Question 1

What is vibe coding?

Accepted Answer

Vibe coding is a term coined by AI researcher Andrej Karpathy on February 2, 2025, describing a development approach where programmers use AI tools to generate code based on natural language prompts while paying minimal attention to the underlying code itself. Karpathy described it as: 'fully give in to the vibes, embrace exponentials, forget the code even exists.' The term was named Collins English Dictionary Word of the Year for 2025. In practice, vibe coding means accepting AI-generated output without deeply understanding or reviewing it, prioritizing speed of output over code comprehension.

Question 2

How much code is AI-generated in 2025 and 2026?

Accepted Answer

Multiple sources confirm that AI-generated code has reached significant scale. ShiftMag reported that 41% of all code written in 2025 was AI-generated. Microsoft CEO Satya Nadella stated at LlamaCon that 20-30% of Microsoft's code is AI-written. Google CEO Sundar Pichai confirmed 25% of Google's code is AI-assisted. Garry Tan reported that 25% of the Y Combinator Winter 2025 batch had codebases that were 95% or more AI-generated. GitHub Copilot has over 20 million users with 42% market share, Cursor reached $2 billion in annual recurring revenue, and Claude Code hit $2.5 billion in annualized billings.

Question 3

Does AI-generated code have more bugs than human-written code?

Accepted Answer

Yes, multiple studies confirm higher defect rates in AI-generated code. CodeRabbit found that AI-authored pull requests average 10.83 issues compared to 6.45 for human-authored PRs, making AI code 1.7x more bug-prone. Veracode found that 45% of AI-generated code samples failed security tests, with Java code failing at a 72% rate. XSS vulnerabilities are 2.74x more likely in AI-generated code. Faros AI found that teams with high AI adoption saw bugs per developer increase 9%, and Cortex reported incidents per pull request up 23.5% and change failure rates up 30%.

Question 4

What is the METR productivity study on AI coding tools?

Accepted Answer

METR (Model Evaluation and Threat Research) conducted a randomized controlled trial in 2025 that produced a striking finding: developers using AI coding tools were actually 19% slower on real-world tasks, but believed they were 20% faster. This represents a 40-percentage-point perception gap between actual and perceived performance. The study used experienced open-source contributors working on their own repositories, controlling for familiarity and expertise. The result suggests that the perceived productivity gains from AI coding tools may be substantially overstated, driven by the psychological experience of generating code faster rather than the actual time to complete working features.

Question 5

How is vibe coding affecting junior developer hiring?

Accepted Answer

Junior developer hiring has declined sharply since AI coding tools became widespread. Junior developer hiring is down 67% since 2022. US programmer employment fell 27.5% between 2023 and 2025. A survey found that 54% of engineering leaders plan to hire fewer junior developers due to AI capabilities. A Harvard study found that junior developer employment drops 9-10% within six quarters of AI tool adoption at a company. This creates a long-term pipeline crisis: if companies stop hiring juniors, they lose the training ground that produces the senior engineers needed to oversee and correct AI-generated code.