Anthropic's Boris Cherny rejects 'vibe coding' label
Key insights
- Cherny, who called engineering 'dead,' now objects to 'vibe coding' as too loose and unserious a label.
- The dispute is substantive: enterprise code quality standards require structured practices, not improvisational workflows.
- Anthropic's own tooling team is signaling that 'vibe coding' misrepresents what rigorous AI-assisted development looks like.
Why this matters
The terminology battle between 'vibe coding' and structured AI development is directly shaping how enterprises write procurement requirements and security policies for AI-generated code, with real procurement and liability consequences. Engineers and founders building AI coding tools now face a credibility problem: their most viral framings are undercutting the professional legitimacy their enterprise customers need to justify adoption. If the dominant cultural label for AI-assisted development implies low rigor, it gives procurement and legal teams grounds to reject or heavily restrict AI code tooling regardless of actual output quality.
Summary
Boris Cherny, the Anthropic engineer who sparked debate by declaring traditional software engineering 'dead' in the Claude Code era, is now pushing back on the term 'vibe coding' that his earlier comments helped popularize.
Cherny's frustration reflects a real split inside AI tool development circles: the people building these systems want rigorous, auditable, structured workflows, while the 'vibe coding' framing has come to imply something closer to improvisational guesswork. That gap matters because enterprises evaluating AI-generated code need clear quality and ownership standards, and a culture built around vibes does not provide them.
Essentially: (Anthropic, Claude Code) are caught between democratizing coding and maintaining professional credibility.
- Cherny previously said engineering was 'dead,' a claim that fueled the vibe coding discourse he now wants to distance from.
- The tension is between AI-assisted development as a serious engineering discipline versus a loose, output-first culture with low accountability.
- Enterprise adoption of AI code tools depends heavily on how liability and code quality are framed, making terminology a practical, not just semantic, issue.
How the industry names AI-assisted development will shape the governance frameworks, procurement criteria, and liability standards that follow.
Potential risks and opportunities
Risks
- Enterprise legal and compliance teams could use the 'vibe coding' framing as justification to block or delay AI code tool adoption in regulated industries, slowing revenue for Anthropic and competitors through mid-2026.
- If 'vibe coding' culture becomes the dominant mental model for junior developers onboarded to AI tools, code review and security audit standards at mid-market companies could erode before governance frameworks catch up.
- Anthropic risks brand inconsistency: Cherny's earlier 'engineering is dead' claim and his current reversal give critics material to argue the company does not have a stable or trustworthy position on AI development practices.
Opportunities
- Coding standards bodies and enterprise software vendors (JetBrains, GitHub, Atlassian) could move to define and brand a 'structured AI development' methodology that fills the vacuum Cherny is pointing to.
- AI code quality and audit startups (Codacy, SonarSource, Snyk) gain a clear marketing opening to position their tools as the antidote to vibe coding's accountability gap.
- Anthropic could publish a formal Claude Code best-practices framework under Cherny's name, converting internal frustration into a thought-leadership asset that differentiates Claude Code in enterprise sales.
What we don't know yet
- Whether Anthropic has an internal style guide or quality framework for Claude Code outputs that defines what 'rigorous' AI-assisted development means in practice.
- How major enterprise customers (financial services, healthcare) are currently classifying AI-generated code in their audit and ownership documentation, given the absence of an industry standard.
- Whether Cherny or Anthropic will propose an alternative term or framework to replace 'vibe coding' in professional contexts, and on what timeline.
Originally reported by timesofindia.indiatimes.com
Read the original article →Original headline: Anthropic Engineer Boris Cherny, Who Said Engineering Is 'Dead,' Now Says He Is Getting Sick of the Term 'Vibe Coding'