OpenAI feared a 'doom loop' for the web

In the ongoing lawsuit with The New York Times, newly public documents show OpenAI's own team was concerned about its training process. They worried that scraping the internet for data would lower the quality of web content, creating a negative cycle.

This doesn't change how you use the app, but it sheds light on the company's early concerns.

A team of security experts used Anthropic's Claude to help them find weaknesses in OpenAI's computer systems. They were able to access employee accounts before responsibly reporting the issues to OpenAI so they could be fixed.

The event shows how AI tools can be used to test and improve security.

FatherPhi tries using ChatGPT to help him 3D print a model of himself, even though he's never done any 3D modeling or printing before.