Building A Global AI Startup From India (40 min)
ai-driven-innovation-economy ai-in-everyday-life ai-in-workforce-disruption ai-investment-trends ai-singularity-speculation ai-utopias-vs-dystopias post-work-ai-society transhumanism-ai-enhancement
- Release date: 2026-03-16
- Listen on Spotify: Open episode
- Episode description:
In this episode of The Lightcone, we talk with Mukund and Madhav Jha, the founders of Emergent - an AI platform that lets anyone build and ship production-ready software. In just eight months, users have created more than 7 million apps on Emergent, with the number doubling in just the last 45 days. We discuss how they built one of the most powerful AI coding agents, why they focused on non-technical users and what it's like building in India for a global audience.Apply to Y Combinator: https://www.ycombinator.com/applyChapters:00:00 - Intro01:06 - What Is Emergent?01:18 - Founder Backstory02:09 - From AI Testing to General Coding Agents02:52 - Getting Ahead of the Market04:18 - The Pivot to Non-Technical Users05:22 - Why Second Movers Can Win in AI09:04 - Building for Production, Not Just Prototypes18:21 - Live Demo: Building Apps with Emergent24:40 - How Emergent Hires and Runs a Lean Team29:04 - Is SaaS Dead? The Rise of Personalized Software34:04 - The Future: Niche Apps, Solo Builders and AI Agency
Summary
- 🚀 Explosive Growth: Emergent hit 7M apps in 8 months post-YC, with 80% non-technical users from 190 countries building business tools like CRMs and niche apps.
- 🔧 Production-First Pivot: From testing agents to #1 SWE-Bench coding, Emergent prioritizes full-stack production-readiness with own infra, multi-agents, and memory—beating prototypes.
- 💡 Domain Expert Unlock: Non-coders bypass dev shops, slashing costs 100x and enabling ‘niche-of-niches’ like psychology-equestrian apps, empowering solopreneurs globally.
- 📉 SaaS Disruption: Custom AI clones kill generic SaaS; 20% apps are agentic, signaling shift to personalized, autonomous software over rigid workflows.
- 🌐 Agent Future: Swarms, 24h horizons, and verifiers herald Cambrian software explosion, expanding markets amid Jevons paradox of more tools fueling more ideas.
Insights
How is Emergent empowering non-technical domain experts to build and ship production-ready apps that power real businesses?
Time: 0:00 – 5:09
Category: AI in Everyday Life, AI-Driven Innovation Economy, Post-Work AI SocietyAnswer: 80% of Emergent’s users have zero programming knowledge yet have built 7 million apps across 190 countries, including CRMs for lawyers and niche apps like EquiMind combining psychology and equestrian coaching. This unlocks ideas previously blocked by tech barriers, enabling solopreneurs to bypass dev shops and reduce costs from $500K to $5K. It fosters a Cambrian explosion of niche software tailored to specific needs. (Start at 0:00)
What pivotal insight allowed Emergent to automate full software engineering cycles?
Time: 2:04 – 10:37
Category: AI in Workforce Disruption, AI-Driven Innovation EconomyAnswer: Solving verification and testing enables agents to loop effectively, automating code reviews, debugging, deployment, and security—replicating top engineering teams. Emergent built its own Kubernetes infra for consistent build/deploy environments and rapid feedback, outperforming prototype-focused competitors. This production-readiness drives user retention for serious business apps. (Start at 2:04)
How is second-mover advantage reshaping AI no-code platforms like Emergent?
Time: 5:22 – 8:07
Category: AI-Driven Innovation Economy, AI Investment TrendsAnswer: Emergent learned from prototypes like Lovable/Bolt, reimagining end-to-end platforms with backend/frontend, async jobs, and influencer marketing for distribution. Each model generation (e.g., Opus) resets opportunities, allowing bold reimagination focused on shipping over prototyping. Strong product + viral growth overtook first-movers despite late entry. (Start at 5:22)
Why does Emergent’s long-term memory from user trajectories outperform static agent skills?
Time: 11:45 – 12:42
Category: AI in Workforce DisruptionAnswer: Agents auto-generate skills via CI/CD on past successful trajectories, enabling continual learning across sessions—e.g., improving calendar integrations over time. This compounds performance without manual intervention, as benchmarks like Skills Bench show human-curated skills beat agent-generated ones, but Emergent automates it effectively. It handles complex apps without context window limits. (Start at 11:45)
Is traditional SaaS facing extinction from customizable AI agentic software?
Time: 29:08 – 30:31
Category: AI in Workforce Disruption, Post-Work AI Society, AI-Driven Innovation EconomyAnswer: Users build clones like Asana/Jira internally, saving costs, and 20% of Emergent apps embed agents for workflows—e.g., lawyer CRMs automating processes. SaaS must pivot to agent-first or risk obsolescence as non-devs create hyper-personalized tools. Software is morphing into agentic, niche systems reducing one-size-fits-all reliance. (Start at 29:08)
What future does agent swarms and 24-hour horizons promise for software creation?
Time: 30:47 – 32:25
Category: AI Singularity Speculation, Transhumanism & AI EnhancementAnswer: With models like Opus enabling long tasks, Emergent experiments with swarms, overseer agents, and fine-tuned verifiers for derailment-proof autonomy. Horizons jumped from 4 to 10 hours; by year-end, hundreds of agents could collaborate. This scales ambitions, powering complex, agentic apps beyond human-led dev. (Start at 30:47)
How are AI tools accelerating a solopreneur revolution and niche-of-niches innovation?
Time: 34:11 – 38:56
Category: Post-Work AI Society, AI in Everyday Life, AI Utopias vs. DystopiasAnswer: Domain experts build intersectional apps (e.g., psychology + horse riding) impossible in limited SaaS worlds, enabling autonomy without VCs/CTOs. This counters job-loss fears by expanding markets—non-devs become creators, teams shrink (1 person = PM/designer/engineer), fostering global micro-businesses. YC-like startup wave extends to everyone via AI. (Start at 34:11)