The Chinese chatbot DeepSeek has sent shockwaves through the global tech industry, causing billions in losses for American tech stocks when it unveiled its R1 program.
OpenAI is finalizing the design of its first custom AI chip, according to an exclusive report by Reuters.
On February 7, 2025, OpenAI improved its o3-mini model's reasoning transparency in response to competition from DeepSeek-R1, enhancing user understanding and performance at lower costs.
Users reported being unable to initiate conversations on ChatGPT, encountering errors like: "Unable to start a conversation" or "Something went wrong while generating a response. If the issue persists, contact us at help.openai.com."
Chinese AI startup DeepSeek launched DeepSeek-V3, featuring 671 billion parameters and surpassing competitors like Meta's Llama 3.1. It aims to bridge open and closed-source AI gaps.
AI development shows no signs of slowing down and is instead accelerating, according to Anthropic co-founder Jack Clark.
Capable of converting text into speech in just 75 milliseconds (excluding network and application delays), Flash is one of the fastest voice AI models available today.
OpenAI has launched the o3 model, enhancing AI reasoning and outperforming its predecessor o1. This follows Google's introduction of Gemini 2.0, escalating competition in AI advancements.
OpenAI's December 2024 update for ChatGPT's desktop app includes enhanced app integration and Advanced Voice Mode, improving task management and coding support for various software.
YouTube, in collaboration with the talent agency Creative Artists Agency (CAA), has introduced a new tool called "likeness management technology" to help celebrities detect and manage deepfake content