Every new trend comes with its own nomenclature.<\/p>\n
In the 1980s, it was yuppie for young urban professional. The 1990s saw the rise of the expression soccer mom. In the 2000s, we had hipsters.<\/p>\n
Now there\u2019s a new term in artificial intelligence programming: vibe coding<\/a>.<\/p>\n Coined by OpenAI<\/a> co-founder Andrej Karpathy<\/a>, vibe coding refers to the writing of computer programs without knowing programming languages. Instead, users use plain language and focus on the vibe of the project.<\/p>\n \u201cThere\u2019s a new kind of coding I call \u2018vibe coding,\u2019 where you fully give in to the vibes, embrace exponentials and forget that the code even exists,\u201d wrote Karpathy in a February post<\/a> on social platform X. \u201cIt\u2019s possible because the [large language models (LLMs)] \u2026 are getting too good.\u201d<\/p>\n There\u2019s a new kind of coding I call \u201cvibe coding\u201d, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It\u2019s possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper\u2026<\/p>\n \u2014 Andrej Karpathy (@karpathy) February 2, 2025<\/a><\/p>\n<\/blockquote>\n For example, a company can use the Code Interpreter<\/a> mode in OpenAI\u2019s GPT-5<\/a> and instruct it to create a chatbot that answers customer questions. The AI model will generate the application for a company to test and refine through more prompts.<\/p>\n Small\u2013 to medium-sized businesses (SMBs) can turn an idea into a working prototype in hours instead of weeks. Tools from platforms like Google AI Studio<\/a> are making this process accessible to non-technical users.<\/p>\n Some people have successfully launched businesses after creating apps using vibe coding, CNBC reported<\/a> in May.<\/p>\n Amjad Masad<\/a>, CEO of Replit<\/a>, a coding platform, said in an interview<\/a> with the Big Technology Podcast<\/a> that anyone from HR folks to doctors and Uber<\/a> drivers can now develop apps based on their ideas.<\/p>\n \u201cEveryone in the world has ideas,\u201d Masad said during the interview. \u201cPeople build so much domain knowledge about their field of work, but they never could make it into software because they didn\u2019t have the skill or capital.\u201d<\/p>\n For example, Maor Shlomo<\/a>, the non-technical founder of Base44<\/a>, used vibe coding to create a no-code development platform. Within six months, it attracted 250,000 users<\/a> and was acquired<\/a> in June by website creator Wix<\/a> for $80 million.<\/p>\n \u201cThis new approach allows people to simply express what they want to build, while intelligent agents do the heavy lifting,\u201d Wix said in a statement.<\/p>\n Here\u2019s one way to start, according to CNBC:<\/p>\n Read also: <\/strong>AI Coding Assistants Give Big-Tech Powers to Small Businesses<\/a><\/p>\n What Are the Risks?<\/strong><\/p>\n In practice, vibe coding changes the role of the human from writing every line of code to guiding the AI, validating results and deciding what\u2019s ready for production. This enables SMBs to experiment with new offerings, automate tasks or build internal tools without a large IT budget.<\/p>\n Vibe coding also gives users a faster way to test ideas, cut development costs and give creative teams the ability to build their own solutions. It can give them a competitive edge, letting small players innovate quickly without the overhead that slows down larger firms, Forbes reported<\/a> Aug. 4.<\/p>\n However, for use in core financial systems, regulated processes or customer-facing platforms that carry legal or financial risks, experienced developers should review<\/a> the code before launching it, according to the paper, \u201cVibe Coding as a Reconfiguration of Intent Mediation in Software Development: Definition, Implications and Research Agenda<\/a>.\u201d<\/p>\n \u201cBlack box codebases, ethical and data protection blind spots, and inconsistent documentation undermine auditability and verifiability, posing significant barriers to compliance in regulated sectors such as healthcare and finance,\u201d the paper said.<\/p>\n This speed and accessibility come with trade-offs. AI-generated code can be buggy, not secure, and difficult to maintain if no one fully understands how it works. Also, the technology isn\u2019t yet reliable enough for systems where accuracy, compliance and security are critical, the paper said.<\/p>\n As a result, companies must carefully weigh the risks and rewards of using vibe coding, ensuring it is used for the right use cases to prevent introducing a new risk to the business.<\/p>\n For all PYMNTS AI coverage, subscribe to the daily AI Newsletter<\/a>.<\/p>\n Read more:<\/strong><\/p>\n Apple and Anthropic Building AI-Powered Coding Platform<\/a><\/p>\n OpenAI Developing AI Agent to Replace Software Engineers, CFO Says<\/a><\/p>\n AI Agents\u2019 Rise Promises IT Revolution, but Readiness Questions Remain<\/a><\/p>\n\n