Y Combinator’s CEO says he ships 37,000 lines of AI code per day. A developer looked under the hood
We love a good ol’ social media roast, and Y Combinator CEO Garry Tan found himself on the business end of a doozie Wednesday. Tan, who in a past life worked as an engineering manager at Palantir and has more recently been a vocal proponent for AI acceleration, bragged that he and his AI coding agents have been deploying 37,000 lines of code per day across five separate projects. “Absolutely insane week for agentic engineering,” Tan wrote in an X post on Monday, adding in a follow-up post that he was on a 72-day shipping streak. Absolutely insane week for agentic engineering 37K LOC per day across 5 projects Still speeding up pic.twitter.com/VR3utsduYx — Garry Tan (@garrytan) March 30, 2026 Two days later, a Polish game developer and senior software engineer who goes by the username Gregorein decided to have a closer look at the actual results of all that shipping and took a peek at Tan’s AI-focused blog . “Here’s what 78,400 lines of AI slop code actually looks like in production,” he wrote on X . Gregorein—who has an MSc in Computer Science, a Polish engineering title, and 13 years in the industry—found numerous examples of bloat and inefficiencies in Tan’s site code, and used a single (Anthropic) Claude session to review the files he downloaded from the website to confirm his observations, which are these: Tan/AI built the website so that when a user visits, their browser makes 169 server requests for various assets totaling 6.42 megabytes
Y Combinator’s CEO, Garry Tan, recently found himself in the center of a social media roast after sharing his impressive claim about the AI coding agents he oversees. Tan boasted that his team has been deploying an astounding 37,000 lines of code per day across five separate projects, with a 72-day shipping streak under their belt. His announcement, made on the social networking platform X, sparked curiosity and skepticism among developers and engineers.
Two days after Tan’s announcement, a Polish game developer and senior software engineer with the username Gregorein decided to take a closer look at the actual outcomes of this rapid deployment. Gregorein, who holds an MSc in Computer Science, a Polish engineering title, and has 13 years of experience in the industry, decided to examine Tan’s AI-focused blog. In a series of posts on X, Gregorein shared his findings, revealing numerous instances of bloat and inefficiencies in the website’s code.
To confirm his observations, Gregorein utilized a single session with Anthropic’s Claude, an AI model. He downloaded the files from Tan’s website and analyzed them. The results were striking. When a user visits the site, their browser makes an astonishing 169 server requests for various assets, totaling 6.42 megabytes in size. For comparison, the minimalist Hacker News homepage, also run by Y Combinator, makes just seven requests for data totaling a mere 12 kilobytes.
Furthermore, the website ships 28 actual test files directly to every visitor’s browser. These test files, weighing in at 300 kilobytes, are pure developer scaffolding that users never requested. Additionally, the site loads 78 different JavaScript controllers for features such as AI image generation, voice extraction, and video tools, none of which are visible on the homepage.
Gregorein’s analysis highlights the contrast between Tan’s claims of rapid, efficient AI-driven development and the reality of the website’s performance. While Tan’s team may be generating a large volume of code, it appears that this code is not optimized for user experience or efficiency. The excessive server requests, unnecessary test files, and unused JavaScript controllers suggest that the AI-generated code may be suffering from bloat and inefficiencies, which could hinder the overall performance and usability of the website.
This incident serves as a reminder that while AI-driven development has the potential to revolutionize software engineering, it is crucial to ensure that the resulting code is not only voluminous but also optimized and user-friendly. Developers and engineers must continue to scrutinize AI-generated code to identify and mitigate such inefficiencies, ensuring that the benefits of AI acceleration are realized without compromising on performance or usability.
In the meantime, Tan’s 72-day shipping streak stands in stark contrast to the reality of the website’s codebase. As the debate continues, it will be interesting to see how Tan and his team respond to these criticisms and whether they can address the concerns raised by developers like Gregorein. Only time will tell if the AI-driven development model can truly deliver on its promise of efficiency and effectiveness in the long run.
