Hacker News Reader

Curious how much AI driven code can be generated vs. human intervention. I would imagine it depends on the app's complexity, but for some basic functionality it would be good to know how much of it can be created via AI. Also curious about the quality. Would you be spending more time correcting inefficiencies vs. having just written it yourself?

View on HN

With these autodevs, we're starting from 100% generation, and limiting the use case UI prototypes to start. But eventually want to move to other use cases.

In terms of % generated vs. % human for larger projects we take on - yes, you're right it's very dependent on the app's complexity, but for basic functionality I'd estimate 70-80% generated; we use humans for hook up around edges.

We do a lot to enforce quality and making it seem "human written" since we also have humans writing code and quality matters for making that efficient

Reply on news.ycombinator.com