Has agentic AI had its moment in the sun? The explosive emergence of vibe coding suggests that might well be the case. According to Gartner’s recent report Why Vibe Coding Needs to Be Taken Seriously, it’s estimated that as much as 40% of new enterprise production software will be developed using vibe coding techniques and tools by 2028.
The appeal of vibe coding is clear; it is simple to use, fast, and creatively empowering, making for an effective tool for novice coders undertaking a solo build or weekend project. This was showcased by the recent story of a father-daughter duo who made a full small business website in six hours.
Businesses are also keen to hop on the trend and explore the benefits of vibe coding. Advocates of the technology praise its time-to-market advantage and potential to overcome the developer skills shortage issue.
Swedish AI unicorn Lovable, for instance, allows users to create working products in minutes with GenAI and expects to hit $1 billion in annualized revenue next year.
Rebecca Beach, the owner of an independent printables business, claims vibe coding has reduced the time it takes to create her products from weeks to a matter of minutes, doubling her monthly income.
These success stories may add to the vibe coding buzz, but agentic AI “fails” are tempering the allure. App-building company Replit was faced with catastrophe when its AI went rogue while vibe coding and deleted a business’ entire database.
So, can we say that vibe coding is enterprise-ready? The answer is easy. No.
Potential and risks
Vibe coding has yet to reach its full potential and presents too many risks. It may be perfect for low-risk personal productivity hacks, but for enterprises looking to create anything large-scale or overly complex, it remains too volatile and unreliable. It is irresponsible for businesses to forgo structure and planning, and simply go off “vibes”.
Enterprise organizations depend on robust financial systems, critical IT infrastructure, and almost inconceivable amounts of data to operate. Enterprise software requires a well-governed tech stack of advanced, fit-for-purpose tools and platforms, run by skilled employees.
Yet, some ambitious companies view vibe coding as a shortcut to lowering costs and headcount, only retaining minimal human employees where needed.
This approach is impulsive and risks becoming rash. While an AI-oriented approach to coding has the speed advantage, it lacks the structure and safeguards necessary for enterprises to operate successfully.
LLMs not ready for decisions
LLMs are proficient at pattern matching and generating plausible output, but they can’t grasp context and lack the emotional intelligence required to make business decisions.
An LLM cannot understand the intent of its own output which can pose immense risk. Just look at Replit; its AI saw no problem deleting all that data, and there were no error messages or governance checks to stop the disaster.
This episode highlights that the absence of human understanding also means a lack of governance and guardrails.
Enterprises must ensure all outputs align with business goals, ethical standards, and regulatory frameworks. Content created via vibe code may fail to take these into account.
Vibe coding is simply incapable of handling the level of responsibility required to enforce proper governance procedures, whether it be protecting customer data or vetting third parties.
Also, not all outputs are created equal. Yes, AI can generate code quickly. But fast does not automatically mean good.
Developers have already expressed their unease: although 76% now use or plan to use AI tools in their workflows, only 43% say they trust the accuracy.
Churning out software
The potential to churn out more software faster has led to a growing inclination for short-term thinking. As AI hype rolls on with the potential of delivering software at a breakneck speed, some organizations are betting on vibe coding replacing engineering talent.
Removing humans from the equation signals a concerning lack of foresight. There appears to be little consideration and appreciation for what only human beings can offer: strategy, nuance, and accountability.
While AI will undoubtedly remain fundamental to the future of software development, businesses must not get ahead of themselves and blindly follow the vibe coding trend without full consideration of how it will affect wider IT strategy.
Organizations must take time to build a comprehensive strategy fit for the long haul. When setting out governance plans, they should keep adaptability front of mind, to ensure they are able to leverage the best AI has to offer now and are ready to implement novel technologies yet to come.
But adaptability should fit in with the governance fundamentals: organizing your data, updating your SDLC to include data engineers and end users, and investing in talented employees who are eager to evolve.
Vibe coding in the world of enterprise software will need to understand the notion of non-functional software requirements, governance, and control. Until then, it’s a short-term con with limited long-term gains.
We’ve featured the best Large Language Models (LLMs) for coding.
This article was produced as part of TechRadarPro’s Expert Insights channel where we feature the best and brightest minds in the technology industry today. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/news/submit-your-story-to-techradar-pro
Add Comment