I'm surprised it's taken me this long to write my first opinion piece about AI and its impact on software development. I think I've held off because I feel like my opinion is just adding more noise to a topic that is getting beaten to death.
I write to better understand my own position and to be able to reflect on my own state of mind. It is obvious that we will all look back on this time as an important time in history. In the future, I want to be able to look back and see what my thoughts were in the middle of the transformation.
This is a huge topic and my views are actively evolving. I will likely spread across several different posts and perhaps by the end I'll be able to better synthesize it all into one coherent post.
It's hard to decide where to begin, but what feels like a good place is the sentiment brewing amongst software developers (and may adjacent roles).
I've followed Hacker News for more than 6 years now and have been watching in realtime as sentiment towards AI is evolving. Recently, I've also found myself poking around Reddit more and seeing more posts and comments in despair.
Here is a recent example of general sentiment:
Software Developers Are Not Okay
It identifies factors at play like an impending AI financial bubble, mental health/burnout in software employees, layoffs, and the future of software devlopment.
I potentially want to start with maybe a controversial opinion...Software developers have been spoiled for most of their existence.
Since the advent of Software as a Service (SaaS), software companies have been able to make huge profit margins off the work of a relatively small number of software developers.
Meanwhile, it was deemed difficult to learn programming which led to not very many people that were competent, and so companies fought over the best talent, further raising salaries and benefits to all software developers.
I graduated from a 4 year computer engineering degree (plus an additional year for internships) in 2015. This was right about the time that coding bootcamps started to take off.
I think the conversation needs to start here, because what we are seeing with "AI" (aka coding agents) had already started way before the rise of modern LLMs.
Coding bootcamps could churn out a web developer in 4 months that was often better than a university grad.
When you increase supply, demand naturally goes down. At the end of the day a software company works the same way now as it did back then. Some kind of request or "prompt" for what needs to get built (usually driven by customer demand) and then some associated costs (e.g. time and money) until that request has been fulfilled.
At the time, it was jarring that after just a few months someone could write web apps as good as I could and maybe even better.
Part of writing this is exploring my own thoughts. I think as I saw the above start happening I already started to lose some of my interest in programming. This naturally led to a shift towards more engineering management.
At the turn of 2026, it became clear that agents could handle writing the majority of code. At this time I was working as a solo developer on a web application.
From January 2026 until now, I covered every aspect of the business from marketing, support, sales, and development.
Traditionally software developers have been a cog in the product development machine of a larger organization. Prior to coding agents, writing the actual code was a genuine bottleneck. In business bottlenecks are important to identify and optimize around. This led to optimizing most things so work was neatly handed off to a developer and they were pretty much left alone until they completed the work. This work was generally broken down into 2 week sprints and like it or not that's how long the work usually took.
I'm constantly blown away by how much I can accomplish with just a single prompt. I spend more time fighting anxiety about "how big a task feels" than I do actually completing the task. I've spent a decade learning to estimate work based on my own typing and thinking speed. My estimate comes immediately, but it is usually dramatically incorrect.
I think you'd be hard pressed to find someone now that doesn't acknowledge that this is the way software will be written moving forwards.
This is a big shift. I'd be lying if I said it wasn't affecting me.
I think my biggest struggles are lack of flow and guilt.
I could enter a flow state planning out a feature for 30 - 90 minutes. Then I would have several flow states to look forward to for actually writing the code.
Now I might take a few minutes to neatly craft my prompt and another few minutes later the work is done. I saved anywhere from 30 minutes to several hours of time. From a business perspective it is obviously the write choice to do it this way.
But I lost the satisfaction of completing the task. Despite the prompt requiring a fair bit of knowledge and expertise, on the whole the experience feels mindless.
In some ways it always was mindless...
The guilt sets in while sitting waiting for a prompt to complete its work. There are some that are parallelizing their work and running multiple changes at a time, but generally speaking I don't feel I can properly work on more than one thing at a time. So I sit there and wait until my prompt is done.
Sometimes it's a minute, sometimes it's 5. Maybe 100 times a day I have to sit and wait.
Objectively the work is getting done faster. But if I take my eyes or mind off the task in those few minutes now all of a sudden I am the bottleneck.
So I'm constantly in a state of hurry up and wait.
This writing has turned much more into a stream of consciousness, which I think is helpful to me as I need to get all of my thoughts out to attempt to organize them later.
I think many software developers (myself included) are left wondering "do I want to be a part of this next phase of the evolution of software development".
Here's where things get tricky. Companies will need fewer developers. In my opinion, those willing to trim their numbers will move faster and be better off.
But firing people is hard and frowned upon.
What if slowly your job became less enjoyable though? You would start to think about whether you want to stay. At some point you decide to leave. This doesn't reflect poorly on the company and they don't need to pay severance.
I don't write this to say "shame on those companies". Yes, in many ways "fuck big tech". But this affects companies all across the spectrum. If there genuinely isn't work for someone to do, a company shouldn't be obligated to keep them.
I think the most natural place my brain goes to is "what is the solution?" I (and others) can write about how I'm feeling all I want, but ultimately it boils down to what's the point of writing about it. As an engineer, the point has always been to fix some kind of problem.
I originally kept thinking that software developers were first to the punch in terms of having how they work uprooted. But then I remember how adamant people were that Radiology was going to die because AI could read scans and spit out reports.
A question that keeps occupying space in my head is "what do I need in order to keep doing this profession?"
I still love software and building things. I think about it pretty near all day long.
The more I think and write about this, the more the same answers I've always thought come up. Answers I formed even before AI.
Everyone in the world should understand how programming works. Much like reading and writing.
I also believe that people should build software for themselves.
I think there's a world where coding agents allow us to untangle ourselves from big tech. Unfortunately, it's unlikely that this is the path taken by the general public.
I'm a bit stuck swimming in thoughts. Perhaps some of the general feelings and vibes right now can be summarized simply as not knowing what the future holds.
This post has been challenging to write. I'm self conscious of the chaotic ordering of thoughts and meandering. I don't feel I can write structurally, because I don't think I know what I'm trying to communicate.
As weird as it sounds, I think I'm mostly communicating with my future self. I need to see all of these thoughts written out to be able to tie them together more clearly.
This is the moment that I typically abort actually publishing, but I really want to allow myself to publish these messy thoughts.
How I will handle revising this into a neater draft...I dunno, but for now it is what it is.