Always-On AI Is Driving Self-Exploitation

AI’s always-on tools are shifting ‘can’ into ‘must,’ fueling self-imposed overwork and spreading 996-like culture into the West. This internalized pressure aligns with Byung-Chul Han’s view of self-discipline and results in burnout that undermines creativity and innovation. The author argues for cultural change: set boundaries and use AI judiciously so it doesn’t consume our rest.

Key Points

AI’s always-on nature turns capability into obligation, creating a guilt loop where not using tools feels like falling behind.
996-style work culture is reportedly migrating to Silicon Valley, justified by a race-to-compete narrative in AI startups.
Historical pattern: technologies that extend capacity (like electric light) shift ‘can work’ into ‘should work,’ making luxuries into obligations.
Byung-Chul Han’s thesis: modern overwork is self-imposed; AI amplifies this ‘I can, therefore I must’ dynamic, driving self-exploitation.
Hyper-productivity is self-defeating: burnout reduces creativity and innovation; genuine creativity requires rest and reflection, making boundaries essential.

Sentiment

The community leans strongly toward agreement with the article's thesis. Most commenters validate from personal experience that automation leads to more work, not less, and that employers systematically capture productivity gains. The frustration is palpable, with many expressing cynicism about corporate incentives. The minority who disagree tend to be entrepreneurs or people reframing the problem as capitalism rather than technology. There is little enthusiasm for AI as a liberating force within traditional employment.

In Agreement

Automation historically increases output expectations and complexity rather than reducing working hours, with multiple commenters sharing personal stories of being punished for finishing work early
AI creates a unique validation burden: unlike traditional automation where quality is easy to check, LLM output is trivially generated but extremely hard to verify, making code review and mentoring more exhausting
The reward for automating your work is simply more work — employers raise the baseline and never share the gains, making the rational strategy to automate quietly and keep the savings to yourself
AI-generated code is degrading codebases with tech debt, filler comments, and poorly understood foundations, while PR authors increasingly cannot explain their own submissions
The always-on nature of AI tools creates psychological pressure: the ability to work transforms into an obligation to work, echoing Byung-Chul Han's concept of self-exploitation
This is Jevons Paradox applied to developer time — efficiency gains in producing code lead to more code being demanded, not less work

Opposed

For entrepreneurs and non-professional coders, AI tools are genuinely liberating — enabling solo MVP development, startup resurrection, and rapid iteration that previously required expensive dev teams
Automation has historically improved quality of life in aggregate: cheaper goods, reduced household labor, and higher overall prosperity, even if gains are unevenly distributed
The problem is capitalism and management culture rather than technology itself — the same dynamics existed before AI with any form of process improvement
Workers can strategically capture automation benefits by keeping their efficiency gains private, treating salary as a retainer for expertise rather than payment for hours
AI is still early and improving — just as modern detergent simplified laundry, AI validation tools will mature and reduce the verification burden over time