Mar 16, 2026 · 39m · y-combinator
AI Is Unlocking Millions Of New Builders · Y Combinator
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the Lightcone Podcast, Y Combinator partners interview Emergent co-founders Mukund and Madhav Jha about how their AI coding platform reached $100M ARR in eight months by empowering non-technical domain experts to build production-ready software.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The partners as informed peer 5.5 Guest teaching 5.2 Guest disagreement 1.1 The partners pushing back 1.7
speaking balance: gold is the partners, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 38:41 Host challenges 'great for YC' framing
The host politely counters Mukund's comment that massive creation of apps is 'great for YC,' pointing out that many of these hyper-niche lifestyle businesses will never need venture capital.
Hardest push from the partners ▶ 32:53 Direct challenge on model lab platform riskThe host presses directly on the existential risk facing wrappers if frontier model companies like Anthropic build their own user-facing coding applications.
Biggest teaching moment ▶ 11:14 Masterclass on continual learning memory architecturesMadhav educates the hosts on how Emergent aggregates session trajectories through automated CI/CD into cross-session skills memory, referencing the newly emerging SkillsBench paradigm.
The partners hold their own ▶ 14:04 Sharp competitive structural analysisThe host articulates an incisive architectural insight comparing Emergent's top-down developer-grade foundation against competitors trying to add deep engineering capability to lightweight prototyping tools.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The partners as informed peer | Guest teaching | Guest disagreement | The partners pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome & Emergent's Skyrocketing Growth | 4 | 4 | 1 | 1 | The host warmly introduces Mukund and Madhav Jha and references their rapid growth metrics. Mukund outlines their background at Dunzo and explains the origin insight that automated testing was the key bottleneck to general coding agents. | |
| Sweeping SWE-bench and Discovering Multi-Agent Paradigms | 5 | 5 | 1 | 1 | The host probes the 2024 competitive landscape. Mukund details how topping SWE-bench led them to pioneer multi-agent orchestration and test-time compute scaling ahead of published research. | |
| Second-Mover Advantage and Building End-to-End Infrastructure | 6 | 6 | 1 | 2 | The host draws a smart analogy to second-mover advantages in legal AI (Harvey vs Legora). Mukund explains their philosophy of ignoring transient problems that upcoming foundation models will solve naturally. | |
| Scaling Distribution Through Influencers and Positioning | 6 | 7 | 1 | 2 | The host questions whether influencer marketing was paired with specific persona targeting. Madhav provides a deep technical breakdown of their proprietary Kubernetes container sandbox and cross-session continual learning memory system. | |
| Balancing Developer Power with Non-Technical UX | 7 | 5 | 1 | 3 | The host synthesizes Emergent's architectural positioning versus competitors (top-down power simplification vs bottom-up UX adding depth) and presses on the risk of foundation models expanding into application layers. | |
| Jevons Paradox and the Expansion of Software Work | 6 | 4 | 1 | 1 | The hosts and guests engage in a collaborative discussion on Jevons Paradox, agreeing that more capable coding tools expand aggregate software demand and compress multi-person roles into single operators. | |
| Demo Showcase: AV Configurator and Legal AI CRM | 5 | 5 | 1 | 1 | Madhav demonstrates the interface and shows live user applications, including an AV room builder and legal CRM. The host highlights how non-technical UX abstractions like managed API keys streamline builder onboarding. | |
| Dogfooding: Replacing Asana with a Custom In-House Tool | 5 | 5 | 1 | 2 | The host probes why an internal team replaced Asana with an Emergent-built tool and asks about version control mechanics. Madhav explains how QA and PM workflows were tailored while abstracting Git complexity. | |
| Hiring High-Ownership Talent in India and SF | 5 | 5 | 1 | 1 | The co-host asks about their hiring model across Bangalore and SF. Mukund details hiring top IIT rankers and enforcing mandatory customer support shifts for all engineers to maintain user empathy. | |
| SF & Bangalore Hiring and the Shift Toward Agentic SaaS | 6 | 6 | 1 | 3 | The host asks whether personalized AI software kills traditional SaaS and challenges the founders on how they defend against foundation model labs entering the app layer. Madhav explains their focus on custom verification layers. | |
| Unlocking Domain Experts: From Equestrian AI to Solopreneurs | 6 | 5 | 2 | 2 | Mukund shares real-world user examples like an equestrian sports psychologist building an app without coding. The host reframes this as an uplifting societal unlock for non-venture solopreneurs, extending classic startup theses. |
Statements from this episode (27)
Mukund Jha: Solving verification enables full automation of software engineering
“If you can solve for verification, which is essentially, you know, you can solve the testing part you can actually automate all the software engineering. That was sort of our key insight that like, you know, verification is the loop, which sort of keeps agent …”
Mukund Jha: Emergent became #1 globally on SWE-bench within two months
“And we built you know, soda coding agents, which became world number one on SweetBench you know, in two months of time.”
Mukund Jha: Emergent invented multi-agent paradigms before academic papers were published
“There was a time when we sort of invented the multi-agent system. We invented memory. We invented like, how do we do agent to agent communication? How do you scale up test time compute? A lot of those things which like were sort of coming out, like we would di…”
Mukund Jha: Emergent dropped enterprise sales after 2-3 months due to slow cycles
“And we spent like two, three months trying to you know, make our agents work within the enterprise. We found that it was too slow.”
Jha: 80% of Emergent users have zero programming knowledge
“Today, 80% of users who are on the platform are non-technical users with zero programming knowledge.”
Automating software engineering requires replicating full developer workflows
“Our key insight was that to automate all of software engineering, you will have to build a platform that replicates what best engineering team do, like code reviews, automated testing, debugging, deployment, security, hosting.”
Emergent Scaled Early Distribution via TikTok and Instagram Influencer Networks
“We built out a large influencer network and that was our, Initial sort of, you know starting point for us. Like we use TikTok, Instagram, and part of this bunch of influencers to really, really spread the word out. And that sort of, you know, kickstarted the w…”
Jha: Matching build and deploy infrastructure reduces agent deployment errors
“If you give your agents the same in front during the build time and the same in front during the deploy time, Then the sort of like during this like deployment phase, you don't encounter those many problems.”
Emergent agents learn across user sessions using long-term trajectory memory
“We were able to figure out, okay, all the trajectories that we are generating, we can kind of aggregate over time and like sort of build in a long-term memory for the agent, which is very unique in the sense that your agent learns not just from your own sessio…”
Madhav Jha: Emergent developers use their own agent instead of Cloud Code
“Our coding agent is so powerful that we basically internally use it as a replacement for cloud code as developers, right?”
Adding deep developer power after building simple UX is structurally difficult
“And I think fundamentally it's like, unless you start from, you know, a starting point, which sort of solves all of these problems along the line, the whole software development life cycle, it's actually really hard to come from the other side and solve these …”
Mukund Jha: Giving foundation models more autonomy improves performance
“And the more autonomy you're able to give to the models, the better they perform.”
Mukund Jha: Coding is only 20% of shipping software to production
“Our view is that I think the coding aspect is only 20% of the job, right? I think like taking an app to production is like really, really hard.”
Emergent extracts 20% to 30% more performance from models using custom harness
“At least with our harness, we're able to extract 20, 30% more on top of these models.”
Emergent tools enable one engineer to replace a five-person development team
“We also are internally seeing like the role sort of combining. So like a PM, a designer engineer, like a single person is doing, you know, like work of all three together, right? So like we have a PM who's By coding internally things. And recently like we so w…”
Madhav Jha: Users build mobile apps for personal use, web apps for business
“What we have noticed is that a lot of personal apps people use, people build mobile apps, but a lot of business apps, they would go and build a web app, right? So that's generally the trend we are seeing.”
Madhav Jha: Emergent QA engineer built an internal Asana clone
“The only other thing I wanted to show was this is an actual Asana Clone that our team built, like one of our QA engineers built internally.”
Madhav Jha: Emergent saves $3,000–$4,000 monthly replacing Asana internally
“And we are also saving like, you know, like 3004 thousand dollars a month in subscription.”
Emergent runs deployment with two engineers and memory with one
“For example, our deployment, which almost mirrors what our cell would look like is done by two people. Like our memory, like where you have like multiple startups solving for memory. It's just built by one person.”
Mukund Jha: Every Emergent employee must do weekly customer support
“Like one of the things that we do really religiously is everybody talks to a customer once a week, twice a week. Like everyone in the entire company. Everyone in the company. Right. They talk to a customer, everybody does customer support.”
Mukund Jha: SaaS companies must pivot to become agent-first to survive
“One is more and more of these SaaS workflows are going to get consumed by an agent, right? Like, so, like you know, unless your SaaS company pivots into like an agent first company you know, I think that's going to be hard to sort of survive.”
Mukund Jha: Roughly 20% of apps built on Emergent are agentic
“A lot of people are building on emergent today, like roughly 20% of them are actually agentic apps.”
Mukund Jha: Hundreds of AI agents will collaborate on single tasks by year-end
“I think by end of the year you'll have, you know, agents which are running for hours. And like maybe hundreds of agents collaborating on the single task.”
Emergent augments foundation models with fine-tuned verification layers instead of building models
“We don't want to, like, build a Opus 4.5 alternative right away, but we do want to augment it through our custom fine tune verification layers.”
Mukund Jha: Foundation AI models will become completely commoditized
“Most of these models are going to get, get really, really commoditized, like, where all of these models will have similar behaviors they'll have, you know, price, price competitiveness between them”