Legora intentionally hires people with high learning velocity ("high Y slopes") over deep experience ("high Y intercepts"). In a rapidly evolving AI landscape, this ensures the team can scale their capabilities as exponentially as the company grows.

When building for a specific domain like insurance, the best hiring strategy isn't to find unicorn candidates with both AI and deep industry expertise. Instead, hire top-tier AI talent and top-tier domain experts and have them collaborate closely, sitting them "next to each other" alongside customers.

Delaying key hires to find the "perfect" candidate is a mistake. The best outcomes come from building a strong team around the founder early on, even if it requires calibration later. Waiting for ideal additions doesn't create better companies; early execution talent does.

When building core AI technology, prioritize hiring 'AI-native' recent graduates over seasoned veterans. These individuals often possess a fearless execution mindset and a foundational understanding of new paradigms that is critical for building from the ground up, countering the traditional wisdom of hiring for experience.

Conventional wisdom favors experienced, mid-career hires, but these programs don't scale. In AI, expertise resides with early-career talent. They can deliver immediate impact on short-term government projects, as there are no established "mid-career AI experts" yet.

While technical founders excel at finding an initial AI product wedge, domain-expert founders may be better positioned for long-term success. Their deep industry knowledge provides an intuitive roadmap for the company's "second act": expanding the product, aligning ecosystem incentives, and building defensibility beyond the initial tool.

While speed is a key business strategy, it's insufficient in a market where the technological foundation shifts weekly. The priority for AI startups should be building high talent density. This enables the company to change direction correctly and quickly, avoiding the trap of moving fast towards an obsolete goal.

The ideal founder profile for AI startups is shifting. Previously, deep domain expertise was paramount. Now, the winning archetype is a scrappy, fast-moving team that can keep pace with rapid model development and quickly productize the latest advancements, outpacing slower, more established experts in their respective fields.

Perplexity's talent strategy bypasses the hyper-competitive market for AI researchers who build foundational models. Instead, it focuses on recruiting "AI application engineers" who excel at implementing existing models. This approach allows startups to build valuable products without engaging in the exorbitant salary wars for pre-training specialists.