Vibe coding is the most discussed phenomenon of 2026 in tech communities. The idea: describe what you want in natural language, and let AI generate the code. Millions of non-developers are already building their first tools.
What vibe coding actually is
Popularised by Andrej Karpathy, former AI director at Tesla and OpenAI: describe your intent in natural language, AI generates, fixes and improves code in real time. A collaboration, not magic.
Dominant tools in 2026
Cursor: the reference AI editor
Cursor integrates Claude and GPT directly into the editor. Over 2 million active users, many non-developers. Particularly powerful for building dashboards, automating reports and creating small APIs.
Claude Code: the agent approach
Claude Code navigates your codebase, understands global architecture and modifies multiple files coherently. Ideal for medium-sized projects with some complexity.
For data profiles: vibe coding is particularly powerful for automating reports, creating interactive dashboards, cleaning data and building small APIs.
What vibe coding does not replace
Vibe coding often produces code that works but is not maintainable at scale. For critical systems, a real developer remains essential.
With care,
Excellent article, this matches exactly what we're seeing with our enterprise clients. The section on inference costs is especially valuable. It's a topic most articles gloss over but it's make-or-break at scale.
Thanks James! Inference cost optimization is often deprioritized during prototyping but becomes critical in production. Feel free to book a session if you'd like to go deeper on this.
Sharing this with my whole team. The distinction between an impressive demo and robust production is exactly the debate we're having internally right now. The human checkpoint advice is immediately actionable.
Great article. I'd push back slightly on the 18-day deployment estimate, in our experience with enterprise security and GDPR requirements, 4–6 weeks is more realistic for a first production agent.
Completely fair point David. The 18 days refers to a scoped first agent in a test environment. For full enterprise production with security constraints, your estimate is accurate.