This really resonated with me. AI can absolutely make development faster, but I think the part we underestimate is the mental cost of constantly reviewing, questioning, correcting, and keeping track of everything it produces.

When you write the code yourself, you naturally build a mental model along the way. With AI, you can get to the result much faster, but sometimes you’re left trying to understand a much larger amount of code and decisions after the fact. That can be a very different kind of tired.

The line between “I’m being more productive” and “I’m just processing more information” feels surprisingly thin here. Maybe the next challenge with AI-assisted development isn’t making AI generate even more code—it’s figuring out how to reduce the cognitive load that comes with supervising it.

Really relatable perspective.