Sometimes we joke that we’ll soon need some form of craftsmanship certification for manually coded systems. And perhaps there is actually a grain of truth to that. In a short amount of time, AI has become a natural part of daily life for many developers and has fundamentally changed how digital product development works.
At Plingot, we clearly see how AI for programming is changing both our pace and our workflows. The question is no longer if AI should be used in software development – but how
Will AI Take Over Developers' Jobs?
This is likely the most common question today, both inside and outside the tech industry. Today, AI can write code, debug, plan features, and structure projects faster than most people thought possible just a few years ago. However, generating code is not the same as replacing a developer or a software architect.
"Today’s AI cannot replace a developer or an architect. It’s simply not at that level yet," says Mio Nilsson, CEO and Developer at Plingot. "With the right guidance, you can get excellent results. But you can't just ask it to build anything and expect it to work flawlessly forever."
Instead, it is the role itself that is changing. AI is excellent at rapidly generating solutions and writing large volumes of code simultaneously. At the same time, this places a much higher demand on quality control, structure, and oversight.
"AI makes mistakes, gets sloppy, and takes shortcuts," Mio explains. "The final product might look perfect on the surface. But when you look closer, there can be structural flaws that make future updates difficult. Sometimes, security vulnerabilities are introduced, making it easier for systems to be breached."
Because of this, human expertise remains absolutely vital in areas such as:
- Architecture and System Overview: Understanding how everything connects.
- Security and Code Reviews: Ensuring data protection and reliable performance.
- Prioritization and Business Logic: Aligning tech solutions with business goals.
- Long-Term Maintenance and Scalability: Keeping the system healthy over time.
AI Doesn’t Make Development Simpler – Just Faster
In many ways, AI has turned the traditional development process upside down. It’s no longer just about writing code manually; it’s about directing, reviewing, and collaborating with AI tools effectively.
"It has reshaped our entire development process, for better or worse," says Mio. "You act more like a conductor, pointing clearly to what needs to be done, and testing everything thoroughly in a way we never had to before. You end up building tools just to manage the process and maintain the oversight you need."
At the same time, there are areas where AI is already incredibly efficient. Emil Johansson at Plingot highlights troubleshooting as a major time-saver.
"It is especially fast at debugging when you need to scan through multiple files and logs to find an issue. It often pinpoints the root cause very quickly."
Today, AI is used for everything from research and prototyping to code generation and feature planning. Tools like Codex and OpenCode have quickly become standard parts of many developers' workflows.
"Codex and OpenCode help me write code, troubleshoot... pretty much everything," Emil says.
However, faster does not always mean easier. Many companies seem to think that AI automatically simplifies development, but in reality, it can be the opposite. When the pace increases, bad decisions become more expensive. Technical debt can build up faster, and incorrect solutions can have costly consequences down the line.
Where AI Still Falls Short
Despite rapid advancements, AI still has clear limitations in software development. A common challenge is that AI can write code that looks correct but isn't sustainable in the long run.
"The code quality can be questionable in certain situations, but generally, it’s quite good. However, AI can easily miss assumptions about how the program functions. You have to be extremely clear when giving instructions and review the generated code extra carefully," says Emil.
AI also tends to make solutions more complicated than necessary.
"Often, you notice that the code is a bit more complex than it needs to be," Emil notes. "It’s similar to how you spot AI-generated text—it might be grammatically correct with the right syntax, but you think, 'a human probably wouldn't have written it this way.'"
In some situations, an experienced developer is still much faster than AI. A quick fix that a human can spot instantly might take AI much longer to analyze, simply because it needs to process large amounts of code and context first.
Additionally, AI is not free. The more advanced the tools and workflows, the higher the costs. For many businesses, the challenge isn't just adopting AI, but doing so in a sustainable, well-thought-out way.
How We Work with AI at Plingot
At Plingot, we use AI as a tool to enhance our developers, not to replace them. How much we use AI varies between projects and team members, but today it is a natural part of our daily routine.
AI helps us with code generation, debugging, research, prototyping, and feature planning. At the same time, we place a heavy emphasis on areas that still require human oversight: quality, security, structure, and long-term sustainability.
"Every stage is affected. AI can assist with practically every step of product development. It’s all about how you apply it and what data you give it access to," says Mio Nilsson.
There is no perfect handbook for working with AI in development today. The tools change rapidly, and new workflows emerge constantly. For us, it’s about finding a balance where AI helps us move faster and more efficiently – without ever compromising on quality.
Mio Nilsson, SEO and developer, at Plingot
The Future of Software Development is Collaborative
AI will continue to reshape the role of the developer at a rapid pace. Automated workflows, AI agents, and smarter tools will likely become standard in the future of software development.
"We will see more AI agents working in the background to solve tasks," Emil predicts. "As a developer, you might not need to worry as much about the minor details anymore—at least most of the time. But you still need the expertise to spot where the AI goes wrong."
At the same time, Mio believes that understanding the underlying technology will become even more critical.
"You are in deep water if you ask AI to build things you couldn't build yourself. Stick to a tech stack you are comfortable with—one where you can critically evaluate the results and provide feedback."
At Plingot, we view AI as a powerful co-pilot in the development process, not a replacement for human talent. Building sustainable digital products still requires experience, problem-solving, responsibility, and a view of the bigger picture.
AI is changing how we build. But humans are still the ones who need to steer the ship.
