the-decoder

German court deepens the split on AI and copyright with its latest ruling

Just days after a major copyright ruling in the UK, a Munich court took a sharply different path, highlighting how unsettled the rules around AI and copyrighted content still are. Late last year, GEMA ...
the-decoder

Germany's cybersecurity agency issues new guidelines to protect LLMs from persistent threats

The German Federal Office for Information Security (BSI) warns that even top AI providers are struggling to defend against so-called evasion attacks targeting language models. Ad In these attacks, ...
the-decoder

Most LLM benchmarks are flawed, casting doubt on AI progress metrics, study finds

A new international study highlights major problems with large language model (LLM) benchmarks, showing that most current evaluation methods have serious flaws. After reviewing 445 benchmark papers ...
the-decoder

Google's Gemini API now supports file search with custom data

Google has added a File Search Tool to the Gemini API, allowing developers to query their own documents using a vector database. The tool manages storing files, splitting them up, searching for ...
the-decoder

Facing lawsuits, OpenAI rewires ChatGPT for safer teen use

OpenAI published a "Teen Safety Blueprint" to better protect young users from harm. The new framework follows incidents where ChatGPT allegedly failed to help users in mental distress. OpenAI has ...
the-decoder

Google boasts 1.3 quadrillion tokens each month, but the figure is mostly window dressing

According to Google, it processes over 1.3 quadrillion tokens per month with its AI products and interfaces. This new brand was announced by Google CEO Sundar Pichai at a Google Cloud event. Google ...
the-decoder

Google's VaultGemma shows the struggle to balance privacy and performance in AI

Google DeepMind has introduced a new language model called VaultGemma, designed with a focus on privacy. It is the largest open model to date trained from scratch with differential privacy, containing ...
the-decoder

Adobe Photoshop integrates Google's image AI "Nano Banana"

Adobe is bringing Google's new image AI, known as "Nano Banana" (officially Gemini 2.5 Flash Image), to Photoshop as an optional tool. The model is designed for editing existing images with a high ...
the-decoder

Google brings ads to AI-generated answers worldwide

Google is putting ads directly inside its AI-generated search answers, marking a major shift from search middleman to content provider. The move could reshape the digital advertising landscape. Now, ...
the-decoder

OpenAI has reportedly misjudged its cash burn by $80 billion

The main reason for these numbers is the soaring cost of computing power needed to train and run OpenAI's models, along with a major push to build out its own server infrastructure. The company plans ...
the-decoder

AI researcher Andrej Karpathy says he's "bearish on reinforcement learning" for LLM training

Andrej Karpathy, a former Tesla and OpenAI researcher, is part of a growing movement in the AI community calling for a new approach to building large language models (LLMs) and AI systems. On X, ...
the-decoder

Google DeepMind's Gemini wins Mathematical Olympiad gold using only natural language

Google DeepMind has claimed its first gold medal at the International Mathematical Olympiad (IMO) with an "advanced version" of its Gemini model running in Deep Think mode. The system solved five out ...