I moved the opposite direction. Few months back, I was absolutely sure that the days of me being a software engineer are numbered. The programming I used to love and fun was gone. But, as I got to work more with Agents, factory & the infrastructure surrounding it, I'm more optimistic about the kinds of problems we could tackle with this, that wasn't possible/cost-effective before. We've only barely scratched the surface with what AI could be leveraged in the real world.
> Agents, factory & the infrastructure surrounding it
Can you be more specific about what isn't just a growing pain/maturity problem that will eventually be baked into the default products once the unit costs come down?
I can one-shot full video production that takes almost an hour on claude right now and it'll run ffmpeg for a good chunk of that. I don't see how the delegation + infra isn't a solved problem a couple years from now.
Why do you think there will be a human in the loop on any of these tasks, and why do you think that you will be that human?
I've watched the robots become better/faster software developers than I ever was over the last year. I've got a lifetime of esoteric knowledge that's completely obsolete now. I can ask robots to do things, but so can any other random person.
What, exactly, do you think we are going to be prompting these things to do? And why would anybody pay us to do that, rather than just prompt them directly?
The difficulty with building software has mostly never been the code itself, but how to tease out what to actually build and how to connect system components together in a sane way that meets business requirements.
It is indeed neat that LLMs with fancy agent harnesses can write code very well at this point, but that doesn’t mean the average person can think through problems in the way that is required to build software that operates at meaningful scale with complex trade-offs and architecture decisions to be made. In my experience with others in other fields, that is not a common skill, and I really don’t see how an LLM is going to solve that.
For this reason, I do not see software engineering going anywhere, but I do believe it is possible/likely there will be negative changes in hiring and wages.
If your views changed that quickly, then I don't know what to say. Your priors definitely need adjusting, and perhaps consider the way you evaluate the future.
It was pretty clear even 2 years ago. Now it's happening right on schedule.
Interesting. Since GPT4 came out I strongly believed LLMs would replace human programming. By November 2025 when opus 4.5 came out I thought it was pretty established already and most people agreed with that premise.
I think the experience of the gpt4/gemini2/claude2-3 era of models was widely variable depending on the use case and how much reference material there was on the internet. I'm told it was very good at producing react code for example, but my experience was it often failed to handle Rust or Kotlin type checking and failed pretty often (in less immediately detectable ways) on PHP. So looking at the trajectory from the initial copilot (could auto compete fast inverse square root and method level problems that you could also just google), to GPT4 (which from my experience, still couldn't code) and hearing that the training data by that point was "most of the internet" it wasn't that clear to me that it would reach the point it has today.
Obviously the models themselves have trended bigger since which has helped and also just the tooling and harnesses around it and the models being trained for that use case has achieved a lot since that q4 2025 window which is about the first time it became functional for me.
The training data back then was sooooo bad. You'd open something at random in Common Crawl and it was truly unbelievable that models (like llama) were able to output coherent stuff. The architecture was truly miraculous if it could train well with terabytes of garbage.
There was so much low hanging fruit back then in terms of training.
I think most people failed at extrapolating, seeing the trend and imagining what is possible. Maybe they didn’t want to. This is still the issue today. Some developers were overconfident about their understanding of how AI works, without really understanding deep learning at all. They assumed the limitations and shortcomings at the time were definitive. Tooling and harnesses were not a big invention at all. They were obvious from day one but took time to build and turn them into somewhat mature products we have today.
I think humans typing code is basically dead. But I also think typing code was always a minority of the work. System design, requirements, the big picture, that’s still engineer led.
LLMs can write code. What they can’t do is build systems based on vague desires from business people. Because nobody can, you need to distill and refine nebulous ideas into technical requirements and design. And that’s what engineers do.
I also think this is inherently a human endeavor. It’s about relationships and understanding. LLMs can’t accomplish this because they don’t have the political power or the ability to push back. And we probably don’t want them to have that power.
You just contradicted yourself. Good developers know what impact they have that is not replaceable by AI. If you don't see that for yourself you're probably simply not a good engineer.
Out of every single reason given for why the bootcamp era was justified the only one I believe is that there was so much focus on it because bigger players wanted to flood the market with cheap supply and drive salaries downward.
This is not any different with offshoring before, and AI now.
When I filled this out myself in July, I was "somewhat unlikely"
If you ask me again now, only a few months later, I'm probably leaning towards "Somewhat likely".
Insane how fast this space is moving.
reply