AI - The end to software engineering?

AI - The end to software engineering?

With the recent explosion of coding AI tools, the question of how this will affect the role of a software developer is being raised all throughout the industry.

A recent article argued for the continuing need for experienced developers in the future. The reasoning being that solely relying on AI to write an application is fraught with danger.

After all, AI is only as good as the information being used to train it and people behind it. However, this article was missing one all important element, the human factor.

AI is a tool, right now it’s allowing developers to speed up delivery. But there’s the trap, speed to delivery. For this reason alone, the future of software developers is numbered.

Businesses want results and they needed them yesterday. Non tech business owners will all be aware of AI and how it can improve their business.

What they will also be aware of is that you as a software developer should now be able to complete tasks in a much faster time frame than before. In a business, time is money.

Most software teams are under extreme pressure to deliver projects on time and within budget. As a software developer you will be under pressure to deliver to tighter timescales. AI will allow you to do this and as a result software teams will have to rely more and more on the code generated by AI. When under pressure the temptation to let AI write the code, despite the potential risks, will win through. This is the human factor.

Let’s not forget that the code now being produced by these AI models is already pretty impressive. In time it will only get better. In fact we could argue that today the code the AI models can generate is better than most developers can create.

Only the top few highly skilled developers will be able to match it. Eventually even they will be surpassed.

As a lead developer you will want to take time to review the code and ensure there are no bugs or security flaws, make the code readable and follow core coding principles.

That is a great value to have and as a software developer you know this to be true, but unfortunately clean code doesn't pay the bills.

As AI models improve, more and more developers will take the easy way out. The tools to test and review code will improve and the code being produced will have far less bugs.

As much as we would like to fight this, the pressure cauldron of fast paced delivery will be the end of software developers. Within the next ten years the skills needed to create software will not be about how good you are at coding, it will be about how well you use the AI systems. How well you craft a prompt to meet requirements.

Understand this from a business owner's perspective. In the past you had to pay top dollar for software engineers because they were highly skilled. However no matter how skilled no system was ever built bug free and the cost of maintenance high. Now those skills are being diluted.

You won’t need to understand code to deliver software, you can do that today already. So why would I pay top dollar for a software engineer when I can employ someone for a lot less who can build a system faster and cheaper using AI.

As a result salaries will drop and learning to code will not be incentivised. The only software developers remaining will be the elite, those working for the AI giants and tech companies.

Businesses will no longer need software developers to build their systems.

In conclusion, the human factor of taking the easy way out, the pressure of software delivery timescales, cost of delivery and the speed at which AI quality is improving daily. The future of coding in its current guise will become a thing of the past. It will be replaced by people who can use the AI tools. These AI Dialogue Architects will be the next generation of software developers.