the-decoder

Waymo's leaked system prompt reveals a 1,200-line rulebook for its in-car Gemini assistant

Prompt engineers, take note: Jane Manchun Wong has uncovered the system prompt for Waymo's unreleased Gemini AI assistant, a specification over 1,200 lines long buried in the Waymo app's code. The ...
the-decoder

AI in practice

Read full article about: Anthropic's AI kiosk agent bought a PlayStation 5, ordered a live fish, and drove itself to bankruptcy The Wall Street Journal ran its own test of Anthropic's AI kiosk, and ...
the-decoder

Matthias Bastian

Read full article about: Report: OpenAI may embed sponsored content directly into ChatGPT responses OpenAI's advertising plans for ChatGPT are taking shape. According to The Information, employees are ...
the-decoder

Jonathan Kemper

Google’s new A2UI standard gives AI agents the ability to create graphical interfaces on the fly. Instead of just sending text, AIs can now generate forms, buttons, and other UI elements that blend ...
the-decoder

"The winner takes it all!" - The US-China AI race explained

In the race for AI supremacy, there are two possible winners and the EU. An overview and assessment of the status quo. The global race to develop the best AI systems is in full swing. For years, the ...
the-decoder

Sakana AI's Darwin-Gödel Machine evolves by rewriting its own code to boost performance

Sakana AI and the University of British Columbia have developed the Darwin-Gödel Machine (DGM), an AI system that improves itself by rewriting its own code and evolving new agent variants through ...
the-decoder

Sam Altman says OpenAI has enough data to train the next generation of AI

In an interview, OpenAI CEO Sam Altman emphasizes the importance of using high-quality data to train AI models, whether it is human-generated or synthetic. OpenAI is experimenting with generating ...
the-decoder

Chatbots are now rivaling social networks as a core layer of internet infrastructure

Generative AI systems reached around seven billion monthly web visits in September 2025, a 76% increase compared to the previous year, according to Similarweb. ChatGPT remains at the center of the AI ...
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

Large Language Models and the Lost Middle Phenomenon

Researchers at Stanford University, UC Berkeley, and Samaya AI have found that large language models (LLMs) perform best when retrieving information at the beginning or end of an input, similar to the ...
the-decoder

Microsoft CEO Satya Nadella says AI models "are getting commoditized"

Microsoft CEO Satya Nadella stated in a recent podcast that AI models are becoming standardized and interchangeable, shifting the competitive edge to how effectively companies integrate them into ...
the-decoder

AI progress in 2025 will be "even more dramatic," says Anthropic co-founder

Anthropic co-founder Jack Clark anticipates a significant acceleration in AI development by 2025, driven by the combination of traditional model scaling and novel approaches like OpenAI's test-time ...