Everyday AI Podcast – Ep 750: The Vibe Coding Boom: Why Vibe Coding isn't Going Away and How it's Both Good and Bad

Date: April 7, 2026
Host: Jordan Wilson

Episode Overview

This episode dives deep into the "Vibe Coding Boom"—the rise, impact, and future of building software via natural language with AI-powered platforms. Host Jordan Wilson examines how vibe coding is fundamentally changing software development both for seasoned engineers and non-technical users, the opportunities it creates, its major pitfalls, and why it’s here to stay. The episode also covers the top vibe coding platforms, the internal crisis many companies are beginning to face due to unstable AI-generated code, and the emergence of new disciplines like "rescue engineering."

Key Discussion Points and Insights

What is Vibe Coding? [03:55]

• Vibe coding means describing your desired software in plain language ("I want an app that...") and having AI build it, handling everything from code to deployment—even for those with little or no technical expertise.
• Coined by OpenAI’s Andrej Karpathy in February 2025, the term exploded in popularity as AI drastically lowered the barrier to software creation.
• "You describe something in plain English... and AI builds software." – Jordan [06:45]

The Vibe Coding Explosion [08:15]

• AI now generates about 41–46% of all new code globally.
• 92% of US developers use AI coding tools daily.
• Not just engineers: most vibe coding growth comes from non-developers or "builders."
• "Most of the people fueling all of this growth are not even software engineers." – Jordan [10:05]

Top 10 Vibe Coding Platforms and Categories [13:20–25:10]

1. Terminal/Command Line Coding Agents

• Claude Code / Claude Cowork (Anthropic)
  • Reads and rewrites entire codebases, offers agentic, autonomous changes.
  • Features a 1 million token context window, context-driven editing, and fast multi-file refactors.
• OpenAI Codex
  • Powers multi-service agent systems via desktop, command line, IDE, and web.
  • "I always have Codex and Claude Cowork running..." – Jordan [15:02]

2. AI-Native Code Editors

• Cursor
  • AI-first code editor; supports multi-models (GPT5.4, Claude, Gemini, etc.), hit $2B ARR with 1M+ paid users.
• Windsurf
  • Agentic IDE with a complex acquisition history; now partly under Cognition and Google.
• GitHub Copilot
  • Most widely used, offers code completion as an add-on to traditional IDEs; cost-effective with broad support.

3. Browser-Based App Builders

• Lovable
  • Prompt-to-deploy full-stack apps for non-developers; includes plan mode and built-in security testing.
• Bolt New
  • Fastest route from prompt to prototype, best for speed demos, offers free tier.
• Replit
  • Cloud IDE with multi-language agent, real-time multiplayer, scalable pricing.
• V0 from Vercel
  • Clean React/NextJS code, easy export, minimal lock-in.
• Google AI Studio
  • Free full-stack vibe coding, powered by Gemini, quick app building with robust multiplayer support.

Honorable Mentions

• Anti Gravity (Google)
• Other rapidly updating platforms cited as emerging but didn't make the top list.

The Double-Edged Sword: Benefits vs. Risks [25:20]

Opportunities

• Speed and Democratization
  • Timelines cut from months/weeks to hours (e.g., Zendesk, Autonom CRM case studies).
  • New products and services launched by non-engineers; entire enterprise stacks replaced by AI-generated alternatives.
  • "Their head of finance built a full CRM on Lovable... used to replace their $40,000 Salesforce contract." – Jordan [28:05]

Risks & Downsides

• Software Fragility & Lack of Trust

  • Analogy: "You can talk your dream house into existence, but walk in and the floors are sagging."
  • Over 63% of vibe coders have never written a line of code.
  • Trust in AI-generated code has dropped: 43% → 29% (Stack Overflow study).
    • Top frustration: "AI solutions are almost right, but not quite right."
  • Security vulnerabilities skyrocketed in 2025.
    • "Of 534 AI code samples, about 25 had confirmed security flaws." (AppSec Santa 2026 study) [32:15]
    • Example: Multbook, an AI-built social network, accidentally exposed 1.5M API tokens.

• The Three Month Black Box Effect

  • After ~90 days, teams and even AIs forget how/why code was built.
  • Leads to "untouchable" architectures; rework becomes expensive (estimated 8,000 startups now in full/partial rebuild).
  • "It's like if a conveyor belt was cranked to 10x speed, but you still have just one quality inspector." – Jordan [33:50]

The Emergence of Rescue Engineering [34:00]

• New tech discipline focused on rebuilding or "rescuing" brittle AI-built architectures.
• Growing market: full rebuild costs range from $50K to several million dollars per company.
• AI sprawl is creating software stacks no one understands or maintains.

How to Survive and Thrive in the Vibe Coding Era [34:45]

• Adopt Governance: Organizations without controls will pay for rescue engineering and data breach costs.
• Treat AI Coding as Engineering: Human oversight and rigorous methods are key.
• Accept permanence: "This technology is not optional. But how your organization governs it is entirely your choice."

Notable Quotes & Memorable Moments

• [06:45] "You describe something in plain English... and AI builds software."
• [10:05] "Most of the people fueling all of this growth are not even software engineers."
• [13:20] "Claude Code... by far it is broadly considered the current leader in agentic coding."
• [28:05] "Their head of finance built a full CRM on Lovable... used to replace [Autonom's] $40,000 Salesforce contract."
• [32:15] "Of 534 AI code samples, about 25 had confirmed security flaws."
• [33:50] "It's like if a conveyor belt was cranked to 10x speed, but you still have just one quality inspector."
• [34:45] "This technology is not optional. But how your organization governs it is entirely your choice."

Important Timestamps

• 00:16 – 03:00: Analogy: Instant AI dream house vs. real engineering risks
• 03:55 – 08:15: Definition and history of vibe coding; explosive growth stats
• 13:20 – 25:10: Breakdown of top 10 vibe coding platforms and categories
• 25:20 – 34:00: Deep dive into opportunities, dangers, and risks (inc. security flaws, collapse of developer trust)
• 34:00 – 34:45: Rise of rescue engineering, 3-month black box effect
• 34:45 – 35:47: The future: rules for survival, importance of governance, closing thoughts

Final Thoughts

Jordan’s tone remains accessible, even as he delves into technical topics, always returning to clear, pragmatic advice for both non-technical and seasoned listeners. The main message: Vibe coding isn’t just a phase—it’s altering how software is built, who builds it, and what kind of oversight and expertise the future will demand.

For anyone touched by AI, tech, or enterprise software, this episode is a compelling guide to understanding and navigating the unpredictable yet exciting landscape of AI-driven coding.

For the full Start Here series and more resources, visit starthere.series.com.