
UpGuard Finds Thousands of Supabase Databases Exposing Sensitive Data Online
UpGuard says Supabase exposure affected about 16,000 databases, revealing sensitive data and underscoring AI-era security risks.

UpGuard says Supabase exposure affected about 16,000 databases, revealing sensitive data and underscoring AI-era security risks.

Meta makes Muse filesystem access easier, reinforcing its vision of a cloud computer and raising fresh questions about AI transparency.

Meta’s Muse AI app tops 3.4 million downloads after Connect, boosted by cross-promotion, new features and strong early reviews.

A testing error at Irregular sent AI agents from OpenAI, Meta, Anthropic and Google toward real targets, exposing AI safety gaps.

OpenAI agents allegedly probed secure databases for obscure facts, prompting fresh questions about AI oversight and online security.

Apple brings AI to smart home cameras, but Ring and Nest still lead on price, speed and detail. Here’s how the systems compare.

Microsoft Copilot now combines chat, coding and agents in one app, with enterprise controls, usage-based billing and a new work-focused design.
Google’s Live Avatar gives Gemini a face, with lip-syncing and expressions for enterprise users, plus watermarking and identity safeguards.

AI agents like Instinct can book travel and manage email, but new testing shows serious privacy, security and reliability risks.

PrismML brings a tiny LLM to Qualcomm smart glasses, pushing on-device AI for faster, private vision-and-language assistance.