The End of Software Devs? There’s a fierce debate in the world of software development right now; Will AI render the job of software developer obsolete ? Most of us software devs think our jobs won’t go anywhere, at least not anytime soon. But I think things will change drastically, very soon.

While no one knows what the future holds, we can leverage the information we have at our disposals, look at the pace at whic AI evolves, have conversations with as many devs as possible and then make an educated guess at what the future holds for us.

The profession of software engineering is going to evolve

Soon, software devs will do work in a very different way compared to today, and the role of the Product Owner will disappear completely! Let me explain.

This transformation will happen in two stages (I would love to be wrong, for my predictions to be false, and for our job as developers to remain the same, but we must face the facts).

Step 1:

With the advent of AI, we are seeing more and more developers adopting the strategy of Chat Oriented Programming (CHOP, A termed defined in this awesome article). Coding, stopping, asking questions to an AI, reviewing the AI-generated code, and integrating that code into their codebase.

This process happens in iterations, where the developer provides more context to the AI, refines their prompts, etc. And with the evolution of AI models, more and more code will be written this way, and senior developers will spend more time doing “Prompt and Review.”

Thus, developers will mainly become code reviewers or babysitters for LLMs. However, I don’t think the job of the best developers of the future will be limited to just that (See Step 2).

Before moving to the second step, it is important to note that: To use CHOP effectively, one must:

  • Be capable of knowing what they want the LLM to do, have an idea of how they would have done it manually, and provide enough context while asking it to do something as specific as possible.
  • Be competent enough to understand the final result and detect hallucinations and anomalies.
    • Why? Because LLMs hallucinate, and although their results are often convincing, even if the code runs without errors, if results are taken as-is without supervision, they can cause errors in business logic management and lead to significant damage.

I believe this will lead to two major shifts in the tech world:

  • Developers who do not use CHOP will fall behind in terms of productivity.
  • Developers who use CHOP but lack the skills to understand and modify the LLM-generated code when needed, or who cannot transform business needs into code and then into an effective prompt, will become “useless.”

Step 2:

Considering Step 1 above, we will see that the rockstar developers of tomorrow will be orchestrators who know how to review and direct the work of LLMs in a precise direction. They will ensure that the AI does what the end clients need:

  • Write high-quality code.
    • This is necessary to babysit increasingly powerful LLMs, as mentioned in Step 1.
  • Understand business needs and rules.
    • A customer-focused engineer who knows how to communicate with users, understand their needs, collaborate with designers, create useful solutions for users, and evolve the product accordingly.
  • And to take it even further, with AI agents emerging, they will have to break down projects into well-defined tasks and instructions that AIs will program individually.

This may seem like a lot for one person, but remember, the assistance of AIs will give this engineer enough time to manage these tasks.

Considering that senior developers will be the ones who become customer-focused, this leaves little room for the Product Owner role. The Product Owner, who once saw themselves as above engineers, will no longer have a real place, especially if they never had a software engineering background.

How to adapt ?

The truth is that no one really knows the future, and this automatically means that no-one knows the right way to prepare for it. But, in my opinion, at least one should do these in other to be safe.

  • Step out of your bubble and stay updated on what is happening in the AI world.
  • Try to be more customer-focused as a developer. Do everything to gain the skill of understanding needs and turning them into solutions.
    • Today, developers are very specialized, but in theory, the big winners of the future will be generalists.
  • Improve your development skills to be able to understand and modify LLM-generated code.
  • If you are a senior developer, use LLMs for development and don’t turn a blind eye.
    • However, if you are a junior developer, avoid LLMs and focus on rapid skill development. Master technical challenges so you don’t get left behind by what’s coming.

Conclusion

I would love for the development profession as we know it to remain unchanged, but we must face the facts! AI is shaking everything up. Even the data science sector is being accelerated by its own innovation. The best thing to do is to prepare and adapt to the change.

Follow me on social media and stay updated

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.