Hacker News Reader

I agree. However simple to understand code with high human efficiency is nearly impossible to measure. Honestly, maybe ChatGPT could give you the best evaluation/measure of that.

View on HN

“write code clear enough that an LLM can look at it and create an accurate summary of the high-level goals” gives me similar vibes to “explain like I’m 5”, which seems like it could be an interesting metric (and my gut tells me that things will go horribly wrong as soon as somebody turns this metric into a target :D)

It looks like a metric of the kind of lines of code.

Hugely useful for measuring a lot of things. But obviously bad for optimizing even if it's not a formal target.

It's easy to measure. You sit a dev down in front of it and time how long it takes them to understand it. That's not just an estimate that's the key value itself.

Reply on news.ycombinator.com