Flapping Airplanes on the future of AI: ‘We want to try really radically different things’

2026-02-16 · Source: AI News & Artificial Intelligence | TechCrunch · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Entrepreneurship & Start-ups, Emerging Technologies & Innovation · Depth: Intermediate, extended

Summary

Flapping Airplanes, a new AI research lab, has secured $180 million in seed funding to focus on developing more data-efficient AI training methods. Co-founders Ben and Asher Spector, and Aidan Smith, believe current frontier models, which rely on the "sum totality of human knowledge," are inefficient compared to human learning. Their approach is a concentrated bet on the commercial value and scientific viability of data efficiency, aiming to create models that require significantly less data than large language models (LLMs) like Transformers. They draw inspiration from the human brain as an "existence proof" for alternative algorithms, but do not aim to strictly replicate it, instead seeking "flapping airplane" solutions that are different from both biological and current AI paradigms. The lab prioritizes fundamental research and seeks creative, often young and inexperienced, talent to explore radical new architectures and optimizers.

Key takeaway

For AI Scientists and Research Scientists evaluating future AI development paths, Flapping Airplanes' focus on data efficiency suggests a significant shift from current scale-driven paradigms. You should consider exploring alternative architectures and learning algorithms that prioritize data efficiency, as this approach could unlock new capabilities in data-constrained domains like robotics and scientific discovery, potentially leading to more intelligent and commercially viable systems that generalize better with less training data.

Key insights

Data-efficient AI training, inspired by biological learning, offers a path to novel, commercially viable, and less resource-intensive AI systems.

Principles

• Data efficiency is a critical, underexplored AI problem.
• The brain is an existence proof, not a ceiling, for AI algorithms.
• Fundamental research can be cheaper than incremental scaling.

Method

Flapping Airplanes focuses on developing AI architectures and optimizers that drastically reduce data requirements, drawing inspiration from the human brain's learning mechanisms while not being constrained by its biological limitations.

In practice

• Explore neuromorphic AI for data efficiency gains.
• Prioritize fundamental research for potentially cheaper breakthroughs.
• Seek diverse, creative talent for paradigm-shifting AI work.

Topics

• Data-Efficient AI
• Neuromorphic AI
• AI Research
• Foundation Models
• AI Funding

Best for: AI Scientist, Research Scientist, AI Researcher, Investor, Entrepreneur

Related on AIssential

• The next step towards AGI — AI progress is accelerating due to diverse innovations, shifting the focus from raw scale to sample efficiency and data compression.
• Bridging the gap between neuromorphic ionic computing and more efficient AI — The brain's efficient neuromorphic computing inspires AI for better efficiency.
• AI may not need massive training data after all — Brain-inspired AI architectures can achieve human-like activity without massive training data, challenging data-centric development.
• AI Isn’t Hitting a Scaling Wall. It’s Hitting a Measurement Wall. — AI's efficiency gap and "emergent" behaviors stem from a "measurement wall," not a scaling limit, due to evaluation methods collapsing complex internal states.
• Hebbian architecture AI model [R] — A Hebbian AI model demonstrates emergent learning and resilience without backpropagation, using minimal parameters.
• Data scientists: Powering the future of AI and analytics — The data scientist role spans the entire ML lifecycle, facing challenges mitigated by unified platforms and evolving with AI agents.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI News & Artificial Intelligence | TechCrunch.