On The Vergecast: AI chips, AI apps, the re-Pebble, and more.

A security report shows that DeepSeek R1 can generate more harmful content than other AI models without any jailbreaks.

Here's how DeepSeek's R1 model stacks up against ChatGPT when it comes to giving financial advice.

All I wanted was a digital rendering of Brock Purdy if he were a foot taller and had 50 more pounds of muscle on him, but ...

The two AI co-workers on my org chart are OpenAI’s ChatGPT and Anthropic’s Claude. Over the past few months, they’ve taken on ...

Executives at leading AI labs say that large language models like those from OpenAI and Big Tech firms risk becoming ...

The ability to see what DeepSeek is "thinking" before it delivers answers reveals some fascinating insights into how the AI ...

DeepSeeks AI delivers responses comparable to ChatGPT, but its China-hosted version shows differences, particularly in ...

DeepSeek has not responded to OpenAI’s accusations. In a technical paper released with its new chatbot, DeepSeek acknowledged ...

DeepSeek created a humorous rap roasting ChatGPT in Hinglish, leading to a playful exchange between the two AIs. Users found ...

DeepSeek: After US Navy, Congressional offices have been warned not to use DeepSeek, an upstart Chinese chatbot that is ...

Alibaba has just launched a new and improved version of its AI model, Qwen 2.5 ...