the-decoder

Project Gumdrop: OpenAI's first AI gadget could send handwritten notes to ChatGPT

OpenAI has reportedly shifted production of its first AI hardware device from Luxshare to Foxconn. The company wants to avoid manufacturing in China, according to Taiwanese business newspaper Economic ...
the-decoder

Moonshot AI closes $500 million Series C to fund Kimi-K3 development and expand computing capacity

Chinese AI startup Moonshot AI, the company behind the Kimi chatbot, has closed a $500 million Series C funding round. The deal values the company at $4.3 billion, according to an exclusive report ...
the-decoder

OpenAI seeks new "Head of Preparedness" for AI risks like cyberattacks and mental health

OpenAI is hiring a Head of Preparedness. The position focuses on safety risks posed by AI models. OpenAI CEO Sam Altman points to the now well-documented effects of AI models on mental health as one ...
the-decoder

Security researchers catch "privacy" browser extensions siphoning AI chats and selling them via a data broker

The extensions intercept conversations with ChatGPT, Claude, Gemini, Copilot, Perplexity, DeepSeek, Grok, and Meta AI. Even when the VPN is switched off, data collection continues in the background.
the-decoder

Scientific AI models trained on different data are learning the same internal picture of matter, study finds

Researchers at MIT have compared 59 scientific AI models trained for different tasks, including chemical formulas, 3D atomic coordinates, and protein sequences. Despite different architectures and ...
the-decoder

Matthias Bastian

Read full article about: OpenAI merges internal teams to fix audio AI accuracy gap ahead of ChatGPT hardware push OpenAI is building its planned ChatGPT hardware around conversation. To make that work ...
the-decoder

AI in practice

Read full article about: Baidu's in-house chip unit Kunlunxin joins wave of Chinese AI firms heading for Hong Kong IPO Baidu's AI chip division, Kunlunxin, has confidentially filed for an IPO in Hong ...
the-decoder

The ARC benchmark's fall marks another casualty of relentless AI optimization

Update from December 25, 2025: Poetiq has raised the bar even higher on the ARC-AGI-2 benchmark. The AI startup achieved a 75 percent accuracy rate on the public test dataset using OpenAI's GPT-5.2 ...
the-decoder

Maximilian Schreiner

Read full article about: OpenAI releases new models for its Realtime API OpenAI has updated its Realtime API with three new model snapshots designed to improve transcription, speech synthesis, and ...
the-decoder

Meta Neuroscientist King: "Some of the concepts like reasoning may need to be re-evaluated"

The neuroscientist Jean-Rémi King leads the Brain & AI team in Meta’s AI division. In an interview with The Decoder, he discusses the connection between AI and neuroscience, the challenges of ...
the-decoder

OpenAI CEO Sam Altman warns users not to trust ChatGPT agent with sensitive or personal data

OpenAI has introduced ChatGPT agent, a new system that can independently carry out complex, multi-step tasks for the first time. CEO Sam Altman advises caution when using the agent for sensitive data ...
the-decoder

Seedream 3.0 beats GPT-4os image generation, says ByteDance

ByteDance has launched Seedream 3.0, a new text-to-image model that can directly generate high-resolution images. In tests, Seedream 3.0 performed particularly well in displaying text in images, ...