BOOK THIS SPACE FOR AD
ARTICLE AD
The internet (and especially my inbox) has been awash with buzz about vibe coding, a term coined by a former OpenAI founder and master programmer who waxed poetic about using an AI to do a lot of his coding's more prosaic work. I went into substantial depth about that post and the definition of vibe coding in my previous article on the topic.
Also: 10 key reasons AI went mainstream overnight - and what happens next
The term, though, is sticky as heck. By sticky, I mean that it resonated enough to be both cool on its own and be an immediate adjective that marketing folks can attach to their coding tools, making those tools immediately sound a notch cooler than they were on their own. I know, because as an AI columnist, I get press releases. I get a lot of press releases.
Unfortunately, the marketing folks who wasted no time adopting the term and vibe-washing their products seem to be applying the vibe-coding moniker to the exact opposite of what it means. Remember, the term was coined by a master programmer who could very easily decide what's appropriate for an AI (and fix it, if it goes wrong). It was not intended for a newbie who wants to type a two-sentence prompt and make a couple of million dollars off the app that results from that mere moment of mental mathematics.
But suddenly, all the so-called no-code and low-code AI development tools (which are basically templates dressed up in AI-painted robes) are claiming they are vibe coding innovators.
As a long-time developer myself and former computer science professor, I shared my opinion on vibe coding in my article last week. But I'm cranky, crotchety, sarcastic, and snarky. I know, how could you possibly tell from my writing?
Also: The best AI for coding (and what not to use)
So rather than just leaving you with my opinion, I decided to reach out to real professional developers and ask their opinions. As you're about to see, some like the concept and say that AIs are saving them a ton of time, but most are concerned that non-pro developers and junior coders might miss important coding considerations and thereby produce code that's somewhere on the spectrum from junk to janky to downright dangerous.
Developers on vibe coding
Todd Olson, co-founder/CEO of usage analytics company Pendo
He shared with ZDNET, "Vibe coding changes how we build, but not why we build. The goal isn't to produce more software; it's to create better software experiences. Without a clear understanding of user needs, vibe coding risks amplifying misalignment at scale."
Also: How to use ChatGPT to write code - and my favorite trick to debug what it generates
Olson continued, "AI makes it easier to ship code, but that doesn't guarantee better outcomes. Without clear alignment to user needs, vibe coding can accelerate the creation of software no one uses. This isn't a new challenge, necessarily, but AI raises the stakes."
Chris Reynolds, developer advocate and software engineer at hosting platform Pantheon
He told ZDNET, "The most important thing for me about vibe coding is that there needs to be a human operator who is actually smarter than the computer. One of the problems of using LLMs for development is they make stuff up that sounds feasible, and they only do deeper research or consider things like security or best practices if prompted."
He pointed out how having limited experience as a developer could prove a challenge when asking an AI to write software, saying, "So, if you're just an average power user without any development experience, you could absolutely get AI to build you a cool thing that looks and acts the way you want, but if you don't know how any of it works, when it has issues or bugs or whatever, you can find yourself deeper and deeper in the rabbit hole."
I love his comments about robots running wild: "Vibe coding absolutely needs the human component. You can't just let the robots run wild without checking what it's doing, or you'll absolutely get yourself into trouble."
Jamie Marsland, YouTube manager and a developer at Automattic
Jamie told me, "We're entering a new era where the web becomes a canvas for dynamic, app-like experiences. As AI, no-code tools, and automation remove technical barriers, creators of all kinds are free to build richer, more interactive digital worlds. The lines between websites and apps are blurring -- and with it comes a surge of creativity that redefines what's possible online."
He also shared a YouTube video he posted where he uses AI to "vibe code" a front end to a website. It's fascinating, but keep in mind, this is a veteran developer instructing the AI. Someone completely new to web development might not have the vocabulary or experience to give the level of direction Jamie does.
Bill Salak is the CTO of AI education company Brainly
He told me, "In tech, vibe coding is all about working with AI in a fluid, intuitive way, letting it handle the repetitive stuff so developers can focus on creative problem-solving. But vibe coding isn't just for coders. It's also about creating with AI even if you don't know how to write code. It's about removing barriers, making technology accessible, and allowing anyone to build something meaningful. So why aren't we applying that same mindset to learning?"
"Vibe coding is proof that AI isn't here to replace creativity -- it's here to unlock it. It turns coding from a technical skill into a creative collaboration, where anyone with an idea can bring it to life."
Salak extended the idea of vibe coding into learning overall. There's something to be said for the lower stress, higher interactivity idea that vibe coding promotes. He said, "Imagine if learning worked the same way, where students weren't just consumers of information but creators, shaping their education in real-time with AI as their guide."
"Just like AI is transforming coding, it's time for AI to transform learning. Less memorization, more creativity. Less pressure, more discovery. The future of education isn't one-size-fits-all. It's personalized, powerful, and finally built for the student (and the teacher, too)."
I do like the idea of AI providing more interactivity in our activities, but as with coding, where we've seen how AIs can generate terrible code, any learning-related activities will need to have substantial guardrails and quality checks. I'm not sure we're ready to unleash students (or newbie coders) into the AI world (and its relatively common hallucinations) without some professional supervision.
Willem Delbare, founder and CTO of Aikido
Aikido is a security company with the best tagline I've seen in a while: "No bullshit security for developers." Delbare doesn't pull any punches about vibe coding, telling ZDNET, "Vibe coding makes software development more accessible, but it also creates a perfect storm of security risks that even experienced developers aren't equipped to handle. SQL injections, path traversal, hardcoded secrets. Many developers consider security a blind spot, so imagine the risks when going from a simple landing page build to a full-fledged product handling user data."
He continued, "It is incredibly easy for AI to write vulnerable code, and vibe coders often don't know what they don't know. When you're collecting user data or handling sensitive information, you need more than just good vibes. Ideally, security isn't an afterthought that ruins the fun; you need to learn the risks and set up security checks so you can keep building without your users becoming collateral damage."
Also: Why OpenAI's new AI agent tools could change how you code
Then he made a very valid point about supercharging risk. He emailed me with, "Sure, Gen AI supercharges development, but it also supercharges risk. Two engineers can now churn out the same amount of insecure, unmaintainable code as 50 engineers."
Jamie Madden, founder of PetFun, a service that creates AI-enhanced pictures of pets
He echoed some of the security and support concerns we've been talking about.
He said, "Vibe coding represents a fascinating shift in software development by demonstrating how AI tools can democratize coding. While it enables newcomers to build applications without traditional programming experience, it also highlights a crucial reality: without fundamental software development knowledge, security and performance suffer."
Zack Katz, president of no-code forms platform GravityKit
Katz said vibe coding has been a game changer. He told ZDNET, "Vibe coding has dramatically accelerated our product development process. Ideas that sat in our backlog for years have come to life in just a week, thanks to AI. I can generate a working prototype of new functionality in a single day -- far faster than starting from scratch. From there, we refine, restructure, and test the code, cutting down development time significantly."
Then, he provided an example: "In a recent AI-powered session, I entered the requirements for a new product, switched back to my other work, and checked back a few minutes later. Within 20 minutes, I had a functional prototype ready for refinement and testing. AI has been a game-changer in our ability to quickly deliver features that customers want, and vibe coding has proven to be my go-to approach for rapidly building MVPs."
Leonardo Losoviz, developer of WordPress data management plugin Gato GraphQL
Losoviz also believes vibe coding can reduce time to market. He told me, "I'm currently integrating my plugin Gato GraphQL to use vibe coding to create extensions, with the goal of creating all the extensions that my customers are demanding in a couple of weeks (normally it'd take a few months)."
John Campbell, director of content engineering at Security Journey
Campbell has both positive and negative thoughts about vibe coding. He told ZDNET, "Vibe coding is set to revolutionize software development by making developers more efficient and enabling individuals with limited technical expertise to expand their access to software tools."
Also: Google just made AI coding assistance free for everyone - with very generous limits
But there are concerns. He said, "Just like a teenager experiencing the thrill of driving for the first time, that freedom would quickly turn to chaos without traffic laws, safety features, and guardrails."
He echoed Delbare's thoughts about accelerating risk. Campbell said, "Vibe coding accelerates development and removes the need for deep technical knowledge of tools or techniques, but at the cost of a near-total lack of system understanding. Because vibe coding inherently disregards code details, it becomes almost impossible to assess the security risks in the software being developed."
Campbell said he thinks the gap between vibe coding and production coding will continue to narrow. But for now, "Vibe coding today excels at creating 'web toys' -- personal tools with a narrow focus and minimal security concerns -- without the stakes and risks of deploying them in a production environment."
Rhys Wynn, freelance WordPress developer trading under the name Dwi'n Rhys
He reported to ZDNET, "I think vibe coding is a phrase invented by people who think that AI-generated code is safe and secure, and want to position themselves in a market they maybe shouldn't be in. Every single vibe coding project I've seen has been insecure, not able to answer a use case, or just emulating better things that exist already."
He, too, shared the security and optimization codes mentioned by most of the other developers I spoke to. He said, "AI can be useful for coding, but getting AI to write out large swathes of code means the code is tending towards being insecure or unoptimized. Ask your AI assistant to fix small problems with well-defined tasks, and it works well. Similarly, to fix problems in your code. Always review, as if you don't understand your code, who will?"
What does it all mean?
Some developers have seen rapid improvements in coding time using AI development tools. I'm on record as saying that "I am fairly convinced it has doubled my programming output. In other words, I've gotten twice as much done by using ChatGPT as part of my toolkit."
But was that vibe coding? It's true that I let the AI code some of the more tedious tasks that involved common programming language, saving me hours. But did I just "wing it" and let the AI do whatever it wanted? Oh, hell no. I tested every single call, every input, and every output.
Also: How to turn ChatGPT into your AI coding power tool - and double your output
Vibe coding, for a very experienced developer, is kind of like being in a flow state. A lot of work can be accomplished by being in flow. But the prevailing opinion amongst nearly all the developers here is that (a) vibe coding is not for newbies, and (b) vibe coding could leave some problematic security and reliability problems.
I don't think the idea of vibe coding, on its own, is a problem. But I would recommend keeping your bullcrap detector on full power, because marketing folks are seizing on the stickiness of the term to claim that newbies can do no-code projects with alacrity.
The problem with low-code and no-code will remain as it's always been: for customizing pre-built scenarios, the low-code and no-code tools work fairly well. But if the code you need doesn't fit the scenarios (or templates) provided, you're going to have to whip out that programming knowledge and write code like you know what you're doing.
Also: If ChatGPT produces AI-generated code for your app, who does it really belong to?
If you can't do that, you're going to run into brick walls or ship products that will fail in very day-ruining ways.
What about you? Have you tried vibe coding yourself, or are you still on the fence about letting AI take the lead in your development process? Do you think AI-powered tools truly lower the barrier to entry for newcomers, or do they just make it easier to create insecure, unmaintainable software?
Are you more excited by the creative possibilities or concerned about the risks? Let me know in the comments below.
Get the morning's top stories in your inbox each day with our Tech Today newsletter.