“AI won’t replace developers, but developers who use AI will replace those who don’t.”
- Andrej Karpathy, AI researcher and founding member of OpenAI
Coding is no longer just writing code
It was unthinkable just 2 years ago. But for thousands of software coders across the world, in large corporations or at startups, hammering out code painstakingly line-by-line has started feeling slightly dated. Many coders now follow what Karpathy famously called “vibe coding”, i.e. coding using AI tools. In a way, they are co-creating software solutions. Compared to normal programming, this may be one of the biggest shifts in coder behaviour in a long time.
Rise of the 'AI collaborator'
Before LLMs (large language models) and ChatBots hit the scene in 2023, all this was unthinkable. But from 2024 onwards, prompt engineering skills rose in demand, not just for the lay users of LLMs but for experienced coders as well. By mid-2025, powerful tools like Copilot and Cursor started enabling engineers to delegate routine coding tasks. Some time was freed up to focus on the design, debugging, and domain alignment aspects of projects. So coding well is now accompanied by prompting well. Man is talking to machine, using natural language. [/lock]
Vibe coding needs new skills
This new wave of coding - Vibe Coding - demands human skills like analytical communication, language precision, and a sharp eye for quality. Coders need to translate complex problems into prompt-friendly English and evaluate AI outputs carefully. Noting is a given. AI makes mistakes, and it readily accepts it too! A smart vibe-coder evolves his own code-hygiene protocol, including checks like "avoid redundancy", "is this really correct", "are the dependencies truly functional", and so on.
Industry-wide transformation underway
All said, this change is surely industry-wide. According to various industry observers and news reports, more than 70% of Indian firms now use AI in at least one function. Companies are using AI not just to cut costs but to compress timelines and boost product delivery. AI coding tools are proliferating, most built on the back of powerful LLMs like Claude or Gemini. Even newbies with no coding background are happily trying their luck with these tools, though not always with gratifying results.
Software development redefined
There is the element of a tech shift, and one of a cultural shift as well. As said, man is being evaluated on his/her ability to work with machines. Tech roles are now demanding communication skills, that form the foundation of how one talks to Vibe Coding platforms. Coders now need a good mix of reasoning, logic, communication and lot of patience.
Summary
AI is fundamentally reshaping the role of software engineers. Job listings now routinely demand AI-collaboration skills. Man is now co-creating with machines. This wave of ‘Vibe coding’ reflects a new reality. It has its downsides, which if carefully managed, can be tackled well.
BH food for thought
If AI now helps write the code, who takes responsibility for bugs—the human or the machine?
Expertspeak
Vibe coding blends prompt engineering with human intuition, requiring developers to master context-switching, linguistic precision, and interpretive logic. As LLMs become co-developers, the traditional debugging-test-deploy loop is disrupted, pushing the frontier toward cognitive orchestration rather than syntax mastery.
Read more
- Harvard
Business Review – “AI won’t replace coders, but coders who use AI will
replace others”
https://hbr.org/2023/09/ai-wont-replace-coders-but-coders-who-use-ai-will-replace-others - MIT
Technology Review – “How GitHub Copilot is changing the way developers
code”
https://www.technologyreview.com/2023/06/13/github-copilot-ai-developer-coding - McKinsey
& Company – “The state of AI in 2025: adoption, impact, and the road
ahead”
https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2025
COMMENTS