Vibe Coding

Alessia quit a "perfect" job, walked away from a stalled biotech pivot, and rebuilt her future inside Y Combinator with Vibeflow.After two failed ideas, three YC applications, and zero safety net, Alessia and her co-founder realized the current wave of AI builders was hitting a wall: frontend magic, backend chaos.Vibeflow emerged from that frustration. It’s a visual backend editor that offers deterministic code generation, giving non-technical founders a way to finally understand and control the logic of their apps without relying solely on black-box prompting.We dig into why vibe coding breaks at scale, how visual backends solve reprompt hell, and what the Y Combinator pressure cooker was really like from the inside.Key TopicsThe Backend Gap: Why AI builders fail on integration layers and get stuck in reprompting loops.Visual Control: Moving from LLM hallucinations to deterministic node graphs.The Pivot: Shifting from microbiome deep tech to developer tools in just weeks.The YC Journey: Three attempts to get in, the reality of the program, and building under extreme constraint.Viral Launch: How honesty and "building in public" drove a #1 Product Hunt launch. TakeawaysBackend Reality Check: Non-technical founders need visibility, not black boxes.Determinism Wins: Node graphs that map directly to backend code eliminate hallucinated logic.Prompting Isn’t Enough: Maintenance and scalability break without visual structure.Speed Requires Sacrifice: Quitting the safe job is often the necessary inflection point for growth.The Future of Vibe Coding: Human-in-the-loop control remains essential for production-grade software. LinksVibeFlowAlessia on X ConnectDiscover the latest vibe coding news and resourcesSubscribe to our newsletterJoin our DiscordFollow on X

Will Simon lost a fortune in the FTX collapse, walked away from crypto, and rebuilt using vibe coding. CFO by day, vibe coder by necessity. Will built a full internal CRM for his payroll company using Lovable—no coding background, just frustration with manual processes and limited dev resources. We dig into his chat mode strategy, why he sees AI having bigger impact than crypto, and his plan to automate every manual payment by next year. Key topics Building production tools without developers: from margin calculators to DocuSign clones Strategic prompting: using chat mode to optimize before implementing Crypto's decade vs. AI's revolution The FTX collapse and walking away from trading Takeaways Chat mode is your friend: spend 30-40% of prompts reviewing plans before implementation to prevent the AI from chasing the wrong problem Security matters from day one: even Lovable's built-in review got a B+ from Will's database engineer with minimal tweaks AI creates jobs, not just kills them: new skills for everyone from founders to entry-level, not replacement Information access is the real shift: from encyclopedias to Wikipedia to ChatGPT—each step didn't kill jobs, it changed what skills matter Know when you need real developers: vibe coding gets you 80% there, but production needs cleanup and optimization   Links The Mainspring   Connect Discover the latest vibe coding news and resources Subscribe to our newsletter Join our Discord Follow on X

Adrian Humphrey beat thousands of developers to win $100K at the Bolt Hackathon—here's how he did it. His approach to vibe coding: strategic debugging, knowing when AI isn't ready, and building for real pain points. We dive into how he built Tailored Labs (an AI video editor) using Bolt and Google's Gemini 2.5 Pro, why his 2018 startup was "too forward thinking," and his framework for turning hackathon wins into fundable companies. Key topics Debugging strategies for vibe coding tools that save time and money Building production-ready apps vs. hackathon prototypes Competing with established players like CapCut and Adobe Takeaways Start with a PRD and user stories: map out use cases before building, then write tests Debug strategically: don't let AI iterate blindly—investigate first, then fix Timing matters more than tech: great ideas can fail if the market isn't ready Production is the real bottleneck: vibe coding makes building easy, but deployment and scaling are still hard Solve time, not features: focus on the biggest pain point (time) rather than flashy AI capabilities Links Tailored Labs Follow Adrian Connect Discover the latest vibe coding news and resources Subscribe to our newsletter Join our Discord Follow on X

Tommy Thomas beat 130,000 developers to win $50K—here's his exact playbook. His approach to vibe coding: product strategy, systematic iteration, and knowing when to pivot. We dive into how he built Keyhaven (an API key manager with analytics) after scrapping two other apps, why he starts every project with a PRD, and his framework for picking ideas that actually make money. Plus: running a tech incubator in Kuala Lumpur, the marathon that traumatized him, and why UBI might be inevitable when everyone becomes a builder. Key topics Why vibe coding is easier but not easy The PRD-first approach to building apps Monetization strategy from day one Choosing the right platform (Bolt vs. Lovable vs. others) Takeaways Start with a PRD: let AI help you think through requirements before you build MVP or bust: focus on one feature, add incrementally Would you pay for it? brutal honesty about your own willingness to subscribe Build for distribution: plan your marketing strategy before you write code Prioritize monetization: cost-benefit analysis beats cool features   Links Opzura KeyHaven Follow Tommy   Connect Discover the latest vibe coding news and resources Subscribe to our newsletter Join our Discord Follow on X

Are AI tools making engineers worse—or just exposing who’s coasting? Robert Welch (Chief Innovation Officer at APM Help) shares how his team ships zero-to-one products fast without letting AI erode core skills. We get into the guardrails that keep quality high (tests, CI/CD, enterprise privacy/no data retention), the difference between designing systems vs pasting code, and why AI should augment—not replace—judgment. Plus: protecting yourself from scams, humanoid robots in property ops, and whether AI music can ever rival human craft.   Key topics AI’s trade-off: speed vs. skill erosion Guardrails that keep quality: tests, CI/CD, code review, metrics Security & compliance in AI dev   Takeaways Augment, don’t abdicate: think first, code second Preserve craft: write tests, run reviews Measure quality, not just speed: defects, rollbacks, coverage Protect data and yourself: approved tooling, least-privilege, verify suspicious requests   Links APM Help Follow Robert   Connect Discover the latest vibe coding news and resources Subscribe to our newsletter Join our Discord Follow on X

I talk with Alessio Carrà about the rapidly closing window for entrepreneurs in the AI coding space. He shares how someone turned a simple domain idea into $2K monthly revenue in just days, and explains why the "easy wins" won't last. Alessio walks me through Context Engineering—his solution to the chaos that happens when vibe coding projects get bigger and AI agents start breaking things. We discuss the weird emotional shift of watching AI write your code, Amazon's new Kiro IDE, and whether we're heading toward a future of local AI models running on laptops. Key Topics The closing opportunity window - Why entrepreneurs need to move fast before all the simple ideas get taken Context Engineering explained - Alessio's MCP that brings structure to vibe coding through PRDs, tech blueprints, and task lists Scaling challenges - How AI coding tools break down as codebases grow larger Comparison with existing tools - How Context Engineering differs from Taskmaster and Amazon's new Kiro The emotional side of AI coding - Missing the satisfaction of solving hard problems vs. gaining focus on outcomes Future of AI development - The push toward cheaper, local, open-source models vs. cloud-based solutions Takeaways Easy business ideas are disappearing fast—the window for low-hanging fruit is shrinking Unstructured vibe coding hits a wall on bigger projects; you need process and documentation Context Engineering brings traditional software development practices to AI coding The satisfaction of coding is shifting from solving technical puzzles to achieving business outcomes Local AI models could eventually eliminate per-token costs and internet dependency AI tools are democratizing software creation, but require new approaches to stay effective Links Context Engineering Follow Alessio Namesnag example Connect Discover the latest vibe coding news and resources Subscribe to our newsletter Join our Discord Follow on X   Chapters00:00 The Urgency of Action in Business02:58 Navigating the Challenges of Vibe Coding06:04 Introducing Context Engineering08:54 The Process of Context Engineering11:58 The Impact of Structure on Vibe Coding14:45 Comparing Context Engineering with Other Tools18:00 The Emotional Experience of Building with AI28:08 Navigating the Workflow of Coding with AI31:31 The Shift in Coding Satisfaction34:16 Focusing on Outcomes Over Outputs37:04 The Impact of AI on Cognitive Skills40:16 The Future of Coding Tools46:10 Local vs. Cloud-Based AI Models

In this inaugural episode of Vibe Coding, host Cosmo Scharf speaks with Ash Gobindram, CEO of Kinemeric, about using AI to create "believable humans" in XR training simulations and how vibe coding is transforming his client workflow. Ash reveals how he uses Lovable to prototype interactive experiences for clients in under an hour—from a lending platform interface for a private equity firm to step-by-step watch repair instructions for luxury brand Richemont. He shares the practical reality of vibe coding: when to start fresh versus iterating, why some projects work better than others, and how he built a full travel agency website in just two hours. The conversation explores the striking parallels between today's vibe coding revolution and the early mobile app era that Ash lived through starting in 2009. From Palm Pilots to iPhones to AI-generated code, we discuss how computing is becoming more democratized and personalized—and what that means for the future of software creation. Plus: Why "believable" matters more than being photorealistic in XR, the philosophical concerns about AI replacing human relationships, and Ash's vision for a decentralized computing future where we interact with technology across multiple form factors. Key Topics: Using Lovable for rapid client prototyping (real examples shown) Creating "believable humans" with LLMs in VR training Mobile revolution parallels to vibe coding adoption When to start fresh vs. iterate in vibe coding projects The democratization of software creation Future of personalized, on-demand app creation   Connect Discover the latest vibe coding news and resources Subscribe to our newsletter Join our Discord Follow on X   Chapters 00:00 Introduction to Vibe Coding and Kinemeric03:10 Believability in Spatial Technology05:46 Practical Applications of XR and AI in Training08:58 Vibe Coding in Early Development Stages11:44 Prototyping with Lovable14:45 Iterating on Prototypes and Lessons Learned17:27 The Future of Software Creation20:38 Reflections on the Evolution of Software Development26:43 The Impact of App Design on Culture29:09 The Evolution of Computing and AI34:18 Exploring XR and AI Integration40:23 Human Interaction with AI44:33 Philosophical Concerns of AI Relationships51:18 The Future of Decoupled Computing