Is coding dead? There's been a lot of talk about AI killing software engineering, and SWEs rightfully lamenting the death of their favorite activity.
Is AI good enough at coding? At the same time, I've also seen respected, senior engineers deride AI because it didn't yield good results when they tried it. I want to share an alternative take that answers both these questions.
Yes, Coding is Dead
Let's get this out the way. Coding is dead, assuming you can build Sufficiently Smart agents.
"This has been said a thousand times before, but allow me to add my own voice: the era of humans writing code is over. Disturbing for those of us who identify as SWEs, but no less true. That's not to say SWEs don't have work to do, but writing syntax directly is not it." Ryan Dahl - https://x.com/rough__sea/status/2013280952370573666
But the real question is not whether coding is dead, it's whether you're skilled enough to context engineer agents that write high quality code. If you are unable to, then you will see little value when using AI to ship real code in production.
Models have advanced enough that the bottleneck is no longer the model itself. It's you, and everything around it that you provide: rules, skills, context, internal search tools, the way you prompt, and so much more. Given the right context, agents can write better code than I can, faster than I can. I've seen first hand, one of the best and most skilled engineers I know, teach Claude how to understand a deeply complex and difficult project, and use it to solve a bug that they've been stuck on for more than a year.
The next unlock has become whether or not you can build a harness around Claude Code or Codex to make it truly work for you. Coding may be dead, but engineering is needed more than ever. Rather than see agents as our replacements, I think of them as a way to make cybernetically enhanced clones of myself and my colleagues. We can give these clones rules, skills, and context based on the knowledge and scars we've gained over the years designing systems and handwriting code.
In 2026, not getting value out of AI is ackshually a skill issue. And you wouldn't be the first. All engineers (including me) who started their careers before AI must learn new skills in order to stay relevant. These are new skills we'll be learning over time.
I'm not going to say that you're going to be left behind, but like any new paradigm, this requires changing mental models and perspectives. Skills can be learned with time (and they will keep evolving), but changing your perspective is the first and most important step.
The Factory Must Grow
I think Factorio is a close analogy describing what's happening with coding.
Up until very recently, SWEs wrote code by hand, painstakingly sweating the details and logic in order to build software. We were mining iron, copper, and stone with pickaxes. We were the artisans putting together fantastically complicated machines that now run everything.
In this analogy, AI heralds the invention of automation. AI labs invented a magical machine that can automate things that weren't automatable before. Instead of mining with pickaxes, we can now build an agent rig that mines for us. Instead of putting together machines by hand, we can build machines that make the machines.
Long Live Engineering
Even though Factorio is a game where the goal is to automate production so you don't have to do anything yourself, you still end up continuously refactoring, extending, and improving systems so they run smoothly and scale easily. It's not fun if you just sit there and watch while the game plays itself. But nobody does that. Instead of manually mining ore and putting it into a furnace, you build drones and train tracks, and send rockets to different planets to setup more robots to do stuff for you.
You start to think about the higher level problems, because you're not stuck spending all your time on implementation details. And if you mismanage things, you'll end up with an inefficient factory that continuously breaks down and requires intervention.
This is what SWE has become. The only question is whether you will help the factory grow, or hang on to your pickaxe.
--
thanks @gweterings for inspiring me with the Factorio comparison
