AI Should Elevate Your Thinking, Not Replace It — Koshy John on the Hidden Divide in Engineering

In a popular HN post (227 points, 186 comments), software engineer and writer Koshy John draws a sharp line between two types of AI use in software engineering. The first group uses AI to remove drudgery, move faster, and spend more time on high-value work — framing problems, making tradeoffs, spotting risks, and creating clarity. The second group uses AI to avoid thinking — pasting prompts, collecting polished output, and presenting it as their own reasoning. John calls the latter a dead end.

The New Failure Mode: Outsourced Thinking

John describes a dangerous pattern: engineers hand a model a problem, receive a plausible answer, and then repeat that answer without understanding it. He compares this to test-copying — good grades on paper, but no underlying structure. When faced with ambiguity, incomplete information, or non-template problems, shallow imitation breaks down.

“Every time you substitute generated output for your own comprehension, you are skipping the exercises / reps that build judgment. You are trading long-term capability for short-term appearance.”

The Calculator Analogy

John uses the calculator as a parallel: an engineer with strong mental math can use AI aggressively because they can sanity-check output, catch errors, and know when something sounds wrong. An engineer without that foundation becomes dependent on the tool and cannot detect garbage.

What the Best Engineers Will Do Instead

John argues that the most valuable engineers are those who “refuse to spend time on work that AI can do for them, while still understanding everything that is done on their behalf.” They use time savings to operate at a higher level, applying rigor rather than outsourcing thought.

The Risk for Early-in-Career Engineers

John warns that junior engineers are particularly at risk — they may look effective in the short term by using AI to generate output they could not produce themselves, but they miss the reps that build judgment. “That always catches up,” he says.

The full post includes a detailed breakdown of the dividing line, organizational implications, and why this matters more than most people think.