DeepSeek Shakes Global Markets with AI Model

Chinese company DeepSeek launched its AI model, R1, causing over $1 trillion in market losses. It’s now speeding up the release of its next model, R2, planned for an earlier launch than May.

R2 Brings Advanced Features

R2 aims to improve coding and support multiple languages. DeepSeek hopes it will challenge big tech companies by offering cheaper, efficient AI solutions.

Founder Liang Wenfeng’s Leadership Style

Liang, founder of DeepSeek, avoids strict management styles common in China's tech industry. His team works fewer hours in a collaborative environment. DeepSeek is also known for paying employees well.

Success with Low-Cost AI

DeepSeek focuses on cost-effective AI technology, using innovative techniques like Mixture-of-Experts and Multihead Latent Attention. This makes their models cheaper but still competitive with costly US models.

Gaining Chinese Government Support

DeepSeek has earned China's support, with its models being adopted by major companies and government projects. Chinese leaders see DeepSeek’s success as a way to compete with the US in AI development.