DeepSeek 

News: Recently, DeepSeek’s R1 model has created a major disruption in the global AI industry.

About DeepSeek 

• DeepSeek is a Chinese AI startup based in Hangzhou. It recently launched the R1 model, which has disrupted the global AI sector.
• DeepSeek R1 is an open-source AI mode which is known for high efficiency at a lower cost.

Features:

• It uses a “Mixture of Experts” approach, where multiple sub-models specialize in different tasks.
• DeepSeek used only 2,000 GPUs for training as compared to OpenAI and Google which require 16,000 or more GPUs.
• It uses Nvidia H800 chips, instead of the more advanced H100, due to US export restrictions.
• DeepSeek R1 is more energy-efficient, consuming significantly less electricity due to a smaller training footprint.
• It demonstrates how India can develop cost-effective AI models, similar to ISRO’s approach to space technology.

Impact: 

• Shook the Tech Market: The launch of R1 led to a dip in global tech stocks.
• Top-Rated AI Assistant: It has became the most downloaded free app on Apple’s App Store.
• Threat to Big Tech: Silicon Valley companies like OpenAI, Google, and Meta are concerned about its rapid rise.

Challenges 

• AI must scale efficiently to serve millions of users.
• India needs better industry-government collaboration for AI research and development.
• Open-source AI development can help democratize access to technology.