Not long ago, I wrote about how I changed my mind on AI in software development. Since then, I've gone a step further - I've actually built and tested a workflow where GitHub issues turn into working pull requests through AI. And it works. It’s not perfect yet, but it's already effective.
Where We Are Now
Today, AI code generation is strong enough to handle small to medium-level tasks. Things like writing boilerplate code, setting up repetitive functions, or even completing entire features at a basic level. But when it comes to complex problems, creative design, or deep architectural decisions, human developers are still essential.
So right now, we’re in a hybrid world:
So right now, we’re in a hybrid world:
- AI handles the routine.
- Humans handle the complexity.
Where Things Are Going
What excites me most is what’s ahead. Models are getting better, cheaper, and faster. That means:
- More automation for routine coding.
- Faster turnaround from idea to working code.
- Lower costs for building and maintaining software.
I expect that, in the near future, AI will move beyond just "helping" and become a core part of how code gets written - especially for companies that care about speed.
Specialized Models
Another big shift I see coming: custom coding models per technology. Imagine a language model trained just for Node.js, or just for Laravel. Instead of being a general-purpose coder, it would deeply understand the best practices, patterns, and quirks of that specific ecosystem.
This could open up possibilities we can't fully imagine yet - from auto-generated code that’s indistinguishable from expert-level work, to project-specific models that "know" your codebase as well as your senior developers do.
This could open up possibilities we can't fully imagine yet - from auto-generated code that’s indistinguishable from expert-level work, to project-specific models that "know" your codebase as well as your senior developers do.
Will AI Replace Programmers?
Honestly, I don't know and nobody really knows (yet). It's probably a billion-dollar question.
What's clear is that the environment has already changed. Senior developers are now expected to boost their capabilities with AI. In a sense, we may be seeing the rise of real "10x engineers", not because of natural genius, but because they know how to leverage these tools.
At the same time, junior and entry-level developers are struggling to find jobs. This isn't only because of AI - market conditions and economics also play a role - but AI code generation is definitely a heavyweight factor.
I don't think we're anywhere near AGI, but what's happening feels like an industrial revolution for software development. The boring, manual parts of coding are disappearing fast. And here's the bottom line:
What's clear is that the environment has already changed. Senior developers are now expected to boost their capabilities with AI. In a sense, we may be seeing the rise of real "10x engineers", not because of natural genius, but because they know how to leverage these tools.
At the same time, junior and entry-level developers are struggling to find jobs. This isn't only because of AI - market conditions and economics also play a role - but AI code generation is definitely a heavyweight factor.
I don't think we're anywhere near AGI, but what's happening feels like an industrial revolution for software development. The boring, manual parts of coding are disappearing fast. And here's the bottom line:
If you're not using these tools, you're already underperforming.
Otar