Hundreds of Billions Are Going Into AI While People Rethink the Meaning of Education and Work
Why does spending hundreds of billions on AI force people to reconsider the value of education, professions, and their own time?
Understand which parts of education and work become cheaper with AI, and which skills—goal-setting, review, and accountability—become more valuable.
What to watch for
Key takeaways
The decision in “Why more young people are rejecting higher education: Soviet and American views of education” depends on one criterion: Gen Z in the U.S. treats becoming an electrician or a plumber not as a bad fallback but as a faster path to income without years of debt; this is not a rejection of knowledge — it is a refusal to keep paying for a system whose value is no longer obvious.
The decision in “Where we're going” depends on one criterion: the old link between a degree and guaranteed income has broken, and online education created an enormous range of very uneven quality; the hosts see the shift toward skilled trades as a wave lasting five to fifteen years, overlapping with the changes in the technology market.
The boundary of the “The $100 Billion OpenAI–Microsoft Supercomputer” case is defined by this point: the hundred-billion-dollar project is not meant to power one more convenient chat: it should provide the computing base for next-generation models and agents and strengthen Microsoft's cloud position against Amazon, Google, and Oracle — data-transfer speed, chip access, and proprietary data centers matter no less than the model itself.
The practical meaning of “New operating systems?” is that the hosts tie new platforms to agents — systems that autonomously handle knowledge-worker tasks that take a person hours; the technology is not there yet, though Sam Altman promises such capabilities as early as GPT-5.
The discussion of “Democratizing AI development: when anyone can program” yields a practical test: if local AI lets far more people program, the barrier to development falls, but understanding the task, checking the result, and taking responsibility become more valuable — a model can assemble a draft or an interface, but it cannot say why it is being built and what counts as a good result.
In the context of “Working with AI and with people,” this criterion applies: when a person works together with a model, it remains open who owns the result and how income should be shared: not everything a model produces can be treated as someone else's, but the source data, style, and contribution of authors cannot be dismissed either.
The boundary of the “Massive investment in AI and declining trust in traditional education point to the same conclusion” case is defined by this point: a degree, a profession, a software product, and a large model now have to prove their value on a real task, and the faster execution gets cheaper, the more valuable a person's ability to choose the task, understand the consequences, and answer for the decision.
What this episode is about
Microsoft and OpenAI are prepared to spend enormous sums on supercomputers, while young Americans are increasingly choosing short, practical qualifications over years in college. The two stories are connected: AI is changing not only technology, but also the value of skills, time, and human labor.
When college means years of life and a large debt, while the labor market offers graduates no immediate advantage, young people start doing the math differently. In the United States, more members of Gen Z are treating work as an electrician, plumber, or another skilled trade not as a bad fallback, but as a faster path to income without years of financial burden. This is not a rejection of knowledge.
It is a refusal to keep paying for a system whose value has become less obvious.
At the other end of the market, Microsoft and OpenAI are discussing infrastructure that could cost around one hundred billion dollars. Projects like this are not meant to power one more convenient chat interface. They are supposed to provide the computing base for the next generation of models and agents, while strengthening Microsoft's cloud position against Amazon, Google, and Oracle.
Data-transfer speed, access to chips, and proprietary data centers are becoming competitive advantages on the same level as the model itself.
But the more powerful the system becomes, the sharper the question is: what remains for the person? If local AI makes programming accessible to many more people, the barrier to software development will indeed fall. At the same time, understanding the task, checking the result, and accepting responsibility will become more valuable. AI can produce a draft, assemble an interface, or process data, but it does not remove the need to know why something is being built and what a good result should look like.
The same problem appears in creative work and copyright. When a person works together with a model, who owns the result, and how should the income be divided? There is no simple answer yet. We cannot automatically treat everything produced by a model as someone else's property, but we also cannot pretend that the source material, style, and contribution of creators mean nothing.
Massive investment in AI and declining trust in traditional education point to the same conclusion: old guarantees no longer work on their own. A degree, a profession, a software product, or a large model must prove its value on a real task.
And the faster technology reduces the cost of execution, the more important a person's ability to choose the task, understand the consequences, and take responsibility for the decision becomes.
As execution becomes cheaper, human judgment becomes more valuable: the goal, the consequences, and accountability cannot be handed to the model with the routine work.
Episode transcript
The episode is in Russian; below is an English reading guide to the transcript (the full EN transcript is a machine translation). Voice matching applied to 36 segments: 29 identified, 1 mixed, 4 probable, and 2 unresolved.
Read transcript on a separate page
Loading…