Mar 13, 2026
The Dracula Effect
Steve Yegge calls it "The Dracula Effect" — AI-augmented coding is vampiric. You ship at full speed, feel great, then find yourself napping mid-afternoon.
The mental model that clicks for me: a developer's day used to be a mix of easy and hard problems. Write a config file, rename some variables, debug a gnarly race condition, scaffold a new endpoint. The easy stuff gave your brain breathing room between the hard parts. Now AI handles all the easy tasks — the boilerplate, the repetitive refactors, the straightforward implementations. What's left is a concentrated stream of hard problems, all day. That's what drains you.
Yegge's take is that three hours of vibe coding at max speed might be all you can reasonably get from someone. Those three hours still produce way more output than a full pre-AI day, so it works out. The key is that everyone — engineers and managers alike — needs to recognize this new reality and adjust expectations accordingly.
This explanation clicked so much with me as I'm working on a big migration and I was pushing myself so hard that I felt I'm breaking and now I understand why. I needed time to assimilate the new information. The "hard" work needed mental space to be sorted and figured.