Zero marginal effort

Zero marginal effort


The internet enabled a new digital economy where the cost of increasing the production of digital goods is essentially zero. However, the notion of 'zero marginal cost' extends far beyond selling. Zero marginal effort means scaling activities that would otherwise cost time.

The internet enabled a new digital economy where the cost of increasing the production of digital goods is essentially zero. Beyond the initial expense of making the digital good (music, art, etc), replicating that good adds no marginal cost. This is in stark contrast to physical goods, whose marginal costs are modeled and closely monitored. However, the notion of 'zero marginal cost' extends far beyond selling more mp3s, jpegs, or other digital artifacts. Zero marginal effort means scaling activities that would otherwise cost time.

Zero marginal effort means multiplying your efforts, without the need for other people and no extra cost. As one example, I sit on a committee to invite speakers to a national conference. I was recently asked to email thirty people to invite them to the conference, with details about their session, time of presentation, suggested title, and a myriad of other details. The worst part, all of this information was included in a PDF, which also needed to be attached to the email invitation. I was annoyed, to say the least. Partially out of spite, and partially because I'm stubborn, I set out to automate it. I wrote a script that scraped the information out of each PDF, stuffed each into a table, auto-composed an email based on a template I created that would populate each email with the relevant details, and finally sent it from my email address. How much time did it take? A couple of hours. Could I have probably manually composed thirty emails in about that much time? Yes. Did I have the satisfaction of building an auto-emailer that didn't resort to manual emailing? Also yes.

Thanks to Ben Neely, I found this handy chart by xkcd that quantifies when your 'zero marginal effort' output surpasses the time you put into it. Candidly, my auto-emailer was borderline. That is, until the thirty emails ballooned into 50 over the next few weeks. I hadn't anticipated how many people would be unable or unwilling to travel to a national conference, and so needed to send several follow-up emails to the 'alternates'. The additional time to send these? Zero. Ok, technically 15 seconds, but certainly less than 1 minute so I rounded down.

The idea of zero marginal effort is not about doing more. Instead, it's about being more efficient to make time for things that matter. As I've written before, building systems allow you to work smarter, not harder. Doing something with less effort allows more time for important things. Because sending 30 emails is never the important thing.  

Two things make zero marginal effort possible. First, learning to write code. This is not why I learned to code (I wanted to solve a simple problem), but this is why I believe everyone needs to learn to code. Being able to give a computer instructions that can be followed with zero marginal effort and zero marginal cost enables you to scale. Which programming language is best? As I've said repeatedly, it doesn't matter. I write most code in an academic-centric language called R, which has a developer ecosystem that allows me to do anything I need: analyze data, build visuals, make websites, scrape websites, make art, render 3D images, and beyond. Second, in addition to programming, building systems around your projects creates focus and streamlines activities. Together, this allows me to scale my efforts. And all doing it without anyone's permission.

But of course, not everything can be scaled with zero marginal effort. This is especially true in science: writing grants (having and then articulating a good idea), making discoveries (testing and validating hypotheses), writing papers (sharing and contextualizing new knowledge) — quintessential human activities that require creativity, context, and empathy.

This is not an article on "computers taking over jobs". In fact, it's just the opposite. Identifying ways to scale with zero marginal effort is a collaboration between humans and machines. Working with the computer meant I had more time for important things instead of copy-pasting emails.

Looking forward, collaborations with computers are poised to scale our efforts in ways we cannot yet imagine. And is part of the reason I'm bullish on Web3. As I wrote before,

But the Web3 excitement is around building an entirely new platform. The iPhone enabled new gestures, new ways of interacting with a device, and a new OS for developers to create upon. Most uses cases for Web3 can not even be imagined yet. But entirely new infrastructure will facilitate experiments and innovations.

Scaling with zero marginal effort is akin to a new platform — a new way of thinking and approaching work. When we get past the focus on efficiency gains and instead focus on collaborations with technology, new experiments and new innovations thrive.

One active area we're experimenting with is scaling education. In my work with the Center for Computational Thinking, our mission is to enable more people to learn computation, both the technical skills as well as digital literacy that are required to be successful in today's digital world. How can we scale from 30 students in a classroom to 30,000 undergraduates, graduates, staff, and faculty across a University? By collaborating with technology. By scaling with zero marginal effort and zero marginal cost. By capturing value from instructors and sharing broadly. By thinking deeply about students' time, learning styles, and activities. By experimenting with new platforms and new approaches. We don't have the results of these experiments yet, but this is our thinking.

The second example of technological collaborations is in the Web3 space. Profile pictures (PFPs) are digital art avatars that signal community membership, interests, and status. Early PFP projects, like the cryptopunks and bored apes, generated 10,000 unique avatars, which are code-based, combinatorial projects of images with various traits. Since those early projects, several other NFT projects have emerged that showcase technological collaborations.  

Each of the images above was programmatically generated, starting with artist-defined building blocks, and then allowing a computer to randomly assign the combination. Pablo Picasso is considered one of the most prolific artists in modern history, who is estimated to have made almost 15,000 paintings over his 75-year career. Today, an artist can create a 10K project in a few dedicated months.

Leverage allows you to accomplish 100x what others can – it is a force multiplier for your effort, skill, and judgement. Today, we live in an age of infinite leverage where our work can be replicated at no cost, and we don’t need permission to get started. – Jack Butcher, The Age of Infinite Leverage

The idea of building systems to scale effort is not about doing more. Instead, zero marginal effort saves time for things that are important and enables new experiments and new products, which were not previously possible.