↤ September 04th, 2026
For a research project, I have been asked to comment on the “AI revolution” in coding. The research has since shifted its focus to understanding the compounding crisis that programmers are currently facing. On the one hand, I hope my response has helped with that shift, on the other hand, what I wrote is no longer required for the project, so I am publishing it here.
After the fast-moving (agentic) AI evolution is completed, what role will software developers have in society?
As an outspoken opponent of the current cycle of LLM-based tooling in all areas of society (including coding) on ethical, legal, financial, ecological and all around humanistic grounds, I have a grim outlook on what the so-called “agentic” future looks like.
These systems are built on exploitation of the commons that the FOSS community spent over 40 years to put together carefully (if not always perfectly). The predatory nature puts a few US-megacorps into a position of power that we were just starting to divest ourselves from.
There is of course a hypothetical LLM future where models are run on green energy, honour FOSS licenses, and contribute back to society, but I don’t see anyone funding this, so this is, in my opinion, an evolutionary dead-end.
But say this comes to pass against all odds, we know from copious studies and forced rollouts in both large and small companies that the suggested gains simply do not materialise and extended use of such tooling has detrimental effects on the mental health and capabilities of the practitioner. Fred Brooks will prevail, there are no silver bullets1.
This lengthy preamble aside, the role of a software developer in society will continue to be understanding both computers and the requirements of the people that use the software they are developing and building solutions that make both sides happy.
Understanding computers is not going away. After all, even if I am wrong about everything in the above, who’s going to build the next generation of these tools.