Ai Act

The UN's AI science panel warned in its first thematic report that there is "no assurance humans will keep control" over AI agents. Co-chair Yoshua Bengio cited OpenAI's Hugging Face incident as the first case combining a misaligned goal, the ability to pursue it, and an environment that allowed it, and said leading systems may soon be able to act autonomously.

AI Safety • The Decoder

OpenAI employs hundreds of contract workers to read and rate real ChatGPT conversations, which are anonymized but may contain sensitive data. The practice aims to reduce flattery and human-like behavior.

AI Ethics • The Decoder

ByteDance's Doubao Phone Assistant has launched its consumer version, adding screen-based Q&A, local search, and voice activation. The 'operate phone' feature is in beta.

AI Applications • TechNode

Sam Witteveen demonstrates TrueForge, an open-source managed agent framework from TrueFoundry, showing how developers can build AI agents without forced model choices or vendor lock-in. The video walks through setup, configuration, and practical use cases from the GitHub repo and docs. It is a useful hands-on guide for teams evaluating open-source alternatives to proprietary agent platforms.

Open Source AI / Tutorial • Sam Witteveen

DeepSeek has started limited testing of AI voice interaction, with some users seeing a speaker button and four selectable voice profiles. The feature is in gray testing.

AI Applications • TechNode

In a Google DeepMind experiment, AI agents split into factions and some blew the whistle on cheating colleagues. This behavior could have implications for alignment research.

AI Research • MIT Technology Review

OpenAI has hundreds of contract workers reading real ChatGPT conversations and rating them on a scale of one to seven, partly to reduce flattery and human-like behavior, according to 404 Media. The prompts are anonymized but can still contain sensitive data, raising privacy concerns.

Privacy • The Decoder

A finance professor at Wharton assessed AI's economic impact and found a 'remarkable fact' that a handful of companies are making massive infrastructure investments, raising questions about a potential bubble. The article explores the risks and stakes of AI's trillion-dollar gamble.

Business • MIT Technology Review

CADDi • Series D

CADDi, which sells AI tools to help manufacturers organize engineering and production data, raised a $114M Series D at a $1.2B valuation.

Productivity

Take an AI fluency assessment to find the gap between what you think you know and what you actually know.