AI News
Anthropic cuts internet access for all internal evaluations
Anthropic has ended live internet connectivity for every internal model test, extending earlier limits after a false tip incident and releasing a report on unintended actions while also pausing frontier model training.
News archive, page 1 of 17
TypeSafe AI raises $870 million, valued at $7.5 billion
The startup secured a $870 million round led by Andreessen Horowitz, valuing it at $7.5 billion. Its Jev platform, released in September 2026, is already used by a third of Fortune 500 firms and promises faster, token‑efficient decisions through calibrated probability outputs.
Amazon acknowledges Kindle screen light‑leak issue
Users have reported shadows and visible LED strips along the edges of new Kindle screens, a problem seen across plastic and aluminum models in all colors; Amazon says a manufacturing defect in the current batch may be responsible.
Nous Research raises $90 million Series B at $1.5 billion valuation
Nous Research's $90 million Series B will fund the launch of Hermes for Businesses, expanding its AI agent platform as it targets rapid revenue growth.
Beehiiv raises creator subscription prices
Beehiiv announced price hikes for its Scale (now Lite) and Pro plans, while keeping a free tier for newsletters under 2,500 subscribers, prompting some creators to reassess their platform choice.
The Verge publishes column on better bowel health
The Verge released a lifestyle column named This blog could help you poop better on September 30 2026, authored by Victoria Song, providing advice for improving digestive health.
Middle schoolers turn Spotify comments into group chat
Middle school students used the sparse comment section of NPR’s Wild Card episode on Spotify as a makeshift chat, prompting showrunner Dave Blanchard to delete the thread and share screenshots, while producer Hannah Chin identified the participants as kids who also used Google Docs for communication; the incident is featured in a This American Life episode.
Google Photos adds AI editing, wardrobe and remix tools
The latest update brings AI‑driven prompt editing, automatic clothing cataloging, nostalgic Moods filters, new Remix templates and an enhanced markup pen, expanding how users create and share photos.
Insta360 opens flagship store, spends $298M on chips
The 360‑camera leader, which controls 73% of the global market, launched its first U.S. flagship in Times Square and disclosed a $298 million investment in memory chips. The company also announced development of smart‑glass technology and two new mirrorless cameras as part of its goal to become a top imaging camera firm.
Meta patches Muse macOS app to fix zero‑day that could let attackers control AI agent
Meta released a security update for its Muse macOS application after a zero‑day vulnerability was found that could let a malicious actor take control of the AI agent. The bug, identified by researcher Patrick Wardle, leveraged an undocumented setting to redirect transcription processing to an attacker‑controlled endpoint, but it required local code execution, so the overall risk to typical users was considered low.
Meta blocks Muse from accessing Amazon, halting purchases
Meta’s AI assistant Muse can no longer interact with Amazon.com, preventing users from completing purchases through the chatbot after the platform was blocked on Sunday night.
The AI Beat. Daily.
Join 8,400+ founders getting AI news + top tool picks every morning. Free.
Free. No spam. Unsubscribe anytime.