Two Ways to Use AI: Homo Economicus vs. Homo Ludens

The academy has a problem, and it's been getting worse for over a century.[1]

We organize knowledge into disciplines—history, psychology, neuroscience, linguistics, economics—each with its own journals, conferences, and vocabulary. This structure, inherited from 19th-century German universities, serves one purpose brilliantly: it lets specialists gather details efficiently within well-defined boundaries.

But knowledge doesn't respect boundaries. The most important questions—How does the mind work? What makes us creative? Why do societies change?—require insights from multiple disciplines. The pattern you need to see often spans several "bins" of specialized knowledge.

Here's the paradox: we've been talking about interdisciplinary work for decades. Universities have interdisciplinary centers everywhere. Yet the actual structure of academic life—hiring, promotion, publication, funding—still runs on disciplinary rails laid down 150 years ago.

Now we have large language models. And we face a choice about how to use them.

The Economicus Approach

One path is to use LLMs to amplify and accelerate current arrangements. Let's call this the Homo economicus approach—the economic human, focused on optimizing production.

In this mode, LLMs become tools for:

• Writing literature reviews faster
• Reviewing papers for journals more efficiently
• Generating incremental research at scale
• Producing more publications per year
• Staying safely within disciplinary boundaries

This sounds productive. More papers, faster reviews, greater output. But it doubles down on exactly what's broken. We already produce too many narrow specialist papers that too few people read. Using AI to produce more of them faster just amplifies the dysfunction.

The economicus approach treats knowledge production like manufacturing: maximize output, minimize cost, optimize existing processes. Stay in your lane. Don't take risks. Generate the next incremental advance.

The Ludens Alternative

There's another path. Call it Homo ludens—the playing human, focused on exploration and discovery.

In this mode, LLMs become tools for genuine cross-disciplinary integration. Not producing papers, but discovering connections. Not automating existing processes, but enabling new formations.

Here's what this looks like in practice:

Strategic Search Across Disciplines

Say you're investigating how language develops in children. Traditional approach: read the developmental psychology literature, maybe venture into linguistics if you're bold.

Ludens approach with LLMs: "Find work from any field that addresses the relationship between motor development and symbolic capacity."

The LLM doesn't care about departmental boundaries. It surfaces relevant work from neuroscience, evolutionary biology, comparative psychology, and anthropology—connections that specialists, confined to their silos, would miss.

Constraint Satisfaction Across Domains

Rigorous integration requires checking whether your ideas satisfy constraints from multiple fields simultaneously. Is your model of language acquisition consistent with what we know about brain development? Does it align with evolutionary timescales? Does it match observed behavior?

An LLM can rapidly check these cross-domain constraints: "Does this cognitive science claim contradict findings in neurobiology? What about developmental timelines?" It doesn't replace judgment, but it surfaces contradictions and connections that would take months of reading to discover.

Pattern Discovery in Unexpected Places

The most valuable insights often come from recognizing that two fields are studying the same phenomenon with different vocabularies. LLMs excel at this kind of pattern matching across terminological boundaries.

"What work in any discipline addresses hierarchical control systems switching between modes?" The answer might come from neuroscience (neural modulation), robotics (control architectures), or organizational psychology (decision-making frameworks). These aren't citations to pad your bibliography—they're genuinely different perspectives on the same deep problem.

The Center-Out Method

Start with a specific case—a text, an event, a phenomenon—and radiate outward to topics it touches. An LLM can help map these connections systematically: given this particular case study, what frameworks from different disciplines illuminate different aspects of it? [2]

This mirrors how actual insight works: you're wrestling with something specific, and you need whatever intellectual tools help, regardless of which department developed them.

Why This Matters

The difference isn't just practical—it's philosophical.

Economicus treats LLMs as labor-saving devices. Do what we already do, but faster and cheaper. This keeps us trapped in the existing system, just at higher speed.

Ludens treats LLMs as exploration tools. Find patterns we couldn't see before. Make connections that disciplinary blinders obscured. Enable the integrative work that institutions make nearly impossible.

The economicus approach optimizes local maxima—you get better and better at what you're already doing. The ludens approach helps you find new maxima you didn't know existed.

The Play Element

There's a deeper reason the ludens approach matters: genuine discovery requires play.

Not play as opposed to serious work, but play in the sense of free exploration before commitment. Trying unusual combinations. Following tangential connections. Seeing what emerges without knowing in advance what you're looking for.

This is how children learn, how scientists make breakthroughs, how jazz musicians create. You need freedom to explore widely before you settle on what's worth pursuing seriously.

The economicus approach eliminates this exploratory freedom in the name of efficiency. It optimizes production, but production of what? More of what we already have.

The ludens approach embraces exploration. You're not trying to write the next incremental paper. You're trying to discover what you don't yet know you're looking for.

The Current Moment

Right now, institutions are moving toward the economicus approach. Using LLMs to review more papers, generate more text, process more grant applications. It's understandable—they're under pressure to handle increasing volume.

But this is a catastrophic missed opportunity.

LLMs are genuinely good at working across disciplinary boundaries. They don't have careers to protect or departments to represent. They can pattern-match across the entire literature without caring which journal it appeared in. They're natural tools for the kind of integrative work that the current system makes nearly impossible.

Using them instead to accelerate existing processes is like using the internet purely to send faxes faster.

What Individuals Can Do

You don't need institutional permission to use LLMs in ludens mode. Anyone doing serious intellectual work can:

• Use them to search for cross-domain connections their specialty would miss
• Test ideas against multiple fields simultaneously
• Find relevant work hiding behind different terminology
• Map unexpected patterns across disciplinary boundaries
• Build integrative frameworks that specialists wouldn't attempt

The work still requires human judgment. LLMs don't understand what matters or evaluate theoretical coherence. But they can surface connections, check consistency across domains, and search spaces too large for individual humans to cover.

The integration happens in your mind. The LLM just makes it possible to range more widely, check more thoroughly, and discover more connections than you could alone.

The Choice

We're at a fork. One path uses AI to amplify the existing system—more papers, faster production, same disciplinary silos. The other uses AI to enable what that system prevents—genuine integration across boundaries.

Homo economicus wants to optimize. Homo ludens wants to discover.

The economicus path is safer, more measurable, easier to justify in budget meetings. The ludens path is riskier, harder to quantify, requires defending exploration against demands for immediate productivity.

But only the ludens path actually advances understanding. Only exploration discovers genuinely new patterns. Only play enables the modal flexibility needed for creative work.

The tragedy would be to have tools that could finally make cross-disciplinary integration tractable, and use them instead to produce more siloed specialist papers at industrial scale.

The opportunity is to use these tools the way evolution uses play—as a way to explore widely, discover unexpected possibilities, and find new formations that optimization alone would never reach.

The question isn't whether AI will change how we do intellectual work. It's whether we'll use it to amplify what's broken or enable what's been nearly impossible. The choice is ours.

Notes

[1] Claude has written this essay at my request. We had been discussing interdisciplinary work, the fact that, in my experience, the academy has been calling for it for decades, and has established interdisciplinary centers all over the place, but still, the traditional 19th century disciplinary arrangements dominate. The contrast between Homo Economicus and Homo Ludens is a central conceptual motif I am developing in my book, Play: How to Stay Human in the AI Revolution.

During this conversation I told Claude how I'd defined cognitive science in my 1978 dissertation, "Cognitive Science and Literary Theory." I said:

The basic problem of cognitive science is establishing a five way correspondence between the following:

1. Brain Geometry: Neuroanatomy is the study of the geometry of the brain. Comparative neuroanatomy is the study of the correspondence between brains of different species.
2. Computation: Different types of computers can perform different classes of computations and the nature of the computations depends on the geometry of the computer.
3. Behavior: Much of psychology is the study of the behavior of organisms. The behavior an organism exhibits is determined by the class of computations which its brain can perform which in turn depends on the geometry of the brain.
4. Phylogeny: Animals at different evolutionary grades have different brain geometries. The brain geometry must be capable of performing the class of computations necessary for survival in the animal’s particular ecological niche. But what is the relationship between moving to a new niche and the emergence of a new brain geometry?
5. Ontogeny: As a child matures different brain structures develop and permit new classes of computation sustaining new types of behavior. But how does phylogeny exploit differential maturation rates to create a new class of computer?

A decade later David Hays and I used that framework in an article we published, "Principles and Development of Natural Intelligence," Journal of Social and Biological Structures, Vol. 11, No. 8, July 1988, 293-322. https://www.academia.edu/235116/Principles_and_Development_of_Natural_Intelligence 

[2] This conception is from a report I prepared for the School of Humanities and Social Sciences during my final year at The Rensselaer Polytechnic Institute, Policy, Strategy, Tactics: Intellectual Integration in the Human Sciences, an Approach for a New Era, Report to the Faculty, School of Humanities and Social Sciences, Rensselaer Polytechnic Institute, 1985. https://www.academia.edu/8722681/Policy_Strategy_Tactics_Intellectual_Integration_in_the_Human_Sciences_an_Approach_for_a_New_Era