Vibe Coding is complete trust in AI agents decision making without informed input from you. People who don't have any experience in software development, have no knowledge of best practices, architecture or the technologies and frameworks being used to develop their ideas. Everything is left up to the coding agent's decision making. So yes, we all use coding agents nowadays to develop our ideas. But those who understand the fundamentals of engineering know how to steer agents in the direction they want based on informed decisions, not just a vibe.
Why do we have to call it vibe coding at all? Why do we have to invent what ends up being derogatory terminology and ultimately name-calling of well-intentional folks trying to do the best they can with an idea, regardless of their experience?
I used "vibe coding" pre-AI to describe companies that build software without any long term strategy or vision - aka "Agile Theater" - just pick tickets off the board and move forward in whatever random direction your manager feels like today. Your manager's vibes are the only thing that matters.
Sadly, AI vibe coding is far worse. Following your manager's vibes, there was at least some semblance of accountability. Now when an AI vibe codes something, I'm not even sure whose ideas we're vibing on, or if they belong to anyone.
The "probleM"(™) I have with this notion and description, is that we're describing a human and socio-political problem, not a technical one. An individual or even organization that uses (for example) Claude Code, is NOT necessarily "vibe coding" or "vibing xyz".
We would have to find another way to describe what you've just pointed out, which is, _very real_, and often exists in many large corporations and large "complex systems".
When the author of the software cannot confidently explain or does not understand the bugs the software has, performance issues, security vulnerabilities or clumsy architectural decisions.
This is before the obvious Claudisms, I immediately can see in a project README which is a sign that the author did not read their own README, and expects others to read it!
But similarly, you would want an experienced captain to fly a commercial plane with passengers from A to B safely over someone with no experience who vibe-coded a flight simulator for fun to take control of a real cockpit.
There is a clear expensive line between someone experienced and knows what they are doing with AI vs someone who does not and vibe codes want they do not understand with AI.
Vibe Coding is complete trust in AI agents decision making without informed input from you. People who don't have any experience in software development, have no knowledge of best practices, architecture or the technologies and frameworks being used to develop their ideas. Everything is left up to the coding agent's decision making. So yes, we all use coding agents nowadays to develop our ideas. But those who understand the fundamentals of engineering know how to steer agents in the direction they want based on informed decisions, not just a vibe.
Why do we have to call it vibe coding at all? Why do we have to invent what ends up being derogatory terminology and ultimately name-calling of well-intentional folks trying to do the best they can with an idea, regardless of their experience?
I used "vibe coding" pre-AI to describe companies that build software without any long term strategy or vision - aka "Agile Theater" - just pick tickets off the board and move forward in whatever random direction your manager feels like today. Your manager's vibes are the only thing that matters.
Sadly, AI vibe coding is far worse. Following your manager's vibes, there was at least some semblance of accountability. Now when an AI vibe codes something, I'm not even sure whose ideas we're vibing on, or if they belong to anyone.
The "probleM"(™) I have with this notion and description, is that we're describing a human and socio-political problem, not a technical one. An individual or even organization that uses (for example) Claude Code, is NOT necessarily "vibe coding" or "vibing xyz".
We would have to find another way to describe what you've just pointed out, which is, _very real_, and often exists in many large corporations and large "complex systems".
"We all use AI at some point when building software"
False. If you don't know what your program is doing, and you're just 'trusting' the AI, you're vibecoding.
When the author of the software cannot confidently explain or does not understand the bugs the software has, performance issues, security vulnerabilities or clumsy architectural decisions.
This is before the obvious Claudisms, I immediately can see in a project README which is a sign that the author did not read their own README, and expects others to read it!
But similarly, you would want an experienced captain to fly a commercial plane with passengers from A to B safely over someone with no experience who vibe-coded a flight simulator for fun to take control of a real cockpit.
There is a clear expensive line between someone experienced and knows what they are doing with AI vs someone who does not and vibe codes want they do not understand with AI.