Is vibe coding dying?

It's only vibe coding during the first ~15 prompts, after that it becomes "managing a dumb senior" coding.

Well, I get impressive results with Kiro [0] by AWS. It's specs-based but takes a lot longer than just generating a whole working program with a single prompt. GitHub has their Spec Kit [1] and there's OpenSpec [2] now, so, things will get better.

[0]: https://kiro.dev/

[1]: https://github.com/github/spec-kit

[2]: https://openspec.dev/

> Remember how in October and in March I told you that vibe coding — in the sense of amateurs using large language models to write code to “build products that would have previously required teams of engineers”

I don't think there's an agreed upon definition for what vibe coding is - but certainly the "amateur" portion of this sentence is _not_ a requirement to be vibe coding. Vibe coding can be done with success by skilled developers with existing experience.

But sure, change the goal posts for your catchy clickbait article title.

Was it ever alive?

a potential speculation: those that used these tools are now satisfied with fewer tokens as models get better? I'm using only haiku now and it serves my needs better than opus did only 4 months ago!