Shallow Work: Be Aware of Striving Over-Productivity

Be aware and careful about «Shallow Work» (also called pseudo work or meta-work).

With all the AI hype and over-productivity we (suppose) to get with OpenClaw and others, we need to be careful not to do pseudo-work all day long and not finish our actual tasks.

Reading your emails, calendar, and all your chats, creating a summary, checking your work, all your stuff while chatting with a small chat interface on WhatsApp or Telegram with your agents, isn’t actually saving you time, except if you have 1000 of them. But maybe then something in your setup is already wrong?

Because a calendar can be scanned instantly, it takes literally 2 seconds for the full week. Emails shouldn’t be an issue or an urgency in daily matters anyway. Anything that comes in via email should be responded to within 2-5 days.

Also, the additional work of checking whether the agent is working correctly, not missing anything, or when sharing your information, leaving out the most important to you, or deleting your inbox might be better spent saving that mental energy for the actual work at hand.

So what is Shallow work actually?

Cal Newport’s concept of “Shallow Work” refers to non-cognitive, logistical tasks (like email, meetings, admin) that are easy to replicate and done with distraction, creating little new value, contrasting with his idea of “Deep Work,” which is focused, cognitively demanding work that creates significant value, with Newport advocating minimizing shallow tasks to maximize deep, meaningful accomplishment.

He also calls it pseudo work here:
The Straight-A Gospels: Pseudo-Work Does Not Equal Work - Cal Newport

“Shallow Work: Noncognitively demanding, logistical-style tasks, often performed while distracted. These efforts tend to not create much new value in the world and are easy to replicate. In an age of network tools, in other words, knowledge workers increasingly replace deep work with the shallow alternative—constantly sending and receiving e-mail messages like human network routers, with frequent breaks for quick hits of distraction.

Larger efforts that would be well served by deep thinking, such as forming a new business strategy or writing an important grant application, get fragmented into distracted dashes that produce muted quality.” Quote by Cal Newport: “Shallow Work: Noncognitively demanding, logisti…” from his book Deep Work (Cal Newport)

From Use less AI:

[..] Notice what all of these use cases have in common: they’re meta-work. Work about work. Organization about organization. The person using Claude to analyze their weekly logs for “productivity leaks” has found a sophisticated new way to avoid doing the thing in favor of doing the things about the thing. Claude Code Won’t Fix Your Life:

The danger of Relying Too much on AI is another danger. As well as we might unlearn how actually learn new things with the constant solution getting presented instantly. No hardship, lessons learned along the way.

As said Will AI replace Humans, it’s not only bad, but we need to be weary of meta work, and avoiding actual work, even more so that AI can make you addicted most simlar to social media, with the constant Dopamine Hit that you’ll do something productive.

When we write more emails with AI, we also use more AI to summarize them. So instead of talking human to human, our communication gets filtered through a layer of AI. We lose the human touch, or real human communication.

Source: AI Written, AI Read cartoon - Marketoonist | Tom Fishburne

• I talk much more in Will AI replace Humans
• Vibe Coding and the the Jevons Paradox cost or running it.

Origin: Cal Newport
References: Jevons Paradox