Why Engineers Shouldn't Let AI Do All The Thinking

AI-powered tools like GitHub Copilot and Cursor are changing software engineering by providing code suggestions, debugging and even writing full sections of code automatically. These tools promise to reduce the workload on developers and provide significant gains in productivity. However, this convenience may come with a cost: Relying too much on automatically generated code can cause engineers to lose their problem-solving skills and creativity.

As AI continues to change the software industry, engineers should focus on cultivating and maintaining their problem-solving skills rather than solely relying on AI. While understanding how to use AI-powered tools effectively is important, the ability to solve complex challenges with creativity will continue setting engineers apart. By focusing on their problem-solving abilities, engineers can ensure they remain valuable and effective, even as AI tools evolve and automate more aspects of software development.

AI In Today's Software Development

Today's AI tools are great at automating repetitive tasks, such as writing boilerplate code and refactoring. By taking care of the routine tasks, AI allows engineers to concentrate on design, architecture and developing new features.

However, while AI can help with straightforward tasks, it usually lacks the ability to understand the detail and complexity that is often present in real-world software projects. AI tools work by recognizing patterns in massive datasets, which means they're great at suggesting solutions to common, recurring problems, but often fail when presented with context-dependent and unique challenges.

Having a tool that can always provide a solution without hesitation can create a false sense of security. Engineers may begin to trust that AI can handle even the most complex issues when, in reality, the proposed solutions can be flawed or completely wrong.

The key to using AI effectively is the ability to critically evaluate its outputs. However, this can sometimes take more time and effort than simply understanding the problem and coming up with the solution yourself. The real value of a software engineer is in this deep understanding and critical first-principles thinking, which AI cannot replace in its current form.

Using AI In Your Engineering Work

To successfully navigate AI-assisted software development, engineers must consider AI as a tool, not a replacement for their problem-solving skills. Here's how you can use AI effectively in your workflow while continuing to develop your core engineering skills:

  1. Use AI for automation: Let AI handle repetitive tasks like autocompleting code, formatting or generating boilerplate functions. This frees up your thinking to focus on more important and challenging tasks. But when it comes to architecting systems, making key design decisions or debugging complex issues, your problem-solving skills are irreplaceable.

  2. Cultivate problem-solving skills: The most valuable asset an engineer has is the ability to solve complex problems, and that skill needs constant work. Engineers should continue to think about problems deeply, asking why something works the way it does and whether there are better ways to solve it. AI tools can help, but they can't replace the critical thinking required to come up with the correct solution.

  3. Understand the technology you're using: Learning the core principles of AI, like how machine learning models are trained, how data is processed and how neural networks function, will help you become an engineer who can navigate in environments where building AI-powered solutions becomes increasingly common. Instead of relying on AI tools as a black box, understanding how they work will not only make you more effective today but also ensure you can develop and implement AI-based solutions yourself, which is becoming a key skill in the future.

An Example From Front-End Development

A useful analogy comes from front-end software development, where many developers rely on software frameworks like Next.js to streamline the development of web applications. While these frameworks are very powerful, they often abstract away core technologies like JavaScript and the browser. Developers who become too reliant on these frameworks sometimes lack the skills and understanding to debug or build custom features outside the scope of the framework.

This is similar to the risk of relying too much on AI tools in software engineering. While AI tools are helpful, they can cause engineers to lose their deep understanding of the underlying principles. Engineers should be cautious about becoming too dependent on any tool, whether it's a framework or an AI assistant. Maintaining strong foundational knowledge is essential for solving complex problems that can't be handled by automation or frameworks alone.

Being A Creative Problem Solver

AI tools like GitHub Copilot and Cursor can be powerful additions to an engineer's workflow. However, engineers should use these tools alongside their problem-solving skills, not replace them. The future of software development will rely on both automation and critical thinking. Engineers who understand the principles behind AI and keep developing their problem-solving skills will be the ones who succeed in this evolving field.