Answer extracted from the Invest Like The Best podcast — listen to the full episode below.
When should a venture investor step back from a research-heavy tech bet?
If you cannot confidently rate a technology bet as an 8 or 9 on conviction, you lack the grounded intuition needed to commit. The distinction matters: backing people on pure pedigree is different from backing people on a technology you truly understand.
Sarah illustrates this with her own experience. She met an interesting company alongside her partners and initially considered backing it, but realized she didn't possess enough understanding of the underlying science to reach genuine conviction. Rather than push forward on faith or reputation alone, she stepped back. This moment of clarity—recognizing the boundary between what she knew and what she didn't—became the discipline that governed the decision.
The core principle is simple: as Sarah explains in the episode, most venture situations require intuition on the business logic itself, not just the caliber of the founder. You can back exceptional people on pedigree grounds—that works in some cases. But for research-intensive or science-heavy bets, that wager becomes much riskier without your own deep understanding of the domain.
Sarah — Early-stage investor focused on AI, biology, defense, and robotics. She co-founded or leads a venture fund with partners Mike, Pranav, and Bella, and is known for her deep engagement with frontier researchers and entrepreneurs. Her investment philosophy emphasizes working with extraordinary people while maintaining discipline about the limits of her own expertise.
The framing of conviction on a 1-10 scale becomes a practical checkpoint. As discussed in this podcast, an investor asking themselves "Would I rate this an 8?" forces honesty. If the answer is no—if you're hovering at a 5 or 6 because the founder is impressive but the science feels opaque—then the right move is to pass or defer. The market is moving fast enough that you don't need to take bets you don't deeply understand.
This applies especially in 2026's AI landscape, where the most competitive and uncertain bets are piling up. Investors who can distinguish between "I believe in this person" and "I understand this technology well enough to bet on it" will make fewer costly mistakes. The discipline to walk away—to admit you lack sufficient domain expertise—is not a weakness. It's a signal of clear thinking in an environment where everyone else is running at 90 miles per hour.
See also
What makes Tony Zhao and Cheng Chi's approach to robotics generalization and data collection particularly innovative?
Sarah credits Tony Zhao and Cheng Chi at Sunday Robotics with generating more interesting ideas in robotics over the last four years than most researchers working in the space.
What regulatory and infrastructure barriers are preventing rapid scaling of compute capacity for AI development?
Sarah identifies compute availability and energy supply as critical constraints over the next five to ten years. A hyperscaler infrastructure leader stated that nothing would move the needle for compute capacity at sufficient scale before 2030.
What psychological challenges do frontier AI researchers face when working at scale in compute-intensive environments?
Sarah observes that many researchers feel disempowered by the massive scale and compute requirements of frontier labs. A large contingent believes either that exponential intelligence arrives in one or two years or that nothing matters until it does.