Garry Tan, CEO of Y Combinator, one of the most influential startup accelerators in the world, sparked a major debate on social media this week after sharing a striking milestone on X: he and his AI coding agents had been deploying 37,000 lines of code per day across five separate projects, on a 72-day consecutive shipping streak. The post went viral quickly. But two days later, a Polish senior software engineer known as Gregorein decided to take a closer look at the actual results, and what he found was quite revealing: Tan's code was full of bloat, waste, and rookie mistakes, even on the public-facing side of the site. **What does this teach us?** The core of the debate is that while AI coding tools make it easy to pump out lots of code, it is really the quality of the code that matters, not the quantity. Code that goes into production without proper scrutiny and testing can cause obvious functional failures, create security vulnerabilities, or introduce issues that surface later and force engineers to track down and fix the underlying problems. As Gregorein put it: "Right now we are in a moment where AI lets you generate code faster than any human can review it, and the answer from people like Garry seems to be 'so stop reviewing'." **The bigger picture: agentic AI in the startup ecosystem** This episode is not isolated. Tan has been a vocal proponent of agentic AI in the startup world. According to him, about 25% of the current YC batch have 95% of their code written by AI, and companies are reaching up to $10 million in revenue with teams of fewer than 10 people. Yet Tan himself acknowledges that human agency and judgment remain irreplaceable. In his own words, "agency and taste are super, super important and humans are going to be a really irreplaceable piece of that." **The real opportunity for those building with AI** Tan also points out that the biggest mistake founders are making today is piling into the saturated coding agent space, which already dominates nearly 50% of all agentic AI activity. The real opportunity lies in the verticals that have barely been touched: healthcare at 1%, legal at 0.9%, education at 1.8%, where AI agents have enormous transformative potential but almost no penetration yet. **What does this mean for IT and technology professionals?** Agentic AI is real and powerful, but it does not replace architecture, code review, and sound engineering practices. The speed at which code can now be generated has already outpaced the human ability to review it. The challenge now is to build quality processes that match this new pace. The biggest open spaces in AI are not in more tools for developers, but in the sectors that have barely been touched. The question is not whether we will use AI to develop software. It is how we will use it responsibly and with sound judgment. --- Source: Fast Company, "Y Combinator's CEO says he ships 37,000 lines of AI code per day. A developer looked under the hood" https://www.fastcompany.com/91520702/y-combinator-garry-tan-agentic-ai-social-media