In the world of AI, companies will need MORE developers, not fewer

The market for developers seems tough right now, and many associate it with the growth of AI tools. However, making such an association is very naive. Let's analyze it and make some predictions.

Why is the developer market tough right now?

Before AI stormed media, I could see a very unhealthy situation in the developer market. Big tech companies hired absurd number of developers, while many domains with low-hanging fruits couldn't hire developers to make products that could actually benefit people. In consequence, developers were often fighting with each other (not actual problems) and feeling they do a nonsense job. Many were leaving to make or work in startups, but one of the big reasons why startups fail is the cost of development.

Let's support those theses with already famous example. Twitter hired roughly 1,500 developers before Musk took it. Main Twitter functionalities are quite simple, and on many universities it is a project to make simplified Twitter clone. Yes, making it for such a scale it much harder, but I firmly believe that a team of good engineers could make it as fast and as stable in a couple of years. So what are those 1,500 developers doing? I believe many were working on features with minor business value. Their development was slow and hard because of the amount of code, practices, meetings, and of course, other developers. The theory I present now is something I could hear from many people, but it was just a theory, lying in the area of conspiracy theories. However, once Must take Twitter, fired 2/3rd of developers, and started finally introducing new features to the product, this theory received a strong validation.

To be clear, a Twitter case isn't a proof, but I believe it made many CEOs think that maybe they need fewer, not more developers. So even if Twitter case leads to wrong conclusions, it already led to different decisions at companies. Additionally, the market is supporting such decisions. The situation in the world is though, and companies look for ways to save some money. I could first observe limiting marketing budgets, travel budgets, training budgets, and not it seens in many companies development budgets are limited. In the end, most developers do not work on maintenance, most developers work on new features, and if a company wants to limit costs, limiting new features development is the easiest way. I believe that those are the key reasons why many companies are freezing hiring or even firing developers. AI is just an excuse, but it might actually help us, developers, but that would require our flexibility and openness to change.

What is AI from a developer perspective?

From a developer perspective, AI is just another tool we use to make our job more efficient. It is one of many tools that emerged in my career, and I don't think it is the most important one. How much more efficient AI tools make us? There are different estimates, some research showed around 30% efficiency improvement, others claimed that improvement is insignificant if we enforce high quality. I believe that those research highly undermine AI tools. I could see how much faster development can be if you know how to use AI agents. You need to write a good specification document, prompt, review what they changed. Done well, I believe AI can make development between 100% to 300% faster, depending on product complexity. I think it might be even better, as those tools and our practices will get better and better, I believe that the claim that development will be 10x faster is achievable. That is a lot, but that is not the first tool offering such a gain. Development started from writing processor-specific commands in assemblers. Comparing to that, high-level languages like Java, with its garbage collectors, optimizers, high-level libraries for everything, are at least 100x improvement. At first, we were developing in text editors. Now our rich IDEs offer us coloring, navigation, type-analysis, refactorings, and many other tools that make us more efficient. Could you imagine working in Notepad instead of IntelliJ? I believe I would be much more than 10x less efficient. We also have other tools, like Git for version control, modern build tools, CI/CD for automating deployments and tests, cloud platforms for easy hosting. Honestly, it seems absolutely impossible to develop anything close to modern applications in assembler (for each device architecture),Notepad, flappy disc to exchange code, and manual deployments.

Do you think that compilers and IDEs made developers redundant and made us lose our jobs? Of course not. On the contrary, they made development more efficient, so they made it more reasonable to hire developers to develop systems instead of using other kinds of specialists to do this work manually. The same with AI, if used well, it will make us (developers) more efficient (used poorly might lead to so many mistakes, it might undermine trust for digital products). It might introduce us to new lands, where our work is actually needed.

The new lands

When I look around, I can see so many areas that are just waiting for developers to improve them. Great, you can buy most products on Amazon, but why do most people need to go to a supermarket to make shopping? I use an app for that, Frisco, but it is quite expensive as is the only well-implemented online supermarket I found (good app and procedures). Why aren't all supermarkets offering online shopping with home delivery? Even better, why can't I buy vegetables in an app, that would come from a farmer living not far from me? Think of it: an app where farmers and factories offer their products, and people order them to their houses. Farmers would have 50% of the final price of their product, not 1%, and we would have local fresh products delivered to our houses. I know it would require crazy logistics, and making such a product would be extremely hard, but this is just an example. In Poland, we can find good private doctors using Znany Lekarz app (known in other countries as DocPlanner or Dottoralia), but why is there is no such a product for finding a mechanic? Why in restaurants I can still see people using unintuitive 90'-style systems instead of modern applications? Why still so many things require going to a physical office instead of completing them online?

For now, I just named low-handing fruits, but there is so much more. So many processes wait to be automatized. So many factories could need only developers to set up processes and technicians to maintain machines. So much work is now done manually, while it could be done by robot arms or drones. We have technology, tools for that are not cheap, but in most domains they would quickly pay back. As it turns out, often the biggest problem is cost and a lack of specialists that could implement that. Us, developers. Yes, most of us do not know now how to program robots or drones, but we could learn that. Thanks to AI, maybe more and more of us will, since entering a new area is simpler when you have AI assistants helping you, teaching you, and making your work more efficient.

Learning AI tools

Treating AI like it is an enemy is not a good strategy. It is a new tool to learn. If I feared losing my job, I would learn AI tools and show even better performance. If I lost my job, I would use it even more. I would use it to make a startup and make some extra money, which would help me learn how to use those tools really well and give me another good position in my CV. If I couldn't find a job, I would look for ways to make those skills profitable. I am sure soon using AI will be among skills tested in job interviews. Learning AI tools are in the center of every case. It would be strange if it wasn't. It is a new tool that can improve our performance, so our job is to learn to use it well. We are developers, we learn new tools! This is why I believe we are actually in a great situation. We are trained in learning and using new tools, what is more, AI is made to work with what we know already (APIs, libraries, MCP servers). We (developers) are at the best starting point in the run for the new best positions in the world of AI. This is an opportunity, use it!

Does it mean there is nothing to worry about?

If your conclusion from this article is that your position is safe, you completely missed the point. In the future there’ll be no need for people who can only search on Stack Overflow and blindly copy-paste solutions. AI can do that. Developers of the future are specialists who know:

1. How to use, communicate with, and shape projects for AI integration.
2. Understand tools and technologies they use, so they can properly correct and properly lead AI.
3. Can communicate clearly and understand business requirements to ensure AI solutions meet organizational needs.

AI rewards experts much more than it rewards beginners. In the future, the importance of education and expertise will only increase.