Post
Indeed, I don't think there's much distinction between "agnetic engineering" and "vibe coding" either. https://simonwillison.net/2026/May/6/vibe-coding-and-agentic-engineering/
The incentives of high output generation with review being the bottleneck will clearly move users towards just vibing and less reviewing. (But that has seemed obvious to me for a long time, anyway)
@cwebber Maybe this is just me, but if I took barely a year to go from "how dare people suggest that vibe coding and ✨agentic engineering✨ are the same thing" to "on further reflection, this is just vibe coding," I would not have the gall to continue pretending to be a Serious Thinker without even breaking stride
@cwebber As it has already been put before: making more code of lower quality had never been a bottleneck.
The incentive to skip review is real, but it doesn't erase the distinction. "Agentic engineering with careful review and tests" is a discipline; vibe coding is skipping that step. The problem isn't the tooling, it's whether you treat review as optional.
@cwebber “it’s the same picture”
@cwebber Time to make agentic automation scalable and that is our key focus of 2026 /s
@cwebber I agree with the blog post that "agentic" isn't much of a middle ground. But I don't think FOSS needs to compete with enterprise head-on in terms of code volume. We've always found ways to punch above our weight and should think harder about this.
Anyone who is pushing FOSS projects to embrace AI based PRs with thorough review should be thoughtful that this logical endpoint will come to their beloved FOSS projects also.
@cwebber Willison's distinction (and previous iterations of it) strikes me as dangerously naive. Based on my own experience with both the tool and, y'know, human nature, nothing but the most draconian guardrails will prevent people from taking the easy way with these tools.
We were never going to do this safely.
@mttaggart @cwebber I think "naive" is being overly generous here. The phrase "responsible use of AI to write code" is slipping a hell of a lot of unevidenced and extraordinary claims in — I read that post as yet more boosterism.
We should reject slop in OSS, whether it's branded as "vibe coding" or "agentic engineering," and following the same ethical imperatives.
@xgranade @mttaggart @cwebber simon's a very neutral party, just ask him
Willison's ... dangerously naive.Yes. See also https://bucci.onl/notes/Word-calculators-dont-add-up
high output generation with reviewseems to be load bearing in the push to widely deploy vibe coding or agentic coding or whatever they're calling it today. "It's OK to have LLMs produce code as long as there is thorough human review" is an argument I've seen trotted out countless times, but should not be given any credence.
Install bonfire.cafe
Get the full app experience
