The sample efficiency black hole Current AI advancements are primarily driven by increased data volume and compute power, rather than significant improvements in sample efficiency. While humans are millions of times more sample efficient, AI’s ability to process enormous datasets and scale compute allows for rapid learning, even if inefficiently. The article posits that AI may eventually automate AI research, tackling the sample efficiency problem to achieve human-like intelligence.

• AI progress is largely due to wider data distributions and scaling compute, not improvements in sample efficiency.
• Reinforcement learning (RL) acts as synthetic data generation, requiring vast compute and human expert data.
• Human expert data is task-specific and requires hundreds of experts for each skill.
• Frontier AI models are trained on trillions of tokens, vastly more than humans see in a lifetime.
• Humans are significantly more sample efficient than current AI models across various tasks like driving and learning skills.
• Evolutionary pre-training is an insufficient explanation for AI’s data requirements; lifetime learning in humans is more analogous to parameter training.
• While current scaling laws suggest parameter scaling can reduce data needs, the discrepancy in sample efficiency between humans and AI remains substantial.
• Despite inefficiency, AI can be trained with immense compute and its learned skills amortized over billions of sessions, making it economically viable for automating white-collar work.
• The automation of AI research itself is proposed as a future step, potentially solving the sample efficiency problem.
• The article questions the common understanding of an ‘intelligence explosion’ and suggests a more nuanced view of rapid AI progress starting from LLMs. Continue reading https://www.dwarkesh.com/p/the-sample-efficiency-black-hole