Flapping Airplanes co-founder Aidan Smith explains why the lab is pursuing data-efficient methodsSample efficiency measures how much useful learning an AI system achieves from a given number of training examples or experiences. and a vertically integrated GPU stackAI compute infrastructure is the hardware, facilities, networks, power, cooling, storage, and software used to train and run AI models.. He contrasts inspiration from biological brains with the practical difficulties of faithfully emulating one.
Smith discusses potential applications in data-scarce domains, model safetyAI safety is the field and practice of reducing harmful failures, misuse, loss of control, and unintended consequences from AI systems. and compute demand, while noting that the company had not announced a model. His claims about performance, economic impact and future capability are forecasts or company claims.
Watch on YouTube
