May 27, 2020

Devs have eaten the world

Rodrigo Mendoza-Smith

Rodrigo Mendoza-Smith

•

12 min read

And still have room for dessert 🍰



We’ve all heard about Marc Andreessen’s famous essay “Software is eating the world”. The essay speaks about the increasingly evident hyper-growth opportunities in the internet and software industries at the beginning of the previous decade. Nine years after its publication, software has finally eaten the world. Smartphones, online banking, and Uber are now mainstream. User interfaces are more fluid than ever and have delivered a degree of human-computer interaction that was not long ago considered futuristic. Software products and services can now be built, delivered, and scaled at an unprecedented rate. The effects of this technological bonanza are well-known to all: TikTok, Bitcoin, AirPods, billionaires. The causes are often attributed to the cooccurrence of key macroeconomic events like the democratisation of software education, the advent of Amazon EC2, and the statistical viability of the startup model. Though these events tell most part of the story, none would have been possible without the time and effort from the people that created, maintained, and popularised the stack over which syllabi, servers, and startups stand. I’m not talking about business icons like Bill Gates or Steve Jobs (although they deserve a honorary mention), but the group of hobbyists, hackers, and programmers that created the tools our economy now depends on — a group commonly know as the open-source community.

On the economic paradoxes of open-source

Though the genesis of open-source can be traced back to the 1950s, its triumph as a paradigm for open and decentralised collaboration is relatively recent. In terms of the economic boom of the 2010s, its importance is multifaceted. To start with, it built and socialised the software tools, coding conventions, and workflow patterns that developers around the world now work under. Thanks to open-source, devs became a globalised workforce that wield the same toolset and parlance. At the same time, it helped programming languages reach a level of universality that is almost comparable to mathematical notation. This didn’t happen overnight. The programs we now summon with yarn, brew or pip are the result of a long natural selection process which subjected their architecture, structure, and APIs to savage competition. During this evolutionary course the most capable projects were not necessarily those that secured a committed pool of maintainers, but those that made their way into the guts of other community projects and turned themselves into dependencies. This created a positive-sum game - a sort of mutualism between codebases in which the guest builds network effects by absorbing parts of its host’s computational skeleton thus making it cleaner, faster, and more maintainable.


Modularising code in dependencies has saved humanity billions of years in development time, but unfortunately under a negative-sum game. Why? Because maintaining dependencies requires developer work which, in an economy that is based on software, is as valuable as gold. A number of monetisation models for open-source work have been put forward in the last decade, but none of them has been able to scale and solve the problem in a fundamental way. The reason being, open-source work is a microeconomic singularity — a paradox in capitalism that can’t be corrected with donations, cryptocurrencies, or freemium models. Indeed, the open-source ecosystem produces economic value at scale, but this value can’t be back-propagated to the creators. The reason for this is more philosophical than technological: if a fair and transparent mechanism to re-distribute value existed, it would incentivise creators to organise around investable legal entities to more efficiently capture a greater chunk of the dependency market share. This would clearly antagonise the values behind the movement and vilify its spirit with for-profit aspirations. Why then, do open-source contributors continue to spend time working on these projects? This is a fundamental question. In some cases, incurred opportunity costs can be mapped to clear long-term incentives like learning new skills, work opportunities or even access to the venture-capital markets. However, in general, motivations are rooted in deeper human aspirations like the willingness to give back to the community, indulging into intellectual pleasure, or even the ego gratification associated with building stuff that peers may find cool or clever.


Open-source is the fabric of our digital cosmos and the raw material software products and services are built with. Our supply chain is built on software, yet code is not a fungible asset one can generally trade or exchange. What then is the real source of value in our software economy? It turns out that the key commodity sustaining our digital world is far from cybernetic. It’s an abstract resource that countries are desperately trying to produce and companies are fiercely fighting for: developer time.

Macro trends

The work of code creators has percolated into every corner of the supply-chain. They have built the databases supporting our healthcare infrastructure, the ML algorithms driving our Teslas, and the computer graphics engines rendering our video-games. In doing so, a number of powerful macroeconomic trends have started to unfold:

The barriers of entry to software development have significantly lowered. MOOCs, Q&A platforms, and coding bootcamps have democratised software engineering knowledge and evangelised open-source tools. Alternative education models have succeeded to the point that the credential-driven model for talent validation has began to shatter. College graduates and high-school dropouts are now competing for jobs that only 15 years ago were reserved for CS PhDs.


The industry has accepted that, for software engineering jobs, remote work works. This trend has been further validated by COVID-19 work-from-home strategies. For example, Facebook and Twitter recently announced permanent remote-work policies for their employees.


The market dominance of git as a version control tool, and of issues as a way to modularise work and manage tech-debt in public codebases have proved that decentralised work is also possible, at least in the software development space. This has contributed to the success of cloud code-hosting platforms like Github, Gitlab, and Bitbucket, which are gradually integrating larger parts of the DevOps delivery pipelines into their products.


The success of open-source has also neutralised skeptics, or at least forced them to apologise and adapt. Geopolitically speaking, open-source has become the new battle-ground where companies signal influence, creativity, and innovation. Facebook, for example, maintains PyTorch and React, both leading projects in their categories. Software is no longer an asset that has value in its own, but as vehicle used to deliver services and products.

The world is changing. The present looks exciting and the future optimistic, but a number of problems have also become evident:

  • Lower barriers of entry to the dev profession have created a noisy supply side with respect to skillsets and experience. Coding education has been standardised, yet writing software remains a highly artistic, textured, and creative feat that requires time, experience, and exposure to be mastered. Talent sourcing and assessment processes need to adapt to this new complexity.
  • Remote work has started to be accepted in the private sector, but the public sector still needs to catch up. Immigration policies and taxing laws impose severe regulatory frictions in the processes of hiring foreign talent. Many startups can’t afford this type of operational overhead. Competition in a global market requires access to the global workforce.
  • Cloud code-hosting providers, especially Github, are gradually controlling more sectors of the ecosystem’s infrastructure. It recently announced the acquisition of NPM, the leading package manager for Javascript. Monopolies that lock-in code, devs, or workflows can cause irreparable damage in the ecosystem and offend open-source principles.
  • Companies are deploying considerable resources into growing their open-source agenda, but the governance of the projects they maintain is not always open. Controlling the development roadmap of a particular tool can lead to the acquisition of unique competitive advantages, often at the expense of the ecosystem.

How does this affect devs and companies? The playground is changing. At the time of writing this article, the world is going through a mass remote-work experiment that has so far validated its feasibility. Developers and companies need to adapt to a future where work is remote and decentralised, where education opportunities are abundant and open, and where experience can no longer be summarised by a CV or a LinkedIn profile.

Dev-Market Interfaces (DMIs)

Developers have eaten the world. Their work has not only become a main driver for economic expansion, but a key supporting vector for civilisation. Still, the labour markets are far from being efficient. Companies are fighting for talent in the same arena where devs are settling for positions that barely align with their ambitions and work-life balance requirements. In an era like this marginal improvements in talent allocation can unlock incredible amounts of economic value. This is not easy to accomplish. The market currently reduces software creators to a few sets of keywords — devops, python, frontend, full-stack, and vets them with theoretical CS tests and subjective interviews. However, the journey and experience of a developer is extremely textured and can’t be appropriately assessed with a LinkedIn profile or a coding test. With so many variables to control, it’s more important than ever to diverge from the establishment and empower companies & devs with new interfaces to the labour markets. Some that can abstract away the noise from the real world and the complexities companies have inherited from dinosaur-era HR practices.


At Quine we’re building technology, machine intelligence, and marketplaces to help programmers and companies interact with the labour markets in a more fluid and human-centric way. In other words, dev-market interfaces that create frictionless experiences between software creators and capitalism. We believe that with the right mix of data, technology, and mechanism-design we can port the remote & decentralised workflow that made open-source flourish to the private sector — all while empowering devs and the open-source ecosystem. We’re a company at the intersection of ML, dev-tools, and economics. We believe in open-source, but make no assumptions on its monetisation strategies. Instead, we treat it as a hugely underserved passion economy with misaligned incentives that is inadvertently guiding us towards the future-of-work.


Keen to hear more? Sign-up to our newsletter — we’d love to keep you updated. Else, shoot us an email at info [at] quine [dot] sh.



Artwork from
 asciiart.eu — Written with ❤️ by Quine

Subscribe to our newsletter