AI, Costs, and the Myth of Inevitable Human Obsolescence

 

 

Why the disruption narrative is more complicated,  and more hopeful than it appears

 

For years, the AI narrative has been relentlessly linear and tinged with apocalypse: models get smarter, cheaper, and more capable — and humans get edged aside, role by role, sector by sector. Then came a headline (1) that disrupted the script.

“Microsoft is limiting internal use of expensive AI coding tools as enterprise AI costs surge.”

It sounded like a contradiction from the company that bet its future on AI, poured $80 billion into data centres, and plastered Copilot across every product it makes. But it was not a contradiction. It was a revelation.

The Economics Behind the Curtain

Inside Microsoft's engineering divisions, Claude Code the AI coding assistant from Anthropic was not cancelled because it failed. It was cancelled because it succeeded too well. Engineers used it relentlessly, and token-based billing, where every prompt, every agentic loop, every code-generation cycle cost real money, ballooned into millions of dollars for a single team. The cancellation deadline is June 30, 2026.

The pattern is wider than Microsoft:

•         Uber burned through its entire annual AI coding budget in four months, after actively incentivising engineers to maximise usage through internal leaderboards.

•         Meta built an internal dashboard called Claudeonomics to track which employees were consuming the most AI tokens.

•         Amazon encouraged "tokenmaxxing" — gamifying maximum AI consumption as a proxy for productivity.

 

The collective result: for many enterprise tasks, AI is now more expensive than humans. The promise was frictionless efficiency. The reality is that when thousands of employees use frontier models without discipline, the economics invert.

Does This Slow the Replacement of Humans?

 

Only partially, and only temporarily. The cost ceiling buys time. It does not change direction.

 

The "AI will replace humans" narrative was always too blunt. AI is not a flat substitute for human labour. It has a cost curve. At low usage it is remarkable. At scale, without governance, it is ruinous. Companies will not replace human beings wholesale. They will replace them selectively, deploying AI where ROI is unambiguous, retaining humans where judgment, accountability, ambiguity, or trust cannot be priced away.

 

The most exposed roles are not those at the bottom of the skills ladder or the top. They are in the middle: paralegals, junior coders, financial analysts, content writers, radiology readers, customer service agents, roles that are routine, pattern-based, and high-volume. These are precisely the tasks where AI delivers the clearest per-unit economics.

 

The Real Disruptor: Robotics, Not Software

 

While enterprises wrestle with token bills, robotics companies are quietly solving a different equation.  and it has no analogue to the cost-per-prompt problem.

 

A humanoid robot is a capital expenditure. You buy it once. Its "salary" is electricity and maintenance. Tesla Optimus, Figure, and Boston Dynamics are all targeting price points that will undercut the minimum wage in developed economies within this decade. And they are beginning with exactly the jobs that employ hundreds of millions globally: fast food, warehouse picking, hotel housekeeping, retail stocking.

 

A burger-flipping robot does not need to be perfect. It only needs to be cheaper than a human over a five-year horizon. For low-wage work in high-cost economies, that crossover is arriving faster than most expect and unlike software AI costs, it will not be throttled by a token budget.

 

The China Factor: The Cost Floor May Collapse

 

DeepSeek's emergence earlier this year changed the global cost equation in ways that have not yet been fully absorbed. It demonstrated that frontier-level reasoning can be delivered at 20 to 50 times lower cost than comparable Western models. Chinese AI companies benefit from a structural advantage: state-subsidised compute and energy, lower infrastructure costs, and no pressure from shareholders to monetise quickly.

 

If DeepSeek-class models gain wide adoption in India, Southeast Asia, Latin America, and Africa, markets where price matters more than geopolitics, the cost barrier to replacing human workers with AI could collapse years ahead of current projections. The West will move more cautiously, constrained by data sovereignty and regulatory anxiety. The Global South may move faster, precisely because it cannot afford to be slow.

 

This creates a two-speed world of AI adoption,  and, by extension, a two-speed world of displacement. That asymmetry deserves far more attention than it currently receives in the global policy conversation.

 

A New Dimension: The Jobs That Don't Exist Yet

 

The debate about AI and employment is systematically incomplete without acknowledging a stubborn historical truth: every major technological shift destroys familiar work — and creates categories of work that were invisible until they became indispensable. The steam engine did not just displace handloom weavers. It created railway engineers, urban planners, and factory inspectors. The internet did not merely kill travel agencies. It created cloud architects, SEO strategists, and UX designers.

 

The honest challenge is that the new jobs are not legible until they exist. But looking at the edges of what is already emerging, by 2030 entirely new categories of work are plausible. and in some cases, already forming:

 

These are not science fiction. They are already emerging at the margins of the economy, waiting for the infrastructure to catch up.

 

AI as an Equaliser: The Welfare Dimension

 

There is an under-discussed possibility that deserves a place in this conversation. If AI dramatically reduces the cost of delivering essential services, the welfare gains could, under the right policy conditions, outweigh the disruption to employment.

 

•         In agriculture, AI-optimised irrigation, autonomous harvesting, and supply-chain prediction could reduce food waste and raise yields in the Global South.

•         In healthcare, AI triage, diagnostic support, and remote monitoring could bring high-quality care to rural and underserved populations who currently have none.

•         In education, adaptive AI tutors could personalise learning for hundreds of millions of children for whom quality schooling remains a geographic accident of birth.

 

If AI reduces the cost of food, health, and education by 50 to 80 percent, the question is no longer simply "who loses their job?" It becomes "what kind of society do we build with the surplus?" That is a question of political will, not technology.

 

Where Humans Still Win on ROI Until 2030

 

Despite the compression of timelines, there are domains where humans remain the higher-ROI choice, and will remain so through the end of this decade:

 

•         Trust-based, relationship-driven work, where clients pay a premium specifically for human accountability — someone they can hold responsible.

•         Novel problem-solving in genuinely ambiguous environments, where no training data yet exists for the situation at hand.

•         Skilled trades in unstructured physical settings, the plumber navigating an unfamiliar home, the electrician improvising under a deadline, where robotics cannot yet match dexterity and judgment at an economic price point.

•         Regulated roles requiring human sign-off, where law or professional standards mandate human accountability regardless of AI capability.

 

The pattern across these safe harbours is consistent. What makes humans irreplaceable is not intelligence alone. It is accountability, physical adaptability, and the fact that in some relationships, the human presence is the product, not merely the mechanism of its delivery.

 

The Question That Actually Matters

 

The debate has moved on from whether AI will displace human workers. That question is settled. The debate now is: which humans, doing which tasks, on what timeline, under which cost structures, and critically, will the new categories of work emerge quickly enough, and be accessible enough, to absorb those displaced?

 

The cost ceiling revealed by Microsoft, Uber, and others is real. It slows the slope. It forces selectivity. It creates space for societies to adapt rather than absorb a vertical shock. But it does not alter the destination. The direction of travel has not changed — only the gradient of the curve.

 

The most important investment any individual, institution, or government can make right now is not in AI itself. It is in the human capacity to navigate the transition: to identify which skills will compound in value, which roles are building toward the new categories, and which paths are quietly narrowing.

“AI will not erase human value. It will redraw the map of where that value lives,and whether we prepare to inhabit that new terrain is the defining human challenge of this decade.”

Note:(1)https://www.linkedin.com/posts/mario-mu%C3%B1oz-serrano_microsoft-ai-costs-share-7464125101605769216-yO2j/?utm_source=share&utm_medium=member_android&rcm=ACoAAAHWSYQBQXp5IHBAbmYtmy0QTDr2QMLjsi8