Worried AI is going to take your job? You’re asking the wrong question

Lessons from the book Reshuffle on staying above the algorithm

Most people are asking whether AI will automate their job. That’s like asking whether the internet would replace typewriters - technically true, but completely missing the point.

Sangeet Paul Choudary’s Reshuffle reframes the question entirely: It’s not about which tasks AI automates—it’s about your position in the system AI is reshaping.

My biggest takeaway from the book: workers operate either above or below the algorithm.

Below the algorithm, you become an interchangeable input. Any productivity gains you make go to the owner of the system—think Amazon warehouse workers whose every action is directed by the same algorithm that also captures all the value created by the increased efficiency. These are just warm bodies doing tasks not yet replaceable by machines.

Above the algorithm, you own the points of control. You rebundle the capabilities that AI has unbundled from experts. You manage the constraints of risk and coordination. You capture the value as systems reshape.

This distinction matters, Choudary argues, because AI doesn’t just automate tasks—it unbundles entire jobs into modular capabilities. When expertise is unbundled from experts, you can rent, recombine, and scale knowledge.

A freelancer doesn’t need a full-time lawyer when AI legal tools handle most contracts. Writing becomes available on-demand through Claude and ChatGPT—I used Spiral to help me quickly produce this very piece. Design and launch of full-fledged apps and websites are now possible for people with little-to-no programming experience.

My own unbundling moment

I experienced this firsthand earlier this year. In my day job as a strategist and data analyst, I spent 15 minutes manually formatting bubbles in a chart before realizing: I could do this in Python. Problem was, I’d never really gotten Python to stick. I knew what was possible, just not how to actually do it.

So I tried Gemini. Told it what I wanted. Ten minutes later, I had a working script. What would’ve taken an hour took ten minutes.

I felt like I’d unlocked something.

A few days later, I faced a massive spreadsheet update. Formulas crashed the file. I went back to Gemini for a Python solution. After an hour of trial and error, I was ready to quit. Then I realized: the problem wasn’t the AI. It was me. I didn’t know enough Python to even ask the right questions.

So I backed up. Worked on the smallest possible step. Then the next one. Slowly, it clicked.

That’s when it hit me: AI really is just another tool. Sophisticated, yes. But still a tool—with strengths, weaknesses, and a learning curve. You have to experiment. You have to be willing to fail. And you have to treat AI outputs as starting points, not final answers.

That experience taught me something that applies beyond just my workflow: AI democratizes capabilities the same way earlier tools did.

When smartphones first got camera apps, instead of killing “real” photography, they let more people take great photos without understanding aperture, shutter speed, or exposure. The app made photography accessible without replacing the photographer’s eye.

AI works the same way. It removes technical barriers but can’t replace your judgment, expertise, or creativity. And if you’re in your early 40s like me, you’ve already adapted to new technologies—browsers, smartphones, spreadsheets. AI is just the next tool in that line.

Once you grasp that AI is learnable—not some god-like entity—Choudary’s system-level implications become clear. Individuals can now remix capabilities on the fly to build unique businesses that solve emergent problems. But only if you structure those businesses with AI as the engine—every workflow and process designed to leverage AI’s strengths and work around its weaknesses, not bolted onto old ideas.

Thinking in systems, not tasks

Most discussions focus on which tasks AI can automate or which jobs are “safe.” The real questions are systemic: How are constraints changing? What is AI unbundling and rebundling? Where is value migrating?

When AI removes old scarcity constraints—like access to specialized knowledge—two new constraints emerge: risk and coordination.

A Harvard Business School study Choudary cites proves this: AI augmentation narrowed the performance gap between low and high performers from over 20% to just 4%. Skill premiums—the extra value specialized expertise once commanded—collapsed. Economic value shifted from workers to tool providers and from scarce skills to capabilities that manage the new constraints.

Choudary’s practical framework:

1. Map your work as tasks with intrinsic, economic, and contextual value. Does this matter if the market vanishes? What does scarcity command? What constraints does it resolve?

2. Identify what AI is unbundling in your domain. Where does coordination get harder? Where does risk increase? Skills matter only in relation to the constraints they resolve.

3. Rebundle your role around the new constraints. The opportunity is designing your position so the system elevates you rather than replaces you.

If you’re struggling to grasp AI’s potential, the key is to start using it. Pick something small—a research task, a quick experiment with ChatGPT. Treat it like any other tool you’ve learned: try, fail, adjust.

Designing your position

The book’s most provocative claim: “If you’re asking what the future of work looks like, don’t keep looking for what AI can’t yet do. Instead, ask what it breaks in the system.”

AI’s greatest impact won’t come from automating tasks but from lowering coordination costs so dramatically that innovation can scale without central control.

When AI converts tacit organizational knowledge into explicit, searchable assets, it breaks coordination bottlenecks. Small teams can deploy capabilities that previously required large organizations. Solo entrepreneurs can operate at scales once available only to companies with dozens of employees.

The old moats Choudary identifies—scarce expertise, specialized credentials, years of experience—are crumbling as AI commoditizes knowledge. The new moat is leverage: the ability to deploy, recombine, and scale capabilities in ways others can’t.

Reshuffle doesn’t predict the future of work. It gives you the questions to design your position in it.

Don’t ask whether AI will take your job. Ask whether you’re positioned above or below the algorithm—and what you need to rebundle to stay there.

---

A note on AI collaboration: I used Obsidian Copilot (a plugin for the popular note-taking app) to help give my thoughts structure and Spiral (an AI writing co-pilot) to help produce and refine this article—practicing the experimental, tool-first mindset I’m advocating for. The ideas, stories, and frameworks are mine; Copilot lowered the barrier to documenting my thought process while Spiral helped me articulate them more clearly.