1. Two Hypotheses In the contemporary discussion about artificial intelligence, two distinct hypotheses intersect and are often conflated.

The first hypothesis describes AI as a thin client between intention and result. Historically, a chain of translators existed between a concept and an artifact. A person formulated a task for a programmer, the programmer wrote code, the code became a program. A screenwriter passed an idea to a studio, the studio hired a VFX team, the team produced a film. A composer worked with musicians and a studio to record a track. AI shortens this chain, allowing a result to be obtained directly from a natural language prompt.

The second hypothesis is more radical. It asserts that AI washes out not only performers but also apprentices. The main function of many professions was not the production of the current result, but the reproduction of knowledge. A junior was needed not because he is useful today, but because in five years he will become a senior. A student was needed not to create value now, but to become an engineer. A doctoral candidate was needed not for brilliant papers, but to undergo the school of scientific thinking.

1. The Destruction of the Apprenticeship Mechanism The classical model of competence growth was built on review. A junior wrote code, a senior dissected it, extracted the substrate of experience, and transmitted professional intuition. Each review was an act of knowledge transfer.

The new model looks different. A person formulates a prompt, AI generates the result. If code of acceptable quality appears immediately, the economic need for a junior declines. Along with it, the mechanism through which knowledge was transmitted disappears.

A structural question arises that goes beyond the labor market. Where will the next seniors come from if the intermediate link does not undergo the path of learning through mistakes and reviews. This is a problem of competence reproduction, not simply automation.

1. The Transformation of Education Historically, the university and school performed the function of an institution of verification. The teacher took lived experience, analyzed it, and taught how to distinguish working knowledge from noise.

Under conditions of mass AI adoption, this function shifts. Teaching increasingly concentrates not on the subject, but on the ability to work with the model: formulating queries, checking answers, assembling agent chains. Knowledge of the subject is assumed to be available on demand, therefore teaching knowledge as such recedes to the background.

Education is turning from an institution of succession into a course on interacting with a thin client.

1. Where the Teaching of Knowledge Goes The teaching of knowledge does not disappear completely, but is pushed to the periphery and distributed across three directions.

First direction: inside models. Knowledge is preserved in the form of statistical weights, without an author, without context, and without a witness who could explain why a solution works.

Second direction: into narrow craft communities. Small laboratories, open-source groups, workshops where the practice of personal analysis and transmission of experience is preserved.

Third direction: into nowhere. A large part of intermediate knowledge simply ceases to be reproduced because the economic incentive to transmit it disappears. There is no systemic reason to teach rotoscoping, syntax, or mixing if these operations are performed by a model.

The paradoxical effect is that access to knowledge has become instantaneous, while learning knowledge has become a luxury. An indirect indicator of this shift is the growth in requests to encyclopedic resources. It is not the number of people who learn that is increasing, but the number of agents that index.

1. Can AI Become a Mentor The key assumption of the second hypothesis is that AI is fundamentally incapable of performing the function of a mentor. Today this assumption has grounding. Models provide answers well, but they form professional intuition poorly. A master usually says: this solution works, but in two years the system will collapse at this point. Such knowledge is based on lived experience of consequences, not on text patterns.

Current models work with corpora, not with experience of operating solutions. This limitation is not proof of a fundamental impossibility of AI mentorship, but it records the current state of the technology.

1. The Problem of Selection, Not Origin The most contentious claim is that new knowledge bases will be filled with statistical noise without verification. Historically, knowledge has never undergone ideal filtration. Universities, scientific schools, and corporations also produced a significant amount of noise.

The problem therefore lies not in who generates the content, a human or a model, but in the presence of a selection mechanism. If high-quality review, testing, replication of experiments, and audit exist, knowledge can be reproduced regardless of the origin of the text.

AI accelerates the production of information faster than society creates new institutions for its verification. It is precisely this gap between the speed of generation and the speed of verification that creates the risk of accumulating unreflective content in knowledge bases.

Conclusion
The analysis shows that the real subject of the discussion is shifting. The first part of the discussion describes AI as a tool for shortening production chains. The second part points to a more fundamental process.

AI removes intermediaries between intention and result, and together with the intermediaries, the institutions through which society reproduced bearers of knowledge disappear. The issue is not so much the automation of the labor of programmers, musicians, or artists, as the possible crisis of knowledge succession.

The key question of the next decade is not whether a model can write code or generate a film, but whether the social mechanism for the emergence of the next generation of specialists capable of understanding why this code and this film work will be preserved.