Learn languages naturally with fresh, real content!

tap to translate recording

Explore By Region

flag Chinese AI startup DeepSeek introduces new method to make large models more efficient, reducing costs and boosting scalability.

flag DeepSeek, a Chinese AI startup, has unveiled a new training method called Manifold-Constrained Hyper-Connections, designed to make large AI models more efficient and scalable while reducing computational and energy costs. flag The technique, detailed in a paper co-authored by founder Liang Wenfeng and published on arXiv, addresses training instability and memory issues in prior models, enabling stable training across 3 billion to 27 billion parameter systems with minimal added compute. flag Building on ByteDance’s earlier work, the approach reflects China’s push for AI innovation despite U.S. semiconductor restrictions. flag The release fuels anticipation for DeepSeek’s next major model, possibly R2, expected around the Spring Festival in February.

7 Articles

Further Reading