Minute
Butterfly

Four possible futures for independent developers
Nov 10, 2026 - Version française

The arrival of generative AI in software development calls into question our business models, our work practices and our very identities. Having spent a good part of my career as an independent free software developer, I have spent quite some time in recent months thinking about the future of my own practice. To be honest, though, I think the same conclusions apply to other intellectual professions and to other entrepreneurial models.

I start from the premise that the business models that have prevailed until now are no longer viable: it is simply no longer possible to charge €500-1000/day over several weeks or several months to create a codebase patiently written line by line by hand. Market expectations have changed. Starting from this premise, what strategies can independent developers adopt? I see four.

1. Switch to a practice of material production or transformation

The first option is to leave the world of software development and take up a practice of production or transformation in the tangible world. In other words: leave the tertiary sector for the primary or secondary sector, sectors on which generative AI is not (yet) able to have a direct impact. This option was chosen by the many developers who retrained as bakers, mechanics, cooks, etc.

2. Take up an artistic practice

As of today, the involvement of generative AI in the artistic field is completely intolerable to almost all audiences. We can see it with the Thélyson Orélien affair, but also with the rejection of AI-generated posters, music and films. This rejection is an opportunity for developers who want to continue their practice but refuse to rely on generative AI to create code or content. For example, video games are a sector with a very large overlap with programming.

3. Achieve productivity gains

This option consists in not radically changing customers, field of activity or even type of production, but in increasing the quantity of production at constant cost. In other words: increasing productivity. It can be summed up as "producing more code in less time". This is clearly the approach adopted by most software companies, and it leads to the proletarianization of developers.

4. Move up in complexity

The last approach consists in shifting developers' work towards tasks of higher complexity, for which there are still few people available to supervise the work of generative AI. If we take the example of Web application development (frontend, backend or devops), there are today enough well-trained developers to supervise the generative AIs to which most of the work will be entrusted. But there are only a few expert engineers able to understand what a generative AI is doing once the complexity goes up. For example: deploying innovative search engines, creating alternative orchestration platforms, building "offline-first" applications, applying recent results in mathematics, etc. This period is favourable to engineers who can train themselves, since they can rely on generative AI both to explore new fields of expertise and to create new tools or integrate them into their practice.

Historical precedents

This is not the first time that a technological revolution has transformed an entire sector of the economy, and the four directions proposed here have already come up in past revolutions. For example, the quartz crisis, which hit the watchmaking industry in the 1970s-80s, led watchmaking groups to choose different survival strategies. The Straumann group moved away from watchmaking and used its expertise in alloy production to make components for dental surgery. Rolex survived the crisis by focusing on the luxury sector and creating watches with a strong identity. Swatch was created during the crisis to offer low-cost industrial watch production. The Concord company, together with three other companies, used the technology brought by quartz to design the Delirium (French), a technological feat and, at the time, the thinnest watch in the world.

There are many other historical examples of technological revolutions we can learn from: the revolt of the Canuts, whose trade was upended by the arrival of the Jacquard loom; the trade of typesetter facing the arrival of the Linotype; and more recently the trade of translator facing the arrival of machine translation.

Collective alternatives

The paths described above for surviving the transformation of the sector work at an individual level, but there are other paths that could help us get through this crisis through a collective effort. For example, after a prolonged strike, the Hollywood screenwriters' union reached an industry-wide agreement aimed at protecting their jobs from generative AI. One could imagine similar agreements being made to protect developers' jobs. To my knowledge, no major industry player has undertaken such an initiative so far.