- “Vibe coding” is the concept that even people who don’t know how to code can create digital products like apps or websites using AI by merely writing prompts. It became an overnight buzzword in Silicon Valley after being coined by OpenAI co-founder Andrej Karpathy. The new attention on AI-powered coding is redefining the barriers to innovation.
This year may be the year that Silicon Valley hits one of its pinnacles—traditional coding is out, and vibes are in.
That’s at least according to OpenAI co-founder Andrej Karpathy, whose post on X last month coining a new term sent Silicon Valley into a frenzy: “There’s a new kind of coding I call ‘vibe coding’, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists.”
Vibe coding works by typing or speaking a few sentences into an AI-powered coding platform (Cursor, Bolt, and Claude are a few examples), and a project that otherwise might take hours or days to code can be created in a matter of seconds. And even if the code quickly grows beyond one’s comprehension, that’s no concern; you can also prompt the AI to fix any bugs or make any desired changes just as easily.
While coding with generative AI is nothing new, its capabilities are improving by the day; in fact, Anthropic’s CEO Dario Amodei predicts AI will be writing all of the code in the next three to six months. So, if you want to jump on the vibe coding bandwagon, here’s what you need to know.
How to vibe code
With a growing number of code editors integrating AI into their platforms, practically anyone can learn to (vibe) code these days—and the process will likely only improve as advanced LLMs are democratized.
Let’s take Bolt, for example. Because it’s browser-based, it can be one of the easiest platforms to experiment with and understand the future of coding. After creating a free account, in as little as one sentence, you can ask the artificial intelligence to design a website, create a mobile app, or develop a video game—and the source code is provided at your fingertips.
For example, I asked Bolt to redesign the McDonald’s website to make it more modern and entice me to eat there:
In less than a minute, I had the foundation to make a decent-looking website. Even though most of the source code is foreign to me, I can make changes to the design by fine-tuning prompts, reviewing outputs, and giving feedback.
Here’s another example. I asked Bolt to create an innovative app to track my fitness goals. After the initial prototype was created, I prompted it to change a specific button red.
While these are two basic instances developed in a matter of minutes, web and app development is just the tip of the iceberg; imagine the possibilities if you have a true business idea. Prototyping that previously could take days of your time can be completed almost instantly.
Ethan Mollick, an associate professor at the University of Pennsylvania’s Wharton School, even recently used Claude to create a video game that includes mechanics of relativity—just by prompting and not touching a single line of code himself.
And while even Karpathy admits vibe coding is not perfect, it signals that a future where new innovations can be developed overnight—and lead to earth shattering business pursuits.
How vibe coding will impact software development careers
Vibe coding and the advanced capabilities of AI just might be the nail in the coffin when it comes to the traditional views of computer programming and software development skills. However, for those who embrace AI, it can open even more doors.
A team of just 10 vibe coders can easily be on their way to building the next multi-million dollar start-up, Garry Tan, CEO and president of Y Combinator, tells CNBC. That work otherwise might have taken 50 or 100 engineers.
“You know, maybe it’s that engineer who couldn’t get a job at Meta or Google, who actually can build a standalone business making 10 or 100 million dollars a year with 10 people,” Tan says. “Like that’s such a powerful moment in software.”
Tan recently tweeted another jarring stat: 25% of the current crop of Y Combinator startups used LLMs (AI large language models) to write 95% of their lines of code.
Even though the term vibe coding is less than two months old, tech experts and students alike have used generative AI for years to create software. However, this new Silicon Valley buzzword is likely to only expedite the need for AI skills. According to LinkedIn, AI literacy is the fastest-growing skill in the U.S. this year.
Computer science is dead—and vibe coding is part of the rebirth
Vibe coding—and AI more broadly—will soon reshape everything in the world of computing, says Jules White, professor of computer science at Vanderbilt University.
“It’s the most exciting time in computer science because all the computer science, in many ways that we’ve been doing for the past decades, is dead,” he says.
Read more from Fortune
And while it’s likely those in tech will be writing less and less code, the ability to read and understand code remains an important skill—at least for now, he adds.
“We don’t do a good job, I think a lot of times, of teaching people to read other people’s code and to understand other people’s design decisions,” White says. “I think that’s going to have to become a much bigger emphasis if I’m using generative AI to write the code.”
The benefits of learning how to code and prompt AI are not restricted to those who want careers in tech. As generative AI is expected to impact all areas of the workforce, learning to use it can give you an edge in any job market.
“It enables a lot of people who might not have a traditional software engineering background or access to it to just bring an idea to life, which I think is super, super exciting,” Antje Barth, a principal developer advocate at AWS tells Fortune.
Overall, Karim Meghji, chief product officer at Code.org encourages young learners in particular to not get distracted by the vibe coding “hype train” and instead focus on building a foundation with their own skills.
“Lean into using these tools. Understand tools, but don’t forget the basics. It’s like we all took a language when we were in school, and then we applied the language,” Meghi says. “That doesn’t mean that we didn’t have tools to help us apply the language, but we learned foundations so we could actually create good stories.”
