Learning to Code in an AI World Requires Slowing Down, Not Speeding Up

Artificial intelligence can solve coding problems in seconds, but it cannot replace the struggle that builds genuine expertise. That's the central tension Chelsea Troy explored in her Craft conference presentation—one that cuts to the heart of how software engineers should develop skills in an era of instant automation.

The Two Modes of Engineering#

Programming operates in two distinct rhythms: the rapid execution of known tasks and the slower deliberation required to understand how systems actually work. Troy argues that engineers often mistake speed for progress, accelerating through the second mode when they should be decelerating. "If doing stuff depends on figuring stuff out, then we're going to be stuck the faster we try to go," she explains. Genuine learning demands we shift into a lower gear.

Why Discomfort Drives Growth#

The neuroscience behind skill development reveals an uncomfortable truth: growth happens in confusion and struggle. Activities that genuinely teach us are inherently inefficient—they demand sustained attention and emotional engagement. "We have to sit with something that we feel bad at, and we have to struggle with it," Troy notes. "That uncomfortable emotion, the confusion, the frustration, is necessary for us to build cognitive capacity."

AI tools designed to accelerate outcomes can actually undermine this process. When a system writes the code for us, we bypass the cognitive work that turns knowledge into intuition.

Three Techniques That Stick#

Troy highlights three evidence-based learning strategies:

• Spaced repetition with sleep: Skill formation requires focused practice interspersed with quality rest. Coaching junior developers means spreading learning across multiple sessions to let their brains consolidate knowledge overnight.

• Prediction before revelation: Instead of passively consuming information, learners should pause at each step and predict what comes next based on what they know. This active forecasting dramatically increases retention compared to simply following along.

• Format diversification: People internalize information more effectively when encountering it through multiple channels—reading, listening, doing—rather than grinding the same format repeatedly. The "aha moment" typically arrives on the second or third exposure through a different medium.