The Knowledge Nobody Teaches
How AI Is Breaking Incentive Structures in Education
The Insight That Changed Everything
Some insights don’t feel like “new knowledge.” They feel like the world rearranging itself. Suddenly things make sense that seemed chaotic before—not because they changed, but because you finally understood what drives them.
For me, one such moment came recently with an insight about stock markets: prices aren’t primarily driven by conviction, but by compulsion. Not because someone believes something is bad, but because someone *must* act—mandates force sales, volatility demands reduction, rebalancing creates pressure. These aren’t opinions. They’re mechanical processes.
This insight was valuable. But even more valuable was the meta-question that followed:
Why did nobody ever explain this to me?
The Incentive Problem of Knowledge Transmission
The answer is uncomfortable: Not because this knowledge is secret, but because nobody benefits from spreading it.
Most financial knowledge that circulates publicly shares one characteristic—it increases activity. It motivates action, generates opinions, creates urgency, produces narratives. Understanding market flows does the opposite. It says: Wait. Do nothing. Endure. Accept that you’ll be right for a while and still lose.
That’s hard to sell. Hard to package. And economically unattractive.
But this phenomenon isn’t limited to financial markets. It’s a fundamental problem of knowledge transmission.
The Three Categories of Untaught Knowledge
1. Tacit Knowledge—The Implicit Dimension
Tacit knowledge encompasses skills, ideas, and experiences that people possess but that aren’t codified and can’t necessarily be easily expressed.[^1] Knowledge management literature frequently reports that a very large proportion (often cited as 85-90%) of an organization’s knowledge resides in employees’ heads and has never been documented though this estimate is more of a frequently repeated rule of thumb than a precise measurement.[^2]
This knowledge is experience-based, rooted in practice, and difficult to articulate.[^1] It’s deeply personal, context-specific, and extremely difficult for competitors to imitate—which is exactly why it potentially offers the greatest competitive advantage.[^3]
Former HP CEO Lew Platt is frequently quoted as saying: “If HP knew what HP knows, we would be three times more profitable”, a quote often used in knowledge management literature to illustrate the challenge of organizational learning, though a primary source is difficult to verify.[^4]
The problem: Effective transfer of tacit knowledge typically requires extensive personal contact, regular interaction, and trust.[^1] These resources are limited—and expensive.
2. Hidden Curriculum—The Unwritten Rules
The term “Hidden Curriculum” was coined in 1968 by Philip W. Jackson in his book “Life in Classrooms.”[^5] He argued that education must be understood as a socialization process, with a mass of unspoken academic and social norms.
The hidden curriculum includes usable strategies for success, but the way formal curricula are structured leaves no room for such life lessons.[^6] Research shows that first-generation students in particular struggle to understand faculty’s implicit expectations they differ significantly from traditional students in their interpretation of deadlines, work standards, and academic norms.[^6]
3. Structurally Neglected Knowledge
Even when evidence exists, it’s often not taught. Example: Research on reading comprehension shows that background knowledge and topic-specific vocabulary are more crucial for understanding text than general strategies.[^7] Yet reading instruction often focuses heavily on decontextualized “skills”, a phenomenon documented in educational research but long not widely discussed.
The delay between research evidence and educational practice is a well-documented phenomenon, though the specific mechanisms are complex.
Why This Knowledge Isn’t Taught
The reasons are systematic:
Economic incentives are missing
Knowledge that preaches patience can’t be sold
Knowledge that reduces activity generates no commissions
Knowledge that brings slow rewards is hard to monetize
Structural barriers
Tacit knowledge is difficult to codify and document
Transfer requires time and personal contact
Organizations often don’t know what they know
Institutional inertia
Curricula change slowly
Evidence is ignored when inconvenient
The system rewards measurable outputs, not deep understanding
The AI Revolution: Access Without Incentives
And this is where artificial intelligence enters the picture.
AI promises to democratize access to knowledge but not just to information that’s already widely distributed. It potentially opens access to knowledge that *nobody has an incentive to teach*.
How AI Could Break the Incentive Structure
1. Reduced Dependence on Financial Incentives
The PERLA Framework (Personalization and Learning Analytics) describes how AI systems can personalize learning experiences.[^8] Large language models make it possible to connect people with knowledge in a conversational style, independent of traditional gatekeepers.
While traditional education programs are costly and limited in scalability, digital platforms can deliver education at scale at low cost though evidence on long-term effectiveness without accompaniment remains limited.
2. New Forms of Knowledge Codification
Modern AI systems show impressive capabilities in tasks traditionally considered “tacit.” To what extent this can be understood as genuine “codification of tacit knowledge” in Polanyi’s sense remains an open conceptual question. But these systems are beginning to make accessible knowledge that previously could only be transferred through personal interaction.
3. Scalability with Reduced Barriers
No-code tools and low-barrier AI interfaces undeniably lower technical hurdles. AI capabilities are becoming increasingly accessible, with both costs and technical barriers falling.
The Limits: What AI (Still) Can’t Do
Of course, AI is no panacea. The quality of serendipitous learning in human conversations, receiving knowledge you hadn’t intended to acquire is a frequently reported experience. Systematic comparisons with AI interactions are still pending.
The risk of spreading misinformation through broader access to information and generation tools is real and intensively discussed in research on misinformation and digital media. People without relevant expertise increasingly question expert statements, a phenomenon facilitated by technology access.
And: Digital tools increase engagement but rarely lead to lasting changes without appropriate follow-up.
What Could Fundamentally Change
Despite these limitations, AI has the potential to change something fundamental: It could increasingly decouple knowledge transmission from economic incentives.
This could mean:
Tacit knowledge becomes more accessible without someone having to invest years of their life in personal transfer
Hidden curriculum becomes more visible without needing access to the right networks
Uncomfortable knowledge becomes more available even when it generates no activity or brings slow rewards
The Meta-Lesson: About Markets and Learning
Perhaps the most important insight from all this is meta-epistemic:
The most valuable knowledge is often not the most complex, but that which nobody has an incentive to teach.
This insight applies not just to financial markets, but to almost every domain of knowledge:
In medicine: Prevention brings less money than treatment
In psychology: Long-term behavioral change is harder to sell than quick fixes
In education: Deep understanding is harder to measure than standardized tests
In career development: Patience and strategic waiting are harder to monetize than networking events
AI doesn’t completely break this structure. But for the first time, it creates systematic access to knowledge that lies outside traditional incentive structures.
A New Coordinate System
Since understanding these connections, I ask myself a new question with every knowledge gap:
Not: “Is this knowledge available?”
But: “Who would have an incentive to teach me this?”
And if the answer is “nobody,” I know: Here lies potentially valuable knowledge.
Because the paradox of learning in the modern world is: The most important lessons are often not those proclaimed most loudly. They’re the ones that stay quiet, not because they’re secret, but because nobody can build a business model around them.
AI gives us, for the first time, the ability to systematically search for this silent knowledge.
---
Sources
[^1]: Polanyi, M. (1966). *The Tacit Dimension*. University of Chicago Press. Polanyi coined the term “tacit knowledge” with his famous statement “we can know more than we can tell” and argued that tacit knowledge is acquired through personal experience and can only be transferred through practice.
[^2]: Smith, E.A. (2001). “The role of tacit and explicit knowledge in the workplace.” *Journal of Knowledge Management*, 5(4), 311-321. [DOI: 10.1108/13673270110411733](https://doi.org/10.1108/13673270110411733); The oft-cited “85-90% estimate” appears in various knowledge management studies as a frequently reported rule of thumb, though the exact empirical basis for these figures is disputed.
[^3]: McAdam, R., Mason, B., & McCrory, J. (2007). “Exploring the dichotomies within the tacit knowledge literature: towards a process of tacit knowing in organizations.” *Management Learning*, 38(2), 147-166. [DOI: 10.1177/1350507607075774](https://doi.org/10.1177/1350507607075774)
[^4]: Lew Platt (former CEO Hewlett-Packard). This quote is frequently cited in knowledge management literature (see Davenport & Prusak, 1998); an exact primary source is difficult to verify.
[^5]: Jackson, P.W. (1968). *Life in Classrooms*. New York: Holt, Rinehart & Winston. Jackson coined the term “Hidden Curriculum” for the implicit norms and expectations in educational institutions.
[^6]: Collier, P.J., & Morgan, D.L. (2008). “’Is that paper really due today?’: differences in first-generation and traditional college students’ understandings of faculty expectations.” *Higher Education*, 55, 425-446. [DOI: 10.1007/s10734-007-9065-5](https://doi.org/10.1007/s10734-007-9065-5)
[^7]: Recht, D.R., & Leslie, L. (1988). “Effect of prior knowledge on good and poor readers’ memory of text.” *Journal of Educational Psychology*, 80(1), 16-20. This classic study showed that background knowledge better explained the difference between good and poor readers than general reading ability.
[^8]: Chatti, M.A., & Muslim, A. (2019). “The PERLA Framework: Blending Personalization and Learning Analytics.” *International Review of Research in Open and Distributed Learning*, 20(1), 243-261. [DOI: 10.19173/irrodl.v20i1.3936](https://doi.org/10.19173/irrodl.v20i1.3936)