What Actually Makes A Startup Durable
YC partners discuss how founders should navigate the AI era, emphasizing the rapid decrease in AI costs, the evolving definition of a 'hard problem' for startups, the critical importance of human connection and co-founders, and practical advice on building AI-native loops, pivoting, and achieving distribution in a world where software is easy to build.
YC partners shared insights on operating as founders in the AI era, stressing that the cost of intelligence is rapidly decreasing, roughly 10x per year. This makes current AI costs a lesser concern, as the economics will improve quickly. An engineer leveraging AI is now considered ‘1,000x better than an engineer without AI’, suggesting that for most programming tasks, models are more effective than humans.
The Evolving Landscape of Startup Challenges
In an era where software development has become significantly easier—‘10 years ago, writing a lot of software was in and of itself the hard bit. That is not hard anymore’—founders must identify a new ‘hard bit’ for their startups to ensure durability and defensibility. This could involve complex B2B sales, regulatory barriers (e.g., banking licenses), or deep tech challenges like building nuclear reactors or hardware. Pure software products that were previously successful (e.g., Calendarly) are now deemed too easy to replicate, lacking a durable moat. Founders should focus on tackling ‘harder and more ambitious’ problems, which may, in turn, justify significant VC funding needs for capital-intensive ventures like nuclear energy or regulated banks.
VC funding remains relevant, especially for ambitious projects that require substantial capital to accelerate growth. While founders might not need to raise as much money for simple software as five years ago, the ability to leverage capital to move faster and tackle complex problems is still a key advantage.
Building an AI-Native Company
For AI-native companies, the distinction between a ‘real wedge’ and an ‘AI wrapper’ is often seen as a derogatory and meaningless term. YC funds ideas where AI is central to solving previously intractable problems. The focus shifts to the team’s ability to add significant value and the depth of the problem space, recognizing that many successful companies historically started as ‘wrappers’ (e.g., SQL or S3 wrappers).
The practical approach to building an AI-native company involves making all information legible and queryable by an AI agent, then focusing on a single, narrow process to automate or enhance. Examples include automated follow-up emails after sales calls or agents prototyping product ideas during customer interactions. The key is to implement, evaluate, and iteratively improve, starting with the narrowest possible implementation to demonstrate value. A significant mistake is using AI without critical evaluation; instead, founders should aim for a ‘self-learning cycle’ or ‘feedback loop’ where the AI’s skills continuously update.
The Irreplaceable Human Element
Despite AI’s advancements, many aspects of startup building remain uniquely human. YC partners believe that founder well-being, community building, and unique knowledge sharing are areas AI cannot replace. The ‘power of witnessing’—having someone acknowledge the struggles of a founder—is vital in a lonely journey. YC partners, as former founders themselves, offer tailored advice and empathy that AI cannot replicate, emphasizing that ’the bet is always the founders’.
The choice between a solo founder and co-founders is addressed, with YC statistcs showing co-founded companies statistically perform better. While theoretically a one-person billion-dollar company might be possible, the marginal benefits of a co-founder (emotional support, modulation, shared workload) significantly outweigh the costs. The criteria for picking a co-founder emphasize shared values, determination, intelligence, and high integrity over complementary skill sets, as many business skills can be learned, and AI lessens the need for diverse technical backgrounds. “When you’re sad, you’re very happy talking to this person and that person can bring your morale up.”
High agency—the belief that one’s actions produce impact—is considered a muscle that can be built. This involves tackling projects that are challenging but tractable, gradually increasing their difficulty, often with a co-founder.
Startup Strategy and Operations
A critical piece of advice for founders is to ’launch early and often’. The biggest regret of YC founders is not launching soon enough, often due to over-researching or fear of rejection. Startups are an ’extremely empirical exercise’: quickly forming hypotheses, testing them with the real world, and adapting based on data. This contrasts with academic success models where extensive theoretical work is rewarded. To succeed, founders must set ambitious goals, continuously reassess them, and consistently overcome their biggest bottlenecks.
Pivoting should be a strategic decision based on strong evidence that the fundamental hypothesis of the business is wrong, rather than a reaction to a lack of enthusiasm. The most effective pivots are often ‘pulled’ by customers, leading to a genuinely better idea discovered through direct market interaction. The ’last thing’ founders should automate is ’talking to customers,’ as this provides crucial context and keeps the founder focused on what’s important.
Standing Out and Ecosystems
With building becoming easier, differentiation and distribution are more crucial than ever. Founders must constantly innovate in how they gain reach and attention. For early-stage companies, this means offering ‘white glove service’ and deeply understanding individual customer needs, rather than competing with mass-market distribution tactics like billboards used by larger, established companies. For founders in less developed startup ecosystems, joining an existing community (e.g., San Francisco, London, Paris) is recommended to benefit from critical mass and peer support. For students facing budget and visa constraints, YC offers significant credits and encourages them to ‘Go build something cool,’ asserting that the fundamental drive to build remains the most important factor.
Regarding YC’s application process, early decision applicants are treated the same as standard ones, but partners will seek to understand the legitimate reasons for an early application, expecting a high degree of commitment from founders who wish to apply far in advance.