In-depth articles with more detailed investigations, interesting thought pieces, and more.
Amazon is reportedly buying large quantities of secondhand and rare books, cutting off their bindings and scanning the pages for artificial intelligence development. A tracking device placed by 404 Media in a book from a bulk Biblio order led to an Amazon warehouse complex in Las Vegas, where workers describe an operation dedicated to scanning …
Google’s AI-generated search summaries are intensifying concerns about the reliability and long-term accessibility of online information. Vass Bednar writes for The Walrus that incorrect AI answers, disappearing webpages and weakened archives could undermine the internet’s role as a public record. Google’s AI Overviews place a generated answer between users and the original source. Bednar points …
Anthropic has disclosed that its forthcoming Claude text watermark uses a secret pattern in the model’s word-selection process, rather than hidden characters or metadata. The method can identify text that Claude likely generated or substantially processed, but it needs enough text to produce a reliable result and can be defeated by a complete rewrite. The …
Text watermarks intended to identify AI generated content can be removed with simple editing or paraphrasing, according to an analysis of upcoming European rules. Sean Goedecke writes in his personal blog that this creates a basic challenge for providers preparing for the EU AI Act’s transparency requirements. The Act’s Article 50 is widely expected to …
AI coding agents can now plan experiments, write software, run tests, inspect failures and revise their work with limited human supervision, according to a detailed personal case study by Google DeepMind researcher David P. Reichert. Reichert writes in his Substack that recent tools such as Anthropic’s Claude Code should be understood less as chatbots and …
Raw benchmark rankings can hide the factors that determine whether an AI agent is useful and affordable in day-to-day work: time limits, token budgets and failed attempts. Laurie Voss writes for VentureBeat that Alibaba’s Qwen 3.8-Max illustrates the gap. Alibaba reports strong coding results with time limits of up to five hours, and as much …
YouTuber and author Hank Green says he plans to slow his video output after acknowledging that his growing use of AI chatbots has become unhealthy. Anthony Ha reports for TechCrunch that Green apologized to his audience and said he needs to better control both his production pace and his impulses around large language models. The …
Google’s SynthID watermark can remain detectable after severe image compression, resizing and screenshots, but it cannot reliably distinguish authentic media from the growing volume of unlabeled AI content. Ryan Whitwam writes for Ars Technica that tests of images generated and edited with Google AI found SynthID still detectable after 300 rounds of simulated sharing and …
OpenAI CEO Sam Altman says the company is preparing for a period of faster model progress while trying to avoid a future in which a small group controls the most powerful AI systems. In an interview with Patrick O’Shaughnessy on YouTube, Altman frames OpenAI’s central task as making intelligence abundant, affordable and broadly available. He …
Restaurants are increasingly using handwritten menus, custom embroidery and lo-fi photography to signal craft at a time when AI-generated marketing images are flooding social media. Bettina Makalintal writes for Eater that restaurants are embracing visible handmade details as a response to AI-generated visual content. The shift reflects a wider cultural reaction to synthetic images that …