When Y Combinator CEO Garry Tan launched his first startup in the 2000s, he worked such long hours that he took anti-narcoleptic medication just to stay awake. Today, he no longer needs pills to power through his schedule thanks to the help of A.I. Tan, who says he sleeps only about four hours a night, wakes up early to spend the first couple of hours checking in on a team of autonomous A.I. agents. “I can do my full-time job, doing like eight, nine hours of meetings, and get 10,000 lines of code done on three different projects right now,” said Tan while speaking at SXSW on Saturday (March 14).
A.I.’s advanced coding capabilities are also changing how YC evaluates applicants for its prestigious startup incubator batches. With intelligence now “on tap,” as Tan puts it, the accelerator is placing more weight on founders’ taste, agency and product management skills. The next big founder “might be an English major,” he said.
YC is known for incubating wildly successful startups like Airbnb, DoorDash, Coinbase and Instacart. Every four months, it runs 13-week batches of promising companies that receive $500,000 in funding and mentorship from YC partners.
Tan himself has moved through nearly every role in that ecosystem. He first joined YC as a founder in 2008 with his blogging platform Posterous, then became a partner between 2010 and 2015. After leaving to build Initialized Capital, the venture capital firm he co-founded with Reddit co-founder Alexis Ohanian, he returned in 2023 to lead the accelerator as CEO.
His tenure has coincided with a period of rapid change, as A.I. dramatically boosts coding productivity. Tan said he is already seeing the impact inside YC: in some batches, roughly half of startups are now producing 10,000 to 20,000 lines of code per day. Not everyone in tech, however, is leaning in. “I have software engineer friends who are in denial,” Tan said, urging coders to adopt A.I. tools to increase their output.
That shift in tooling is reshaping what YC looks for in founders. “I care less and less what university someone went to, or even where they worked at,” Tan said. The traditionally impressive markers—Harvard or Stanford degrees, stints at Google or Meta—no longer carry the weight they once did. Instead, he is more interested in a candidate’s GitHub repository, a living record of their code, files and documentation.
A.I. is also changing YC’s long-standing preference for multi-founder teams. Co-founders remain ideal, said Tan, because they can support one another through the stress of building a company. But he added that A.I.’s broad capabilities are making strong solo founders more viable, pointing to Peter Steinberger, the solo founder of OpenClaw, who has launched dozens of other projects, as an example. Steinberger recently joined OpenAI to lead its personal agents division.
Because YC is often a bellwether for Silicon Valley, these shifts ripple beyond its own portfolio. The accelerator has backed more than 5,000 startups with a combined valuation of $1 trillion, with more than 5.5 percent going on to become unicorns. Its alumni network includes figures like OpenAI’s Sam Altman (who served as president of YC from 2014 to 2019) and Airbnb’s Brian Chesky.
YC’s footprint may soon expand geographically as well. While the incubator has long been synonymous with its San Francisco headquarters, Tan suggested that new hubs could be on the horizon, naming Austin, Texas and Boston as potential locations. The organization already has a presence in the latter, having hired Ankit Gupta last year as its first full-time partner based in Cambridge, Mass, adjacent to Boston.
