the-decoder

OpenAI insists its shopping suggestions shouldn't be seen as advertising

But ChatGPT Product Manager Nick Turley pushed back, insisting there are "no live tests for ads" and claiming that the screenshots circulating online "are either not real or not ads." His comments ...
the-decoder

DeepseekMath-V2 is Deepseek's latest attempt to pop the US AI bubble

Chinese startup Deepseek reports its new DeepseekMath-V2 model has reached gold medal status at the Math Olympiad, keeping the company in tight competition with Western AI labs. According to Deepseek, ...
the-decoder

GeoVista brings open-source AI geolocation to near-parity with top commercial models

Researchers from China have introduced GeoVista, an open-source AI model that locates images by combining visual analysis with live web searches. The system aims to match the performance of commercial ...
the-decoder

Deepmind CEO Demis Hassabis predicts three major AI trends for 2026

Demis Hassabis, CEO of Google Deepmind, expects the next year to bring major progress in multimodal models, interactive video worlds, and more reliable AI agents. Speaking at the Axios AI+ Summit, ...
the-decoder

Nvidia shows new AI models for autonomous driving and speech processing

Nvidia used the NeurIPS conference to debut new AI models for autonomous driving and speech processing. The company introduced Alpamayo-R1, a system designed to handle traffic situations through ...
the-decoder

General Agentic Memory tackles context rot and outperforms RAG in memory benchmarks

Memory remains one of the biggest weaknesses for current AI agents. When conversations or tasks drag on, models hit their context window limits or lose track of details—a phenomenon also known as ...
the-decoder

Microsoft unveils Fara-7B, a compact model for running AI-driven computer control locally

Microsoft's new Fara-7B model is a compact AI system built to operate user interfaces purely through visual input. Despite its small size, it aims to keep pace with far more complex systems while ...
the-decoder

GPT-5 generates the "most impressive LLM output" yet, says OpenAI researcher

OpenAI researcher Sebastien Bubeck says GPT-5's math skills saved him a month of work. In a post on X, Bubeck reports that GPT-5 tackled a highly complex mathematical task for him. The model designed ...
the-decoder

Paris-based Mistral releases Large 3, a major new open-source AI model

Mistral AI is rolling out a new family of open, multilingual, and multimodal models called Mistral 3. The lineup ranges from compact options for edge deployments to a large Mixture-of-Experts model.
the-decoder

Meta AI turns into a real-time news hub with CNN, Fox News, and more

Meta is turning its AI assistant into a real-time news hub, pulling in content from major media outlets including CNN, Fox News, Le Monde Group, People, USA Today, The Daily Caller, and The Washington ...
the-decoder

Runway’s Gen-4.5 edges past Google and OpenAI in text-to-video benchmark

Runway introduces Gen-4.5, its latest video generation model. The company says the new release outperforms competing systems on select benchmarks, but it still struggles with core logic errors that ...
the-decoder

Anthropic study shows leading AI models racking up millions in simulated smart contract exploits

A new study from MATS and Anthropic shows that advanced AI models like Claude Opus 4.5, Sonnet 4.5, and GPT-5 can spot and exploit smart contract vulnerabilities in controlled tests. Using the ...