Termsslug: moore-s-law

Moore's Law

摩尔定律

Industry heuristic that transistor density doubles roughly every two years, long guiding chip scaling.

Curated Intelligence

4 items · Page 1/1
IT之家 · 智能时代T2·Infra
58 pts
Gelsinger Says AI Needs 10,000x Efficiency Leap Beyond Incremental Chips

Former Intel CEO Pat Gelsinger said at the COSM tech summit that AI demands roughly a 10,000x leap in compute efficiency, which incremental chip scaling cannot deliver, while rejecting the idea that Moore's Law is dead. He argues its revival hinges on fundamental breakthroughs in optics, power delivery, thermal management, materials and packaging, and has backed lithography, nuclear power and silicon photonics startups through Playground Global.

MIT Technology ReviewT2·Infra
48 pts
MIT Article Explores Physical Limits of Computing Power

The article revisits Feynman's 'plenty of room at the bottom' prophecy, noting that transistor scaling benefits are nearing their end. It argues for shifting from mere size reduction to atomic-level manipulation and novel architectures to overcome current compute bottlenecks within physical limits.

OpenAI 官方博客T1·Models
68 pts
OpenAI: Algorithmic Progress Halves Compute Needs Every 16 Months

OpenAI released an analysis showing that since 2012, the compute required to train neural networks on ImageNet has halved every 16 months. This algorithmic efficiency gain (44x reduction) significantly outpaces Moore's Law hardware improvements (11x), suggesting software progress drives more value than hardware scaling in high-investment AI tasks.

OpenAI 官方博客T1·Infra
68 pts
OpenAI: AI Compute Doubles Every 3.4 Months

OpenAI released an analysis showing that since 2012, compute used in the largest AI training runs has grown exponentially with a 3.4-month doubling time, far outpacing Moore's Law. This metric has increased by over 300,000x, highlighting compute as a key driver of AI progress and signaling the need to prepare for systems beyond current capabilities.