YC evaluates applicants by reviewing their Claude and Codex transcripts. This reveals a founder's systems thinking, planning ability, and understanding of feature completeness—a more direct signal of building capability than a resume or GitHub profile.
Moats built around the sheer number of integrations or data connectors are vulnerable. AI coding assistants can now replicate this work in a fraction of the time, threatening incumbents like Salesforce and creating opportunities for new, nimbler challengers.
Despite massive ecosystem changes, YC's leadership intentionally keeps the core program consistent. They view the original Paul Graham-designed experience as a "great product" that shouldn't be fundamentally altered, focusing on timeless principles for founders.
Traditional SaaS is obsolete. According to Tan, companies must now adopt an "agentic" approach, using AI to radically compress decision-making and development cycles from months to hours. Those that fail to embrace this new paradigm will be outcompeted.
YC advises founders to avoid market mapping and competitor analysis in the beginning. The sole focus should be on executing "make something people want." Worrying about rivals is a premature distraction from finding the initial glimmer of product-market fit with users.
Inspired by a Steve Jobs quote, YC partner Garry Tan looks for founders who obsess over details others won't see, like a carpenter perfecting the back of a cabinet. This unseen craftsmanship, like a smooth UI scroll, signals deep product taste and commitment.
YC believes AI assistants will enable more founders like Parker Conrad (Rippling)—strong product thinkers who aren't traditional engineers—to build sophisticated applications. This expands the fundable talent pool beyond classic computer science archetypes.
YC doesn't use a rigid framework for pivot decisions. Instead, partners gauge a founder's excitement and energy. If a founder seems unenthusiastic about their idea after a few weeks, it's a stronger signal to pivot than any metric, as passion is a prerequisite for perseverance.
While AI enables startups to reach $1-2M ARR with almost no hires, post-PMF companies are raising larger rounds than ever. Capital is still a weapon for scaling faster, and the surface area for AI products is so large that teams feel constrained even with enhanced productivity.
To scale its batch size without diluting the experience, YC has decentralized. Each partner runs their own "pod" of ~30 companies, effectively operating multiple small, 2008-era YC batches simultaneously. This parallel structure allows them to increase throughput without major operational changes.