Why AI Hasn’t Replaced Software Engineers, and Won’t
11th June 2026
There is great anxiety and uncertainty about AI replacing jobs. How can we move past vague warnings and bombastic predictions and bring data to bear on this question? One good way is to look at the profession where AI capabilities are furthest along and adoption has been exceptionally rapid: software engineering.
In this essay, we argue that there is enough evidence to reject the narrative that once AI capabilities reach a certain threshold, it will cause mass layoffs. Given that this is true even in a sector with very few regulatory barriers, most other professions are likely to be even more cushioned.
We also have a good understanding of why this is the case. We can think of many kinds of knowledge work, including software development, as a “decide-execute-deliver sandwich”. AI compresses the “execute” layer — the middle of the sandwich — but the other two layers resist automation in a way that will not be overcome by capability improvements alone.
We conclude on a note of cautious optimism about the future trajectory of demand for software engineering. This essay is the first in a series, and the next one will look at reasons why individual software engineers’ careers might be rocky even if overall demand is healthy. The series is based on the published literature in economics and software engineering, our own evaluations and observations of AI agents, and many software engineers’ reflection on the present and future of AI impacts on their profession, gleaned both from published writings and our interactions with the community.
June 11th, 2026 at 10:38
The first iteration of any new technology is always less-capable than future models. When they were sending text messages over leased AT&T lines in the first ‘internet’, it didn’t seem like it was a big deal, but you have to look at the technology and ask what is its potential. As if to reinforce that, I have heard from 2 camps on this, that are professional software engineers and development people (I am not, so I claim ignorance to these specifics). 1 comment is the AI software cannot create original thought, or code, and you have to structure the prompts and input very precisely to get anything good out of it. 2nd comment is it DOES write good code, it can smash out simple projects very very quickly (relative to humans), and puts together simple-solutions to more complex problems (e.g. structures the solution and code extremely efficiently).
My own personal interactions at the retail level are sub-optimal. Grok and its ilk lie all the time, make shit up they don’t actually know, and are really nothing more than a fancy google-search. However if you assume the summation/results are accurate, it does do a good job of sifting through in 30 seconds what would take me 30 minutes to do myself with a browser and a fast internet connection.
So back to my earlier statement, when this technology matures, will it be useful? Well, yeah, I mean, it already could be used to ‘save time’, and if given the right feedstock, produces acceptable results in its current form. When I try to imagine what future versions of this technology will look like….? Yeah, I see world-changing possibilities, too!