To Think or Not to Think

All my early exams were in-person, requiring me to use my domain knowledge to solve problems within a well-defined framework. There wasn’t much outside the taught syllabus.

During my master’s, I encountered complex take-home assignments that required researching online for similar problems. From those resources and my domain knowledge, I constructed answers. While I offloaded some of the thinking to existing frameworks, I still felt enough resistance to foster learning.

When I started coding before ChatGPT, the process felt similar—I would search the web for existing solutions and then apply my domain knowledge to tailor them to my needs.

Likewise, in my professional life, engineering solutions often involved modeling equipment using in-house tools that were still relatively primitive. Understanding the principles behind these tools was essential to using them effectively.

The Shift with AI and Advanced Tools

With the advent of ChatGPT and more sophisticated modeling software, the need for deep, manual problem-solving has decreased. While we still need to correct, validate, and test AI-generated code, much of the tedious, boilerplate work is handled automatically. We simply provide inputs and receive structured outputs.

Similarly, modern equipment-sizing tools require only input parameters, and the software determines the appropriate specifications. The process has become more streamlined—feed in the right data, and the tool does the rest.

What is “Under-the-Hood” and how does it Grow?

It represents everything that countless researchers, engineers, and thinkers have spent years developing. These individuals were like explorers navigating uncharted territory, pushing the boundaries of knowledge so others could build upon their work. Taking it further in the next stage.

It’s the thoughts, insights, and relentless pursuit of knowledge by those who challenge the unknown. Their thinking and discoveries fuel the expansion of our collective understanding. This collective thinking is then used to build models so the next generation can use them faster.

What’s the Concern when People Think Less?

This shift isn't inherently problematic. Experienced coders and engineers still understand what happens under-the-hood—the intricate logic and principles that power these tools. However, as more processes get automated over the next 10–20 years, future engineers may have less exposure to the fundamentals. They might rely on these tools without fully understanding their inner workings.

If thinking declines, the pace of progress slows. While research will continue, a vast majority of people may disengage from deep problem-solving, relying entirely on prebuilt solutions.

Impact on Jobs and Solution

Early-career roles traditionally serve as a pipeline for developing future experts. When AI replaces these roles, companies lose not just junior engineers but also their future senior engineers, creating a long-term gap in expertise that cannot naturally be replaced.

There’s no simple answer. However, engineers and professionals must actively engage with both domain knowledge and the software they use. Their expertise in how things work should not diverge significantly from their understanding of how tools generate solutions.

Rather than treating projects as mere production tasks, they should be seen as opportunities for knowledge exploration. Progress should be accompanied by a conscious effort to learn, question, and refine—not just accept outputs at face value.

We must think at every stage and challenge each output, even when there are no immediate rewards for doing so. The countless thinkers before us didn’t do it for recognition, yet their efforts shaped the future. In the same way, even the smallest act of critical thinking today can influence the future.