Even ignoring the semantic drift that has happened since he coined the term (on which there have already been a few HN threads), the key part of Karpathy's definition is "...and forget that the code even exists." Which is why I was careful to phrase it thus:

> So this is still engineering, but it will be vibe coding in the sense that we almost never look at the code, we just look at the results.

It is pretty clear that "giving in to the vibes" is simply "looking at the results." But I'm predicting that it is going to be an engineering discipline in itself. Note that I started with (emphasis added):

> I think all coding will become vibe coding but it will be no less an engineering discipline.

And then I went on to explain the engineering aspect as extensive technical validation. There is a role called Validation Engineers in many industries including semiconductors, and I posit that it's going to be everybody's primary role soon.

> Responsibility is what cannot be vibe-coded. ... That isn't the point. Someone was held to account for the outages and had to explain why it happened.

I never implied a loss of accountability anywhere, but I completely agree, and have posted about it before: https://news.ycombinator.com/item?id=46319851

That is still orthogonal to vibe-coding. People have been sloppy without vibe-coding and were still held accountable. The flaw is assuming all vibe-coding is slop, because my point is that validation will matter much more than the code, which means soon we may never look at the code. In fact, extensive automated validation is probably a better signal for accountability than "We looked at the code very, very carefully."