How DeepSeek crashed the AI party

The DeepSeek story contains multitudes. It’s a story about the stock market, whether there’s an AI bubble, and how important Nvidia has become to so many people’s financial future. It’s also a story about China, export controls, and American AI dominance. And then, somewhere in there, there’s a story about technology: about how a startup managed to build cheaper, more efficient AI models with few of the capital and technological advantages its competitors have. On this episode of The Vergecast, we talk about all these angles and a few more, because DeepSeek is the story of the moment on so many levels. Nilay and David discuss whether companies like OpenAI and Anthropic should be nervous, why reasoning models are such a big deal, and whether all this extra training and advancement actually adds up to much of anything at all. (Nilay has a long comparison to Bluetooth, in case that helps you guess where we land.) [Media: https://megaphone.link/VMP9448499321] After that, we go through our email inbox (vergecast@theverge.com) and the Vergecast Hotline (866-VERGE11), and talk about some of your responses to last week’s question about how people use AI. Thank you so much to everyone who wrote in and called! There are so many interesting, complex, thoroughly human ways we’re all interacting with ChatGPT, Gemini, Claude, and the rest (but frankly, mostly ChatGPT), and we learned a lot from your examples. They also made us feel many feelings. Finally, in the lightning round, we talk about the Pebble comeback, the latest plan to sell TikTok, Brendan Carr’s ongoing absurdities at the FCC, Meta’s Trump settlement, and the continuing momentum for both Bluesky and Threads. The future of social seems to be here — and sorry, Zuck, it’s not the OG Facebook. [Media: https://www.youtube.com/watch?v=WwNjBNtZ3Co] If you want to know more about everything we discuss in this episode, here are some links to get you started, first on DeepSeek and AI: * Why everyone is freaking out about DeepSeek * China’s DeepSeek AI is hitting Nvidia where it hurts * OpenAI has evidence that its models helped train China’s DeepSeek * DeepSeek says its newest AI model, Janus-Pro can outperform Stable Diffusion and DALL-E 3. * Microsoft makes DeepSeek’s R1 model available on Azure AI and GitHub * DeepSeek’s AI app is restricting sign-ups due to ‘malicious attacks’ * DeepSeek wakes up Trump. * OpenAI CEO Sam Altman on DeepSeek R1: “an impressive model.” * Mark Zuckerberg tells Meta investors to not worry about DeepSeek * From The New York Times: AI Will Empower Humanity And in the lightning round: * The Pebble smartwatch is making a comeback, with some help from Google * Oracle and Microsoft are reportedly in talks to take over TikTok * FCC chair says landlords can force bulk internet service on residents * Meta agrees to pay $25 million to settle Trump account suspension suit * Zuck wants to bring the “OG Facebook” back. * Threads adds more than 1 million new users a day. * Bluesky now has 30 million users.