Andrew Ng Pushes Back on AI “Vibe Coding,” Calling It Hard Work, Not Hype

At a recent AI conference, renowned artificial intelligence expert Andrew Ng argued that the buzzword “vibe coding” misrepresents what developers actually do when programming with AI tools. Ng – a Stanford University professor and co-founder of Google’s Brain research lab – cautioned that using AI to write code is far from a casual endeavor, describing it instead as a “deeply intellectual exercise” that can leave engineers drained by day’s end businessinsider.com. Despite the laid-back name, Ng says this new style of coding is a real, demanding job – and he’s urging both companies and individuals to take it seriously. In the same breath, he praised AI-assisted coding as a “fantastic” boost for productivity, while insisting that “everyone should learn to code” even as AI systems become more capablebusinessinsider.com businessinsider.com.

• read more in our blogs: Vibe Coding: Karpathy’s Viral Term, Ng’s Reality Check, Klover’s Early Pioneering, https://www.klover.ai/vibe-coding-karpathy-viral-term-ng-reality-check-klover-first-mover-advantage/
• read more in our blogs: Vibe Coding – AI-Assisted Software Development, https://www.klover.ai/vibe-coding-ai-assisted-software-development/
• read more by Dany Kitishian, Google Gemini: “Vibe Coding” Uproar-Navigating the Realities of AI-Assisted Software Development, https://medium.com/@danykitishian/google-gemini-vibe-coding-uproar-navigating-the-realities-of-ai-assisted-software-development-c60c62eeac61
• read more in our blogs: Beyond the Vibes: Mastering AI-Assisted Coding in the New Era of Software Development, https://www.klover.ai/beyond-the-vibes-mastering-ai-assisted-coding-new-era-software-development/
• read more by Dany Kitishian, Google Gemini: Vibe Coding Revolution, https://medium.com/@danykitishian/google-gemini-vibe-coding-revolution-b84296e94e88
• Google: Vibe Coding & AI- A Paradigm Shift in Software Development, https://medium.com/@danykitishian/google-vibe-coding-ai-a-paradigm-shift-in-software-development-d305be86dbb4

Criticizing the “Vibe Coding” Label at LangChain Interrupt

Ng’s remarks came during a fireside chat in May 2025 at LangChain Interrupt, an AI agents conference in San Francisco. When asked about the emerging trend of “vibe coding,” Ng didn’t mince words about the terminology. “It’s unfortunate that that’s called vibe coding,” he said, explaining that the phrase can mislead people into imagining that engineers just “go with the vibes” – simply accepting or rejecting suggestions from an AI – rather than doing real engineeringbusinessinsider.com. The former Google Brain lead warned that the term “vibe” makes it sound like programmers are coding on autopilot. In reality, guiding an AI to write useful software “is a deeply intellectual exercise” that requires significant thought and oversightbusinessinsider.com. “When I’m coding for a day with AI coding assistance, I’m frankly exhausted by the end of the day,” Ng noted, underscoring that it takes mental energy to steer AI tools effectivelybusinessinsider.com. Ng’s critique is a response to the growing hype around AI-assisted development. “Vibe coding might sound chill, but [the name] is unfortunate,” he said, pushing back on the idea that one can program simply by feeling things outbusinessinsider.com. He worries the label downplays the skill involved: “It’s misleading a lot of people into thinking, just go with the vibes, you know – accept this, reject that,” Ng explainedbusinessinsider.com. His message at the conference was clear – even if an AI is doing the typing, the human in the loop is not on cruise control.

What Exactly Is “Vibe Coding”?

The term “vibe coding” has gained traction in Silicon Valley this year, originally coined by AI researcher Andrej Karpathy in February 2025. Karpathy – an OpenAI co-founder and former director of AI at Tesla – used the phrase to describe a new approach to programming where instead of writing every line of code, developers collaborate with AI assistants through prompts and feedbackbusinessinsider.com businessinsider.com. In vibe coding, you tell the AI what you want in plain language or pseudo-code, and the AI generates the actual code. Karpathy provocatively described this style as “fully give in to the vibes” and “forget the code even exists” businessinsider.com, meaning a programmer focuses on high-level ideas while trusting the AI to handle the low-level implementation. It’s an almost hands-off strategy: “It’s not really coding — I just see stuff, say stuff, run stuff, and copy-paste stuff, and it mostly works,” Karpathy quipped of the experience businessinsider.com. In practical terms, vibe coding means using AI-powered tools to do the “heavy lifting” of software creation

businessinsider.com. A developer might sketch out an app by conversing with an AI agent, rarely touching the keyboard except to guide or correct the machine. Proponents say this lowers the barrier to entry for programming – even people with little formal coding training can build simple applications by “just a few prompts” to an AIbusinessinsider.com. In one example, a Block product designer with no engineering background managed to create a dog identification app in two months through vibe coding techniquesbusinessinsider.com. The rise of vibe coding has shaken up how people think about software development, sparking excitement as well as anxiety. Some engineers wonder if smarter code-generating AI could “put them out of a job,” and some tech investors debate whether future startup founders really need traditional programming skillsbusinessinsider.com. This is the backdrop for Ng’s skepticism of the term – he wants people to realize that there’s more going on under the hood than just “going with the flow.”

AI-Assisted Coding: A Powerful Tool That’s Mentally Taxing

Despite his gripe with the name “vibe coding,” Andrew Ng is enthusiastic about the practice of coding with AI. He is quick to point out the benefits of these AI coding assistants. “It’s fantastic that developers can now write software faster with these tools,” Ng said, sometimes completing features while “barely looking at the code” at allbusinessinsider.com. By handling boilerplate code and suggesting solutions, AI systems like GitHub’s Copilot or OpenAI’s Codex can dramatically speed up development cycles. Ng noted that his own teams have embraced AI helpers to the point that “my teams and I just hate to ever have to code again without AI assistance”, calling the productivity boost addictivebusinessinsider.com. In Ng’s view, these tools are not making programmers obsolete – they’re making programmers more effective. He even urges more companies to let their engineers use AI when coding, as many are still hesitant. “It’s been interesting to see how many companies – CIOs and CTOs – still have policies that don’t let engineers use AI-assisted coding,” Ng observed, arguing “we have to get past that” mindsetbusinessinsider.com. However, Ng stresses that using AI doesn’t make the job easy. On the contrary, directing an AI to craft code can demand intense focus. The human developer must constantly review AI outputs, integrate them with project requirements, and catch subtle errors. Ng described coding alongside an AI as cognitively exhausting work, reinforcing that there is “a lot of thinking” involved at every stepbusinessinsider.com. In other words, the AI can take over typing the syntax, but it’s still up to the human to ensure the program does what it’s supposed to do. This mental heavy-lifting is why Ng finds the “vibe” branding misleading. Far from just coasting on machine-generated code, developers are engaging in a new kind of high-level problem solving – they’re architects and editors of the AI’s output. Ng’s balanced view is that AI-assisted coding is both powerful and demanding: it can accelerate software development and handle routine code, but it also requires human oversight and leaves the developer with plenty of complex decisions. The end result, Ng emphasizes, is worth the effort, as long as one approaches these tools with the right mindset – as partners that still need guidance.

“Everyone Should Learn to Code,” Even in the AI Era

A key point Ng wants to drive home is that the advent of AI coding tools does not erase the need for human programmers. In fact, he believes the opposite – that basic coding knowledge is becoming more important than ever. Ng openly disagrees with advice suggesting people skip learning to code because “AI will automate it anyway.” “Over the last year, a few people have been advising others to not learn to code on the basis that AI will automate coding,” Ng said. “I think we’ll look back at [that] as some of the worst career advice ever given.”businessinsider.com

Instead, Ng argues, “everyone should learn to code” – at least one programming language – no matter what field they work inbusinessinsider.com. Why bother learning to code if AI can write programs for you? Ng’s reasoning is that understanding the fundamentals of coding dramatically improves your ability to use AI effectively. “One of the most important skills of the future – for developers and non-developers – is the ability to tell a computer exactly what you want so it will do it for you,” Ng explainedbusinessinsider.com. In essence, coding is a form of precise communication with machines. Knowing how software works under the hood helps people craft better prompts and instructions for AI, much like knowing a bit of a foreign language helps when using a translation app. If you can read error messages and grasp what the AI is doing, you won’t be mystified when things go wrong. Ng noted that understanding how computers work – including how to spot and interpret errors – helps people write more precise prompts for AI coding assistantsbusinessinsider.com. Rather than rendering programming skills irrelevant, the rise of AI elevates the value of computational thinking in all sorts of jobs. Ng practices what he preaches at AI Fund, the startup incubator he founded. He revealed that at his company, everyone knows how to code, “including the chief financial officer, the general counsel, and even the receptionist.”businessinsider.com

 They’re not all software engineers by trade, but each employee is encouraged to pick up some coding literacy to enhance their primary role. “It’s not that I want them to be software engineers. They’re not,” Ng said. “But in their respective job functions, by learning a little bit about how to code, they’re better able to tell a computer what they want it to do.”businessinsider.com

 Even a small amount of coding knowledge empowers non-engineers to automate tedious tasks, analyze data, or prototype an idea without waiting for IT support. According to Ng, that clarity in instructing computers is “driving meaningful productivity improvements” across his company

businessinsider.com. In short, Ng sees coding as a new literacy for the workforce: just as spreadsheets or word processors became essential skills, being able to bend AI tools to your will can make any professional more effective.

A Career at the Forefront of AI Innovation

Ng’s perspective on AI-assisted programming carries special weight given his influential career in technology. A pioneer in the field of machine learning, Andrew Ng co-founded and led Google’s deep-learning research initiative known as Google Brain, which in the early 2010s demonstrated the remarkable potential of neural networks at Googleen.wikipedia.org. He also spent years as a professor at Stanford University, where he directed the Stanford AI Lab and taught one of the university’s most popular computer science courses. In 2012, Ng co-founded Coursera, the online education platform, through which his machine learning courses have reached over 8 million students worldwideen.wikipedia.org. These accomplishments have made Ng one of the most influential figures in the AI community, widely credited with democratizing AI education and researchen.wikipedia.org. Today, Ng continues to shape the AI landscape as the founder of DeepLearning.AI and AI Fund, organizations dedicated to training talent and launching new AI startups. His broader impact extends from cutting-edge research to hands-on mentorship of the next generation of AI practitioners. This background helps explain Ng’s stance on vibe coding: having guided thousands of budding developers, he’s keenly aware that effective programming – even with AI in the mix – relies on human creativity, logical rigor, and learning. The software development world is evolving rapidly with tools that can generate code on the fly, and Ng’s insights offer a grounded perspective amid the hype. Yes, AI helpers are changing how software is built, making coding more accessible and fast. But Ng’s core message is that the human element remains critical. As the era of “vibe coding” unfolds, Ng is championing a future where developers and AI work hand-in-hand – with humans providing the vision and understanding, and AI providing the brute-force speed – to create software in ways we only dreamed of a few years ago. In his view, the best outcomes will come to those who embrace these powerful tools without losing sight of the skill and intellect that real coding, vibe or otherwise, will always demand businessinsider.combusinessinsider.com.

Citations

1. Andrew Ng Says Vibe Coding Is an ‘Unfortunate’ Term – Business Insider. https://www.businessinsider.com/andrew-ng-vibe-coding-unfortunate-term-exhausting-job-2025-6
   
2. Silicon Valley’s Next Act: Bringing ‘Vibe Coding’ to the World – Business Insider. https://www.businessinsider.com/vibe-coding-ai-silicon-valley-andrej-karpathy-2025-2
   
3. Andrew Ng – Wikipedia. https://en.wikipedia.org/wiki/Andrew_Ng