Author: Luiza Jarovsky Source: https://www.luizasnewsletter.com/p/ais-acceleration-paradox Published: March 6, 2026 (edition #278)
"AI's Acceleration Paradox" is a March 6, 2026 newsletter essay (edition #278) by Luiza Jarovsky arguing that the AI industry's acceleration narrative — framed around productivity, abundance, and curing all diseases — rests on what she calls false premises about humans as living, finite, embodied entities with hard biological and psychological boundaries. Jarovsky contends that sustained AI-powered acceleration is dragging society toward a dystopia in which people constantly minimize biological boundaries, ignore psychological needs, and devalue human expression in order to thrive in a machine-prioritizing world. The full framework is covered at AI Acceleration Paradox.
Summary of argument
Jarovsky's central claim is that acceleration cannot deliver on its own promises because human beings are bounded. She frames the limits in two registers, biological and psychological.
The biological boundaries she cites are an average human lifespan of 73 years, roughly 7 hours of sleep per night, roughly 3 liters of water per day, and roughly 2,000 calories per day. These, she argues, cannot be increased with more compute, larger datasets, new training methods, or more efficient chips. The psychological and emotional needs she enumerates — safety, love, belongingness, esteem, and self-fulfillment — are presented as equally hard constraints, in that failing to meet any of them can, in her account, lead to illness or death.
To anchor the argument in finitude, Jarovsky invokes the regrets of the dying. She writes: "I am sorry to break the news to the AI industry, but it has nothing to do with acceleration or productivity. I have never heard of anyone on their deathbed lamenting not being more productive or accelerated." She cites Bronnie Ware's Regrets of the Dying, whose top five regrets concern courage, working less, expressing feelings, friendship, and happiness — observing that people regret not having worked less, not more. She presents the deathbed-regret material as a structural argument against the framing of productivity as an end in itself, rather than as an abstract appeal.
Effects on the individual
Jarovsky describes what AI acceleration means in practice for an individual worker. Delegating tasks leads to losing control of work processes and outputs. Using AI for creative and intellectual tasks can prevent the development of essential skills, and constant AI automation can produce deskilling. Workers may end up managing many AI tasks without understanding the underlying processes, and may experience greater isolation and disconnection from human teams.
She supports these points with two empirical references: a February 2026 Harvard Business Review item arguing that AI does not reduce work but intensifies it, and an arXiv paper on AI's negative impact on skill formation, especially among junior employees.
Two possible futures
Jarovsky frames the choice ahead as a binary between AI as an end and AI as a tool.
Under "AI as End" (Option A), AI fully controls decisions in the name of "progress and innovation," humans are treated as limited entities that must delegate, and AI is regarded as a "special entity" or "alien" that must be respected. Jarovsky argues that the AI industry is currently fostering this option. She names specific instances she reads as part of this trajectory: the "Claude Constitution," Yuval Harari's "alien" framing, and Sam Altman's "abundant intelligence." She forecasts uncontrolled chaos as the failure mode if acceleration continues without meaningful human control, oversight, and scrutiny.
Under "AI as Tool" (Option B), AI requires strict human oversight and control and serves as a tool to foster human flourishing, with rules focused on human-led innovation, well-being, and fundamental rights, and limits on certain use cases to prevent harm. Jarovsky places herself unambiguously in this position.
Relationships
- authored-by: Luiza Jarovsky
- introduces: AI Acceleration Paradox
- related: Cognitive Friction (prescriptive response), AI Divides (Literacy / Occupational / Ethico-Philosophical) (social structure), Technical AI Policies (faster-than-law governance), America's AI Action Plan (oppositional framing)