AI Isn’t Breaking Work. It’s Already Broken.

In one sentence

AI acts as a "magnifying glass" for pre-existing organizational inefficiencies, exposing the fact that our current systems for managing work—coordination, context, and judgment—are already broken.

Why read this (and who it's for)

This piece is for leaders, managers, and high-output workers who feel overwhelmed by the "noise" of digital tools and the hype of AI. It solves the frustration of "doing more but achieving less." After reading, you will be able to distinguish between pseudo-productivity (performing work) and systemic productivity (achieving results), and you will be able to identify where "productivity drag" is slowing down your team.

The big idea

The prevailing narrative is that AI is a revolutionary force that will fundamentally break the way we work. Newport argues the opposite: the problems we see today—toggling between apps, waiting for responses, and performing "work" for others to see—existed long before LLMs. AI is simply accelerating these pre-existing issues.

The core issue is a failure of Systemic Optimization. We often focus on "local" gains (how much faster can one person generate a draft?) while ignoring the "systemic" costs (how much more coordination, verification, and rework is required to make that draft useful?). Because our organizational systems for sharing context and exercising judgment are underdeveloped, increasing the volume of output via AI often leads to a net increase in noise rather than a net increase in value.

The "silver lining" is that because AI is so new and exciting, it provides a unique "Trojan Horse" opportunity. Because leaders are currently obsessed with AI, they may finally be forced to look at the underlying organizational structures that have been broken for years.

Key ideas

The Illusion of Time Savings (Botsitting & Toggling)

• What it is — Workers often overestimate AI's impact because they fail to account for "botsitting" (waiting for AI agents to complete tasks) and the "toggling cost" of switching between multiple tools to get a usable result.
• Why it matters — If you don't account for the "hidden" time spent managing the tool, your productivity gains are an illusion.
• In the book — Newport cites a survey of 6,000 digital workers where respondents claimed AI saved them 11 hours a week, yet only 13% reported any improvement in company performance.

Workplace Theater (Pseudo-Productivity)

• What it is — This is the act of "visibly performing work for bosses and colleagues" rather than focusing on the actual grind of getting things done.
• Why it matters — It creates a culture of "busy-ness" where the appearance of effort is rewarded over the actual completion of meaningful tasks.
• In the book — Newport notes that this isn't new; it was a central theme in his book Slow Productivity, where he refers to it as "pseudo-productivity."

Productivity Drag

• What it is — The extra cognitive load and time wasted when a recipient has to sift through "textual fluff" or half-baked ideas generated by AI that haven't been properly reviewed.
• Why it matters — AI can actually decrease productivity if the output requires more human effort to "clean up" than the original task would have taken to perform manually.
• In the book — A reader example highlights a coworker using AI to turn 4 bullet points into 4 paragraphs of fluff, forcing the recipient to spend more time hunting for the actual action items.

Local vs. Systemic Optimization

• What it is — The difference between making an individual task faster (local) and making the entire workflow more effective (systemic).
• Why it matters — Faster production of outputs does not translate to better results if the requirements for verification, coordination, and rework increase at the same rate.
• In the book — Newport argues that AI is a "multiplier." In a well-designed system, it amplifies effectiveness; in a poorly designed system, it amplifies confusion and coordination costs.

The Opportunity of the "New"

• What it is — Using the current AI hype cycle as a catalyst to fix old, broken organizational problems.
• Why it matters — It is easier to get leadership buy-in for "fixing the system" if it is framed as "optimizing for AI."
• In the book — Newport suggests that as leaders try to understand how to make AI effective, they may finally recognize the systemic flaws in how information, judgment, and responsibility flow across the organization.

Frameworks & tools

The Systemic Optimization Audit

Use this to determine if a new tool (like AI) is actually helping or just creating more "noise."

1. Identify the Local Gain: What specific task is the tool making faster? (e.g., "Writing a summary").
2. Calculate the Systemic Cost: How much more time will be spent on:
   • Verification: Who has to check the output for accuracy?
   • Coordination: Who needs to be informed of this new output?
   • Context-Sharing: Does the output actually contain the necessary "why" behind the task?
3. Measure the Net Result: If the Systemic Cost $\geq$ Local Gain, the tool is creating "Productivity Drag."

The "Workplace Theater" Filter

Apply this to your daily schedule to ensure you are doing "real" work.

• Is this visible? (Am I doing this so others see I'm busy?)
• Is this impactful? (Does this move the needle on a core objective?)
• Is this "Pseudo-Productivity"? (If the answer to both is "No," stop doing it.)

Memorable examples & stories

The 11-Hour/13% Paradox

A survey of 6,000 digital workers showed that while workers felt they saved 11 hours a week using AI, only 13% saw any actual improvement in company performance. This illustrates the gap between "feeling" productive and actually being effective.

The "Productivity Drag" of AI Emails

A worker uses AI to turn 4 bullet points into 4 paragraphs of text. The recipient now has to spend more time "hunting" through the fluff to find the actual information than they would have spent reading the original bulleted list. This teaches us that "more" is not always "better."

The Worker A vs. Worker B Example

A comparison of two workers: Worker A is efficient and knows the shortcuts; Worker B is inefficient but has more "time" to spend with AI. The lesson is that if a company only has one task per day, the company cannot be more productive just by giving Worker B a tool; the system itself must change to allow for higher-volume output.

Key quotes

"AI isn’t so much creating new problems as it is magnifying the types of problems that have long existed." Why it matters: It shifts the focus from "fixing the AI" to "fixing the organization."

"Organizations have long underestimated the coordination, context-sharing, and judgment required to turn individual productivity gains into system-level performance improvements." (Paraphrased) Why it matters: It highlights that "speed" is useless without "direction" and "context."

"AI is less a replacement for organizational capability than a multiplier of it." (Paraphrased) Why it matters: It reminds us that the quality of the output depends entirely on the quality of the underlying system.

How to apply it

This week

• Audit your "Toggles": Track how many times you switch between apps/tools to complete a single task. Identify which tools are causing "toggling drag."
• Identify one "Productivity Drag" moment: Find an instance where an AI-generated (or automated) response actually made your job harder because it was too long or lacked context.
• The "Theater" Check: Look at your calendar. Identify one meeting or task that is purely for "visibility" and see if it can be replaced with a brief, high-context update.

Ongoing

• Prioritize Systemic Flow: When evaluating new tools, ask: "How does this affect the flow of information and judgment across the team?" rather than "How much time does this save me?"
• Context-First Communication: To combat "Productivity Drag," always prioritize bulleted lists and clear action items over "AI-style" fluff.
• Practice "Slow Productivity": Focus on the quality of the system (the "how") rather than just the speed of the output (the "how much").

Common pitfalls

• The "AI-Skill" Fallacy: Assuming that because a worker is "good at prompting," the organization will automatically become more productive. Newport argues that without systemic changes, this just leads to more "noise."
• Ignoring the Human Cost of Reading: People often forget that someone has to read the AI-generated content. If the content is low-quality, the "cost" is shifted to the reader.
• Over-optimizing for Local Speed: Getting caught up in how fast you can generate a draft while ignoring the fact that the draft might be useless without proper human judgment and context.

Connects to

• Slow Productivity (Cal Newport): The foundational text for understanding why "faster" isn't always "better" and how to manage the "pseudo-productivity" of the digital age.
• Bullshit Jobs (David Graeber): Relates to the concept of "Workplace Theater" and the systemic issues of why certain jobs exist only to perform for others.

If you remember only three things

• AI is a multiplier of existing organizational flaws, not a cure for them.
• Systemic Optimization (coordination, context, judgment) is the bottleneck, not the speed of task completion.
• Productivity Drag occurs when the cost of processing "AI fluff" exceeds the benefit of the automated output.

Internalize it

• Reflection questions
  • In what ways am I currently "botsitting" or "toggling" instead of doing deep, focused work?
  • Is my current workflow optimized for my speed, or for the system's ability to process my output?
  • When was the last time I felt "Productivity Drag" from a piece of communication that was too long or lacked clear action items?
• Flashcards
  • Q: What is "botsitting"?
  • A: The time spent waiting for AI agents to complete tasks, which is often excluded from "time saved" calculations.
  • Q: What is "Workplace Theater"?
  • A: Performing work for bosses and colleagues (pseudo-productivity) rather than focusing on the actual grind of getting things done.
  • Q: What is "Productivity Drag"?
  • A: The extra cognitive load and time wasted by a reader when dealing with AI-generated "fluff" or unreviewed content.
  • Q: Why did only 13% of workers report improved company performance despite saving 11 hours a week?
  • A: Because local optimization (individual speed) does not automatically translate to system-level performance if coordination and verification costs increase simultaneously.
  • Q: How does AI act as a "multiplier"?
  • A: It amplifies the existing design of the system: it amplifies effectiveness in well-designed systems and confusion/drag in poorly designed ones.

Fidelity note

This summary is based strictly on the provided blog post and the comments/replies included in the text. I am confident in the definitions of "botsitting," "workplace theater," and "productivity drag." The "Worker A vs. Worker B" example and the "Productivity Drag" specific definition were derived from the reader comments provided in the source material. The "Systemic Optimization" framework is a synthesis of Newport's argument regarding the 13% performance statistic.