Reading List

The most recent articles from a list of feeds I subscribe to.

Adobe Modifies Your Hosts File for Their Analytics

Thom Holwerda (via Hacker News): If you’re using Windows or macOS and have Adobe Creative Cloud installed, you may want to take a peek at your hosts file. It turns out Adobe adds a bunch of entries into the hosts file, for a very stupid reason. […] If the DNS entry in your hosts file […]

Apple Scraping YouTube for AI Training Data

Joe Rossignol: Three established YouTube channels have sued Apple, alleging that the company violated the U.S. Digital Millennium Copyright Act (DMCA) by unlawfully accessing and scraping millions of copyrighted videos from YouTube to train its AI models. […] Apple “deliberately circumvented” YouTube’s protections against video scraping and “profited substantially” by doing so. Apple’s research papers […]

Perplexity Privacy Lawsuit

Ashley Belanger (via John Gruber): Perplexity’s AI search engine encourages users to go deeper with their prompts by engaging in chat sessions that a lawsuit has alleged are often shared in their entirety with Google and Meta without users’ knowledge or consent. “This happened to every user regardless of whether or not they signed up […]

Apple Granted Stay Over External Purchase Fee

Sarah Perez: Apple is preparing to take its App Store fight with Epic Games back to the Supreme Court. In a new filing, the iPhone maker said it plans to ask the U.S. Supreme Court to review another aspect of this long-running case over App Store fees. In the meantime, Apple sought to pause the […]

Anthropic’s New Claude Mythos Is So Good at Finding and Exploiting Vulnerabilities That They’re Not Releasing It to the Public

Anthropic’s Frontier Red Team:

Earlier today we announced Claude Mythos Preview, a new general-purpose language model. This model performs strongly across the board, but it is strikingly capable at computer security tasks. In response, we have launched Project Glasswing, an effort to use Mythos Preview to help secure the world’s most critical software, and to prepare the industry for the practices we all will need to adopt to keep ahead of cyberattackers.

This blog post provides technical details for researchers and practitioners who want to understand exactly how we have been testing this model, and what we have found over the past month. We hope this will show why we view this as a watershed moment for security, and why we have chosen to begin a coordinated effort to reinforce the world’s cyber defenses.

Our new model is so good, it’s too dangerous to release to the public” is a message that sounds like it could be marketing hype. But it seems like it’s probably true. Examples cited by Anthropic include finding and exploiting a 27-year-old OpenBSD bug (that can crash any device running OpenBSD) and a 16-year-old bug in the widely used FFmpeg media processing library.

See also: Techmeme’s extensive roundup.