Countless posts on LinkedIn poke fun at the recent emergence of a new job title adorning profiles: Vibe Coding Clean-up Specialist.
Analysis conducted by Short Fuse tracked down over 350 profiles marketing themselves with the expertise. There are very few job descriptions, however, mentioning vibe coding at all, let alone touting for skills needed to cleanup the mess of non-technical code development. In part, the job descriptions seem to stem from discontent amongst technical staff, either disgruntled with the perceived decrease in the value of programming skills in the wake of vibe-coding platforms like Lovable. Or genuine fatigue with being pushed to unpick the chaos of user-created code.
But aside from the amusement of industry commentators, the growth in the job title paints an interesting picture. Rabid adoption of vibe-coding by enterprises has succeeded, relative to previous iterations of citizen development, such as low-code, no code. Lovable AI now enjoys a valuation in excess of $1bn. Cursor AI is just shy of $10bn.
Research on the challenges with scaling vibe coding paints a less rosey picture. Analysis highlights myriad issues. AI-generated code, especially from vibe coding platforms, is frequently plagued by logical errors that are difficult to detect since developers did not write the original logic.
AI has also been accused of prioritising correctness over efficiency, leading to performance bottlenecks and making applications slower and less scalable. Security vulnerabilities are also a significant concern, as AI fails to anticipate adversarial threats, potentially exposing software to hackers. But of most pressing concern given pace of adoption is vibe coding can result in bloated and hard-to-maintain code, creating challenges for both initial development and future rewrites—in some cases cementing legacy and exponentially increasing costs.
It’s unclear whether vibe coding cleanup experts or better platforms will solve these issues, but increased scrutiny on AI, a major vibe coding disaster could be what bursts the bubble.
