DeepSeek: Has a Chinese company opened a new chapter for AI?

It’s not an exaggeration to say that the artificial intelligence revolution has been dominated by American companies. The likes of OpenAI and Nvidia have been fuelled by billions of dollars of investment that was deemed necessary to continue to lead the AI race. But that theory is now being questioned, following the release of the latest large language model from Chinese company DeepSeek. Reportedly built for a fraction of the cost, it seems to have comparable performance to OpenAI’s models. In this week’s episode we’ll discuss: - How DeepSeek, a private company, blindsided many investors and caused a stock market collapse - Whether spending on chip infrastructure will be impacted - The technology behind DeepSeek's latest large language model - Open source vs closed source - Whether this is good news for Chinese chip firms - Whether DeepSeek is a serious threat to the American firms’ dominance - Government guardrails around DeepSeek's chatbot 00:00 - Introduction 00:52 - Welcome 01:54 - What is ASML and why is it relevant? 04:58 - What is DeepSeek and why has it caused such a stir? 09:17 - Why has DeepSeek had such a big impact on markets? 10:36 - Can we trust DeepSeek's numbers? 11:17 - Will DeepSeek impact future spending on AI infrastructure? 14:20 - The tech behind DeepSeek's model 16:29 - Open vs closed source 18:11 - What's the benefit to DeepSeek? 21:07 - Can DeepSeek maintain competitiveness? 23:09 - The reaction from Trump, U.S. companies and investors 29:04 - DeepSeek's limitations