F

Flapping Airplanes

AI · San Francisco, United States · Founded 2025

Flapping Airplanes is a foundational AI research laboratory focused on solving the data efficiency problem in artificial intelligence. Founded in 2025 by brothers Ben and Asher Spector and their co-founder Aidan Smith, the company launched in January 2026 with a $180 million seed funding round co-led by GV, Sequoia Capital, and Index Ventures. The firm is headquartered in San Francisco and operates on the belief that AI models can be significantly more data-efficient, challenging the prevailing industry approach of scaling models by using massive datasets and computational power. The founders bring a mix of backgrounds to the venture. Ben Spector, aged 25 at the time of founding, previously started Prod, a talent incubator for students, while his older brother Asher Spector, 26, was a Stanford statistics Ph.D. student and former debate champion. Aidan Smith is a Thiel fellow who gained experience at Neuralink for three years while studying at Georgia Tech. Together, they aim to create a "young person's AGI lab" that provides an environment similar to a Ph.D. program, emphasizing research independence and long-term horizons to attract top talent who might otherwise pursue academia or quant firms. Flapping Airplanes' core mission is to develop AI that learns with a fraction of the data consumed by current systems, inspired by the efficiency of human learning. The company's name serves as a metaphor for its willingness to question fundamental assumptions and explore new paradigms in AI development. Instead of focusing on compute-heavy scaling, the lab's strategy is "research-first," prioritizing long-horizon bets on novel algorithms and architectures. Their research areas include data-efficient pretraining, retrieval-augmented systems, synthetic data with verification loops, and modular architectures. Although it does not have a commercial product, the long-term goal is to achieve breakthroughs that could enable applications in fields like robotics, retail, and science. Keywords: AI research, data efficiency, foundational models, artificial general intelligence, AGI lab, machine learning, new learning paradigms, algorithmic efficiency, research-driven AI, long-term research, retrieval-augmented systems, modular architectures, synthetic data, compute-first alternative, biological intelligence, AI breakthroughs