Search
One text can break your phone: Android's 'SMS of Death' explained
A flaw in Android 12 and later lets a specially crafted SMS drop calls and even make a phone unusable. The fix is in October's security patch.
IBM unveils 0.7nm nanostack chip technology as a research milestone
The approach builds transistors upward in three dimensions. The node name is not a literal measurement of every feature on the chip.
Google examines how to measure a coding assistant's usefulness before it writes code
Research connected to Jules focuses on finding the right problem and timing an intervention, rather than treating diagnosis and successful repairs as the same result.
John Jumper leaves Google DeepMind for Anthropic
The Nobel-winning AlphaFold researcher is moving to Anthropic after nearly nine years at Google DeepMind.
iOS 27 code hints a foldable iPhone is close
Foldable-related code was found in the iOS 27 beta, and Bloomberg says the device may be on track for a fall release.
ChatGPT adds Lockdown Mode to protect sensitive data
OpenAI's new setting disables some features so prompt injection attacks can't steal sensitive data.
Microsoft criticized for threatening a security researcher
A researcher known as Nightmare Eclipse published unpatched Defender and BitLocker bugs; Microsoft threatened a criminal investigation and the researcher's accounts were banned.
Claude Opus 4.8 arrives with Dynamic Workflows
Anthropic released Claude Opus 4.8 at the same price, plus a research-preview tool that runs hundreds of parallel subagents.
Harvard study: AI beat two doctors at emergency room diagnoses
OpenAI's o1 gave the exact or near diagnosis in 67% of ER cases at triage, versus 55% and 50% for two physicians, in a study published in Science.
Meta buys humanoid robotics startup Assured Robot Intelligence
Meta acquired Assured Robot Intelligence (ARI), whose team joins Meta Superintelligence Labs to work on humanoid robots.
Anthropic launches Claude Design for creating quick visuals
Type a prompt to build prototypes, slides and one-pagers, then export to PDF, PPTX or Canva.
Adobe patches Acrobat zero-day that hackers exploited for months
Opening a malicious PDF could give attackers full control of a computer. Update Acrobat and Reader now.
Meta debuts Muse Spark, the start of a ground-up AI overhaul
The first model from Meta Superintelligence Labs is live on the web and in the Meta AI app, with a Contemplating mode for hard questions.
Google unveils TurboQuant, shrinking AI working memory by at least 6x
A new compression algorithm cuts the memory AI models use while running. The internet is already calling it "Pied Piper."
Anthropic releases Claude Opus 4.6 with agent teams that work together
Opus 4.6 has a 1 million token context window and a new "agent teams" feature. Claude now also works inside PowerPoint.
OpenAI launches Prism, a free AI workspace for scientists
Prism is a scientific writing editor that works with LaTeX and GPT-5.2, and it is available to anyone with a ChatGPT account.
iPhone has its best year ever in India: 14 million phones
According to Counterpoint, Apple's share of India's smartphone market reached 9% while the overall market stayed flat.
A WhatsApp and Gmail phishing campaign targeted high-profile people across the Middle East
Through a link sent on WhatsApp, attackers stole passwords and 2FA codes from a Lebanese minister, journalists and others. Here's how to protect yourself.
US imposes 25% tariff on Nvidia H200 chips headed to China
Trump signed a proclamation taxing advanced AI chips that pass through the US on their way abroad. AMD's MI325X is also covered.
Anthropic launches Cowork: Claude works on your computer's files, no code needed
Cowork is a simpler take on Claude Code for everyone. You give it a folder and describe the job in plain language.
OpenAI is preparing a screenless device where voice comes first
According to The Information, OpenAI has merged several teams to overhaul its audio models for a device expected in about a year.