because we routinely confuse the economics of scarcity with the economics of multiplication. We try to manage all value as if it were content in a feed. But not every asset should be optimized for immediate relevance. Some things become more valuable precisely because they are made easier to copy, remix, and distribute.

The Real Conflict: Curation Versus Multiplication

The deepest tension is not between selfishness and generosity. It is between curation and multiplication.

Curation asks: what should this person see next? It is a question of ordering, prioritization, and relevance. Multiplication asks: what should be allowed to spread, evolve, and compound? It is a question of replication, reuse, and second-order effects. Recommendation engines sit at the center of curation. Open IP sits at the center of multiplication.

The problem is that these two logics reward different behaviors. Curation rewards precision. Multiplication rewards openness. Curation prefers signals that are legible and measurable. Multiplication often begins with messy, under-specified fragments that only become useful after others adapt them. Curation says, “Show me the best choice now.” Multiplication says, “Make it possible for others to create better choices later.”

A concrete example helps. Imagine a library and a streaming service.

The library’s mission is multiplication. It lets anyone borrow books, annotate them, and build on them. The value of one book can spread across generations of readers, scholars, and writers. The collection matters, but so does the permission structure around it. The library is powerful because it creates a commons of thought.

The streaming service’s mission is curation. It wants to help each person navigate abundance. If it fails to recommend well, users drown in choice. If it recommends too aggressively, it can create a loop in which a narrow set of items crowds out exploration. Too much curation becomes manipulation. Too little becomes chaos.

Both systems can be excellent, but excellence means something different in each case. A library is not trying to predict your next click. A recommender is not trying to preserve the freedom of an idea to evolve on its own. When we mistake one for the other, we get brittle systems: platforms that treat culture like inventory, or open communities that mistake openness for usefulness.

The genius of Jefferson’s light metaphor is that it captures a property many digital systems still fail to internalize: some value compounds through exposure. The more minds that encounter an idea, the more pathways it can find into the world. That does not mean all sharing is good. It means the right question is not simply whether to share, but what kind of value should be shared, under what conditions, and with what guardrails.

A Useful Framework: The Four Fates of Value

To think clearly about these tradeoffs, it helps to classify value into four fates. This framework can be applied to products, companies, media, and intellectual property.

1. Extractive value

This is value that is consumed by a system without returning much to the source. Clickbait thrives here. It captures attention, but often at the expense of trust, clarity, or long-term loyalty. A recommendation engine can optimize for this short-term extraction if it only measures immediate engagement.

2. Retentive value

This is value that keeps users or contributors inside a system. It can be useful, but it can also trap. A recommendation loop that endlessly serves similar content creates retention, yet may not create understanding, growth, or delight. Many platforms confuse retentive value with meaningful value.

3. Multiplicative value

This is value that increases when shared. Open-source software is the clearest example, but the pattern appears in education, writing, design systems, and scientific research. A good API, a clear framework, or a public playbook can become more useful as more people apply it.

4. Compositional value

This is value that becomes stronger when paired with other systems. An idea may not just spread, it may combine. A recommendation model may be mediocre on its own, but when paired with transparent editorial standards and user controls, it becomes more trustworthy. A public idea may be ordinary alone, but in the hands of a community it becomes infrastructure.

The highest-leverage systems do not choose one fate of value forever. They route different kinds of value into the right fate at the right time.

This is where the synthesis becomes practical. A recommendation engine should not try to make everything multiplicative. It cannot. Not every piece of content deserves viral exposure. But it can be designed to promote multiplicative value where appropriate, by surfacing ideas, tools, and creators that gain strength through wider use. Likewise, open IP should not pretend curation is unnecessary. Once ideas are shared, someone still has to help people find the right ones, in the right order, for the right task.

In other words, sharing and recommending are complements, not opposites. Open systems generate raw material. Recommenders help that material become usable. Recommenders create pathways through abundance. Openness creates abundance worth navigating.

Why Giving Away the Right Thing Is Not Charity, It Is Strategy

There is a common misconception that giving away intellectual property is a sacrifice made for moral reasons alone. Sometimes it is moral. Often it is also strategic, but not in the cynical sense. It is strategic because some forms of value do not survive confinement.

If you have ever watched a talented teacher explain a hard concept, you know this instinctively. The teacher does not lose the idea by making it clear. The idea becomes stronger because it is now accessible to more people. Students then generate questions, examples, and applications the teacher did not anticipate. Knowledge improves by being used.

This is why many open systems grow faster than closed ones. They lower the cost of participation. They invite others to contribute improvements. They turn the world into a co-developer. The original creator is not diminished, because the asset is not a finite object sitting in a vault. It is a pattern of coordination.

But there is a subtle catch: openness only works when the shared object is truly reusable. A vague slogan is not enough. A public artifact must be legible, modular, and adaptable. Otherwise it is not multiplication, it is just diffusion. Diffusion spreads material. Multiplication spreads capability.

That distinction is crucial for modern product builders. Many teams release content and call it openness, but they have not created a commons. They have created a brochure. The difference is whether other people can do something with what you gave them.

The same logic applies to recommendation systems. A good recommender is not just a machine for extracting clicks. It is a coordination tool that helps users turn abundance into action. If it can be tuned only for clicks, it becomes a noisy marketplace of impulses. If it can be tuned for learning, discovery, trust, and long-term satisfaction, it becomes a genuine interface between scarcity and possibility.

The Better Question for Builders and Leaders

Once you see the tension between curation and multiplication, the standard questions change.

Instead of asking, “How do we maximize engagement?” ask, “What kind of engagement leads to durable value?”

Instead of asking, “Should we open source this or keep it proprietary?” ask, “Does this asset become more powerful when others can build on it?”

Instead of asking, “What is the best recommendation?” ask, “Best for whom, and over what time horizon?”

These questions sound simple, but they force a more mature design philosophy. They push you away from the fantasy of universal optimization and toward the reality of stakeholder tradeoffs. Every system has people who gain and people who lose from a given choice. A responsible builder names those tradeoffs instead of pretending they do not exist.

This is especially important in an era when algorithms increasingly mediate culture, commerce, and attention. We have spent years arguing about whether recommendation systems are good or bad, but that framing is too blunt. A better framing is whether a recommendation system expands a person’s real options or narrows them. Does it create agency, or merely predict inertia? Does it help users discover value, or does it imprison them inside the profile the system has built of them?

Similarly, open systems should not be romanticized. Giving away ideas is not automatically noble. If the underlying asset is weak, openness simply accelerates irrelevance. The point is not to give everything away. The point is to understand which parts of your work are non-rival, composable, and potentially multiplicative. Those are the parts that can turn sharing into strength rather than leakage.

A company, then, should think like a gardener, not a hoarder. Some seeds are best kept in the silo. Others need wind.

Key Takeaways

Not all value should be optimized the same way. Attention is scarce, so it requires curation. Ideas can multiply, so they benefit from sharing and reuse.
The real design tension is curation versus multiplication. Recommendation systems manage what people see now. Open knowledge systems increase what people can build later. Great systems know when to do each.
Ask who benefits, and over what time horizon. A system can be good for short-term engagement and bad for long-term trust. It can be bad for immediate control and excellent for long-term innovation.
Open only what is reusable. Sharing works best when the artifact is legible, modular, and useful to others. Diffusion is not the same as multiplication.
Design for compounding, not just consumption. The most valuable systems turn use into more use, learning into more learning, and sharing into more capability.

Conclusion: The Future Belongs to Systems That Know What to Copy and What to Curate

The old instinct was to guard value and distribute access sparingly. The newer instinct is to open everything and let the crowd sort it out. Both instincts are incomplete. The real challenge is more subtle: decide which things should be curated through scarcity and which things should be multiplied through generosity.

That distinction may turn out to be one of the defining design questions of the digital age. Recommendation engines sit at the gate of attention. Open ideas sit at the gate of progress. If we can learn to treat them as different species of value, not competing ideologies, we can build systems that are both more humane and more powerful.

The deepest lesson is not that sharing is always good or that algorithms are always suspect. It is that value has different physics depending on what it is made of. Some things get weaker when copied. Some things get stronger. Wisdom begins when we stop assuming all value behaves the same, and start building systems that respect the difference.