Will AI Replace Software Developers?

First of all: this is just a humble, personal opinion. I'm not an expert in generative AI or economics, but I find this debate incredibly interesting — and, to be honest, it's also fun to make predictions to look back on a few years from now 🙂

Right now, as usual, the conversation seems pretty polarized; here are some statements I've read recently:

“Thanks to AI, engineers will be freed from the boring parts and focus on creative work.” “Replacing engineers is impossible because AI only predicts — it doesn't really build.” “Within two years, 90% of engineers will be redundant.” “AI will give birth to new business models we can't even imagine yet.”

With all due respect, I think most of these views are exaggerated and fragile, even though each of them has a grain of truth. As with most things, the truth probably sits somewhere in the boring middle. Here's my take:

The argument that AI will free us from tedious work so we can focus on creativity sounds a bit naive to me. If AI makes an engineer 50% more productive, companies will probably need 50% fewer engineers to deliver the same business value. It's sad to say, but I doubt most companies will pay us just so we can have more fun at work. The real question is whether investment in digitalization will stay stable — if it does, we can do more with the same resources. Maybe this is my bias, but I still see many unexplored opportunities to create value with technology. I'll come back to this point in more depth shortly.

Dismissing AI as just a prediction system doesn't feel right either. That's an oversimplification. After all, isn't our brain essentially “predicting” the best lines of code to write based on our knowledge, too? The line between creative reasoning and merely “spitting out tokens” is pretty thin.

“If 90% of code is generated by AI, 90% of developers won't be needed” — this is another fallacy. Most code being written with AI assistance doesn't mean there won't need to be a human behind it. Taken to the extreme: 100% of code is written with keyboards… and yet we're still here. Even so, I believe today's AI paradigm is inherently limited when it comes to fully replacing a human end to end. The lack of determinism (for now) prevents full automation from scaling. Of course, that doesn't mean the productivity gains will be insignificant. We'll see what happens with new paradigms (magic word: agents) — it looks like just a matter of time.

“AI will bring new business models we can't even imagine” — maybe, but this sounds a bit like wishful thinking to me. It's true that this often happened in past technological transformations; however, there are also economic models that question the idea of infinite growth. Although past innovations have often brought economic expansion and job creation, it's unclear where the limit of all this lies. Personally, I see saturation even in the entertainment sector.

Frankly, all of these narratives are shaped by actors with their own interests — especially the overly optimistic ones, which often come from companies that directly benefit from the hype around AI's potential. Needless to say: we should approach these messages with caution.

I have somewhat mixed feelings myself. On the one hand, I recently tried some AI coding tools and was impressed — this will clearly and permanently change the way we build software. On the other hand, I can't shake a feeling most of us in the industry know very well: the last 10% of the work takes 90% of the time. It seems we're almost at the point of having AI build software from natural language prompts alone… but maybe that final stretch will be the part that takes the longest. Think of autonomous vehicles — people have been saying “almost there” for the last five years.

As for the future of jobs in software development — I think nobody really knows. With my limited knowledge of economics, I believe macro-level trends (consumer demand, international conflicts, pandemics, geopolitical instability…) are inherently unpredictable and lie beyond the control of 98% of the population.

So, going back to point (1): yes, we will probably need fewer engineers to build the same capabilities. But will total investment in technology stay the same? If it does, maybe we'll invest in more ideas with the same number of people. There are many possible scenarios:

Microsoft Copilot Health Unveiled: Your New AI-Powered Health Assistant

Microsoft's Copilot Health analyzes medical records, health history and wearable data on one platform so patients arrive better prepared for appointments.

Strategic AI Agreement Between Mistral AI and the French Military

French AI startup Mistral AI signed a strategic AI agreement with France's Ministry of the Armed Forces centered on data security and technological sovereignty.

There's the Jevons Paradox, which everyone has been referencing lately. According to it, when something becomes more efficient (in this case, software development), its use increases rather than decreases. Until recently, digitalization often meant adopting SaaS. But if building gets easier, companies may find it more valuable to develop their own customized capabilities. …

In short:

Yes, I believe AI will eventually replace some software developers — just as every technological evolution replaces certain kinds of jobs. The real question is when that will happen.

In the short and medium term, I think the impact will be gradual and far from dramatic. As with most big changes, it will seep into our lives slowly and almost imperceptibly. As I mentioned before: especially in a deterministic field, the last 10% of the work takes 90% of the time. So let's stay calm and not give in to panic or anxiety. There are plenty of incentives pushing us toward FOMO and haste.

In the medium to long term, it will depend on broader and more controversial questions — such as whether capitalism can sustain continuous growth, or whether we'll reach a level of productivity and automation at which we may need to rethink everything (universal basic income, radically shorter work weeks, etc.). If that happens, every cognitive job could become replaceable.

But what is certain is that AI will keep improving exponentially and will definitely transform software development roles. We should accept this reality — not ignore it — but calmly and thoughtfully.

And the million-dollar question: would I recommend that my children study software development if they showed an interest?

Absolutely, yes. Building things is an enjoyable experience in its own right. Even if we don't know what developing software will look like in the future, the process itself is deeply satisfying. And I think understanding the fundamentals is essential to getting the most out of abstractions.

Extremely interesting times lie ahead of us; for now, let's just enjoy writing code!

Thanks for reading — especially if you made it to the end of all my rambling. The opinions are entirely my own 😊