📈 Compounding work (work that keeps making you money)

Hello there 😎

I've been circling around a definition for "compounding work" for the past five years.

In 2021, I first started consulting on the side. In May that year, I wrote:

None of those things I listed was the deliverable anyone paid me for. They were incidental. They were the leftovers.

Two years later I tried to give the leftovers a name:

Examples:

Reforge. I still have early drafts of our Reforge courses. I also have early slide templates, and from 2021 to 2023 I used them a lot. They were great for rebuilding a framework and explaining concepts.

Noom. I kept all my "Ddocs," the experiment write-ups that the growth team wrote and pored over every week. I still use them as a reference guide on writing and analyzing experiments.

I Will Teach. Early on, I Will Teach taught me how to write positioning documents that captured the hearts and minds of the people you're building for.

Over the years, these reusable components have saved me hundreds of hours across W2 roles and as a consultant.

♻️ None of this is new

Jason Fried has his own name for these. He calls them by-products.

To summarize: in 2009, Fried wrote about how the lumber industry sells what used to be waste, the sawdust and chips and shredded wood, for a pretty profit. It ends up in fireplace logs, particle board, mulch, pet bedding. His own company set out to build Basecamp and got a book and Ruby on Rails out of the scraps.

In the physical world, by-products or reusable components are things like sawdust and woodchips.

In the digital world, they're unused code, swipe files, or checklists.

Either way, workers have held onto these forever for the same reason: save time and make more money.

⚡ How AI has changed the power of reusable components

Take the Reforge courses again.

Before AI, what was valuable was access to the courses. It's why we were a company valued at $100m that sold $2,997 courses.

In an AI world, what's valuable is everything that goes into building the courses: the slides and frameworks, the subject matter expert interviews and the relationships behind them, and the learner feedback telling you what was useful, what wasn't, and what they want to see next.

If you have those, you can spin up any number of courses, each one hyper-personalized to a single learner.

That's the flip. The deliverable is the commodity now. What's valuable are the inputs that created it.

AI didn't make work compound. Work always compounded. What changed is which half holds the value.

It also changed how we should think about reusable components. You can bucket all reusable components into three types:

All three work in order. The sources shape what you understand, context acts as a guardrail for building, and together they steer the workflow.

Two more quick examples.

🍽️ The meal planning system

Two weeks ago I sent you my meal-planning system. In three parts:

Sources. About thirty mains tagged in Roam. Next week's Home Appétit menu. Two Google Calendars, family and life. The AnyList we shout at Alexa all week.

Context. A night with something between 4 and 8 isn't a cook night; it's slow cooker or takeout. Same for weekends. What the kids won't eat. For grocery stores, we shop at Aldi first, ACME as backup. When to give me a new slow-cooker recipe.

Workflow. Thursday morning, it reads the calendar and the menu and emails me two blocks of options. I reply with numbers. Thursday afternoon it fills the carts and stops before checkout.

By Sunday, the cart and menu this system produced for me are worth nothing. But the three-part system gets more valuable to me every week.

🎓 Clay University

I build courses for Clay's ecosystem (think: education) team. Same three parts:

Sources. Our old university videos. Loom walkthroughs from the product managers who build different features.

Context. Why a product decision went the way it did. Why something shipped half-baked, and whether that's temporary or just what it is. Where the company is heading next. Dozens and dozens of Slack messages. Hundreds of Notion files documenting how everything else at Clay works together.

Workflow. One saved workflow reviews a product video. One turns a recording into a transcript. One turns a transcript into a written guide. One turns an outline into scripts. One helps me edit the video.

This system has been evolving for the past 15 months and counting. It started with brain dumps and a lot of manual writing. Then we did it through a series of Clay tables. Now, it's Notion and half a dozen skills.

Today, the system is the most valuable part of the whole process.

💼 What does all of this mean for a portfolio career?

First, what's a portfolio career? I wrote the long version here, and I'll update it soon.

For now, a few takeaways that are top of mind:

1/ Context switching went from impossible to hard. Working with two fast-moving AI companies... It's not easy. There's still a cost every time I have to switch mentally. But now the context lives in Markdown files instead of all in my head, which makes what would be impossible (for me) possible. (Still difficult, though.)

2/ Cheat whenever you can. Matt Greenberg gave me that advice when I was at Reforge. I still reference it a lot. He didn't actually mean cheat; he meant leverage any and all advantages you have. Pre-AI, the main lever was transferable knowledge, i.e., for me, education, courses, and learning. Post-AI, it's transferable knowledge and those reusable components: sources, context, and workflows. If you have those, keep doubling down on them.

3/ Take even more shots on goal. It's cheap to launch and cheap to test, so the right number of side experiments you should launch has gone up. If you have any idea you're interested in, you can test it in a weekend.

4/ Do things in parallel, not in sequence. For a portfolio career, I wrote that you should start with a service and productize it later. This is no longer true. These workstreams can happen in parallel now. Your ability to context switch is far greater (see 1 above). The real constraints are 1/ how well you can use the tools and 2/ your ability to execute.

5/ I'm more bullish on a portfolio career. It's still a two-way door, and it swings fast. If your AI skillset is sharp and you have access to reusable components, leaving a company costs you less than it used to. You can build something for a client or a new employer faster than ever.

🌍 Remote work was a step change. This is the next one.

Remote work meant we could decide where to live and handed back the hours we used to spend trapped in a metal box shuttling from point A to point B. Nothing since has come close until AI changed the math behind compounding work.

If a portfolio career is something you've wanted to pursue, now is the time.

🤝 When you’re ready, here’s how I help you land your next job:

I love meeting new people. Say hi on LinkedIn 👋