Lan Xuezhao is a venture capital investor at Basis Set Ventures who combines a PhD in psychology and master's in statistics with hands-on experience building Dropbox's first data science team from scratch when the company had only a few hundred employees. She was among the earliest institutional backers of flagship AI infrastructure companies like Scale AI, well before the current AI boom, and applies a systematic, experiment-driven approach to help founders treat everything from sales motions to product development as testable hypotheses with clear metrics. 

• Appreciation for high-agency founders and lean, efficient teams. She highlighted a specific instance where a founder offered to pick her up at the airport to pitch in the car because she was short on time, and she frequently praises startups that reach massive revenue figures with very small staff counts.
• Focus on the deeper evolution of artificial intelligence beyond surface-level mimicry. She expresses interest in technology that moves from simple knowledge retention to actual cognitive architecture and memory, viewing this as the necessary next step for LLMs.
• Strong advocacy for San Francisco as the essential hub for building startups. She repeatedly mentions the influx of young founders moving to the area and suggests that physical presence in the city is required to stay in the flow of important industry shifts.
• Investment strategy centered on finding undervalued assets rather than just seeking non-consensus deals. She likens early-stage investing to treasure hunting, looking for opportunities that others miss because the market feels too early or the founder lacks a traditional background.
• Attention to AI infrastructure and operational efficiency. She actively tracks metrics related to cost and latency, highlighting companies that offer solutions for margin expansion in an environment where compute costs burn through cash quickly.