OpenAI loses a safety employee, while Capcom prepares to work with AI

In today’s roundup: why a former OpenAI employee publicly spoke out about the company’s problems and how AI is changing the work of game studios. Also included: a study of Google AI Overviews and System76’s rules against AI-generated code.

🔥 Hot:

🔹 OpenAI safety employee quits and publicly warns about the risks: David Robinson prepared safety reports for major model releases. He is now criticizing the company in an editorial in The Atlantic. 🔹 Capcom prepares to develop games with AI: At a conference, the company discussed plans to develop gaming tools with AI in mind.

➡️ News:

🔹 Amazon says it no longer uses NDAs in data center discussions: This is how AWS is responding to public distrust of data center construction. 🔹 AI agents for work, family, and travel are appearing in messaging apps: TechCrunch collected notable services that can be used directly in text conversations.

➡️ Useful materials:

🔹 Study: Half of Google AI Overviews’ sources do not appear on the first page of search results: The author examined more than a thousand Russian-language queries: an AI Overview appeared in 93% of cases. 🔹 How to move from one AI agent to an entire AI organization: Based on his experiments, the author explains how to manage multiple autonomous systems and projects.

➡️ Discussions and case studies:

🔹 System76 bans submitting AI-generated code to Pop!_OS and COSMIC: The rule also applies to comments and textual descriptions in changes and bug reports. 🔹 Simon Willison calls for strict spending limits by default: He believes such restrictions are needed almost everywhere.

📝 If you would like to add other news and materials to the list, write in the comments.