Series: The machine that never sleeps #

1. AI is making us work more (Current)
2. AI and the Devaluation of Effort
3. The Sisyphean Task of Trying to do Everything

Intro #

I was listening, recently, to an episode of The Pragmatic Engineer podcast with Armin Ronacher, and something he said really resonated with me. He pointed out how paradoxical it is that AI was supposed to free us and allow us to work less yet, somehow, we find ourselves working more than ever before (timestamp).

I have seen the same trend in my own work patterns. This increase in work is not necessarily in the sense of being busier because there’s more work to do. Instead, it manifests as a psychological compulsion to keep going. The very existence of these hyper-capable tools seems to have created a new kind of pressure, an existential duty to keep working, to keep driving the machine that never sleeps.

996 in the Valley #

The 996 concept - working from 9AM to 9PM, six days a week - used to only be associated with Chinese tech companies, having been championed by the founder of Alibaba, Jack Ma. The rest of the world, however, used to see it as nothing more than an extreme, sustainable model of hustle culture.

In 2025, however, there have been several reports of this (toxic?) work culture migrating west.

According to a recent Wired investigation, AI start-ups in Silicon Valley have begun to adopt 996-style work schedules. The rationalisation, of course, is “to stay competitive”. Leaders at these companies believe that with things moving so fast, especially in the AI space, and with developers having access to tools that can keep going so long as there’s a human there to steer them, people must worker harder to keep up with that. It has become not at all unusual to see job postings that make it clear to prospective hires that they should look forward to long, gruelling hours.

When the machine never tires #

Throughout human history, the natural constraint on work was always human fatigue. We stopped working when we were tired. And, we took time to rest because our minds and bodies needed to recover.

However, the introduction of generative AI and agents fundamentally shifts the status quo. These tools never tire, never switch off, and never lose motivation. even as you rest, the tool keeps running, always available to generate ideas, code, text, designs, etc.

As a result, a new psychological loop forms:

“Every moment I don’t spend prompting, I’m falling behind.”

The system insidiously guilts you for not leveraging it constantly. Allowing AI to sit there, just waiting, feels like a waste. It’s subtle, but corrosive. Any downtime becomes a missed opportunity, and rest turns into inefficiency. Within this framework, leisure becomes a moral failure.

From the lamp to LLMs #

Of course, this is not an entirely novel development.

From lamps to lightbulbs, and now with LLMs, certain human advancements have had the capacity to transform the nature of work. This transformation comes in the form of leverage, allowing us to do more, to transcend limitations, and manifest more than we ever could. In the case of the aforementioned illumination technologies, people who had to set their tools down at sundown because it was now too dark to work could now go on well into the night.

And, with that came a subtle shift as “can work” transformed into “should work”. Such is the nature of advancement - recently acquired luxuries quickly transform into necessities, and pushing our newfound capabilities to their limits becomes an expectation.

With this, the 996 culture re-emerges, not because someone imposed it but because the tools themselves, and the culture surrounding them, make rest feel like lost potential

“One of history’s few iron laws is that luxuries tend to become necessities and to spawn new obligations.” - Sapiens: A Brief History of Humankind Book by Yuval Noah Harari

The Burnout Society, Upgraded #

The Korean-German philosopher, Byung-Chul Han saw this coming. In his book, The Burnout Society, he argued that modern work culture replaced external oppression with internalised “self-discipline”. We don’t need anyone shouting orders at us anymore. Instead, we do it to ourselves in the name of “productivity”, “passion”, and “self-actualisation”.

AI amplifies this dynamic. It extends what Han calls “the excess of positivity” - what I would call the tyranny of can - which is the idea that we can always do more, and therefore must.

“Possum, ergo debeo.”
“I can, therefore I must.”

There is no external whip required here. Just a quiet, internalised command: the machine can keep going, so why can’t you?

The Psychological Toll #

The recurring narrative within the tech community is mostly the same: infinite tools + infinite leverage = infinite potential. But, I find this to be a corrosive mentality with potentially devastating effects on our psyches.

Where feeling tired used to be a signal to rest, now it’s a sign of weakness. Every break you take feels like a gap in your potential productivity.

What we become, then, is our own taskmasters. From the moment we wake until our heads hit the pillow in the evening, we must make use of every available moment. We internalise the myth of this infinite potential and, under the guise of our own search for freedom, we exploit ourselves, putting constant pressure on ourselves to prompt, iterate, and improve. All to keep up with a machine that never sleeps.

Sadly, as a recent NextGen Hero analysis pointed out, this “hyper-productivity loop” is self-defeating. As burnout rates climb, creative output falls, and the majority of teams working 996 often become less instead of more innovative when compared to balanced ones.

And, as the machine continues to improve, so too does the baseline expectation of what you should get out of it. What once looked like “good output” becomes “not enough”. We struggle to improve our prompts, and mould the deluge of mediocre output into something useful.

Meanwhile, as a consequence of how the loudest proponents of this technology speak about it, we absorb the belief that we can (and should) be orders of magnitude more productive, we can (and should) build lucrative businesses, and we can (and should) multiply our abilities with AI tools. Any kind of doubt, scepticism, or negativity is shunned. Any failure on your part to use AI tools to elevate yourself is, at best, a skill issue and, in the worst case, stubborn refusal to embrace a revolutionary technology.

Rest as resistance #

In such a world, rest itself becomes an act of rebellion. If AI gives us near-limitless productive capacity, the truly radical act might be to not use it at all. By setting realistic targets for ourselves, and putting up boundaries, we can slow the erosion of our humanity.

Similarly, AI creates nigh-infinite consumption potential. OpenAI’s Sloptok can generate endless content, which could just as easily trap us in a Sisyphean-loop of consuming limitless content.

In itself, saying “enough” might be a kind of innovation.

Because, ironically, most research supports the idea that innovation and creativity emerge from reflection not exhaustion. Our best ideas usually come not from being “on” all the time, but from being free enough to ponder and let our minds wander.

So, what now? #

The way I see it, this is ultimately a story about culture.

Tools don’t demand work; people and systems do. The machine doesn’t care whether or not you sleep. What keeps us awake are the stories we tell ourselves about what it means to be “productive”.

So, maybe, the challenge of the AI era isn’t just learning how to leverage these tools. It’s learning to live with them in a healthy way that doesn’t allow them to enslave us. And, as I will explore in the second part of this series, it’s also knowing when and how to use them in a way that doesn’t diminish our efforts.

Because, just as the light bulb extended the day, AI is extending the workday. The question is whether we’ll allow it to extend the night, too.