AI assistants can feel like a superpowered thinking partner, helping solve problems at lightning speed 1. But they have a critical flaw: they often produce answers riddled with hallucinations, outdated information, or gaps in domain knowledge.
It’s clear that Generative AI is here to stay, and we need to treat it like any other powerful tool in our professional toolkit. A developer masters their toolchain, a designer masters their software, and we must all learn to leverage AI.
But a tool is only as good as the person using it. Gen AI isn’t perfect; it makes mistakes and can misinterpret requests because natural language has inherent expressive limits 2. Trusting its output blindly is a shortcut to poor quality, whether in code, analysis, or strategy.
This is why fundamentals are more critical than ever. When you have a deep understanding of your domain, you can guide the AI, spot its errors, and correct them. This understanding transforms you from a passive consumer into an active collaborator, turning a flawed tool into a powerful one.
This is especially vital for the next generation. We owe it to our children to teach them foundational principles. Just as we guide them on using the internet safely, we must equip them to be masters of these new AI tools, not the other way around.
Footnotes
-
It feels amazing when an agent helps you quickly fix an issue that you know is there, that you know will take a few hours of coding, in a few minutes. ↩
-
Spec-driven development can help remediate many of the downsides noted above. However, when using a chat to interact with the AI, which is what most of the people do, the guardrails are hard to set. ↩