The air gap is a speed bump, not a vault: RAM can whisper
Research shows malware can turn a computer's memory into a radio transmitter. Isolation still matters, but high-value systems need defenders who model physical leaks too.
Research shows malware can turn a computer's memory into a radio transmitter. Isolation still matters, but high-value systems need defenders who model physical leaks too.
Brazil's top appeals court fined lawyers for hiding commands aimed at its AI. Sanctions are a fine first move, but every document a machine reads is now an attack surface.
A former Meta researcher says he deleted his copy of a pirated books dataset as litigation mounted. AI labs should make data provenance auditable by default, not erasable on demand.
Suncatcher's first reply from orbit is a genuine engineering step, but a haiku from one TPU at a time says little about whether AI compute in space can beat Earth.
Hardware pacing offload won't make headlines, but it shows where kernel performance work is heading: away from flashy features and into the plumbing between CPU and NIC.
GPT-5.6 Sol launched a rocket in peaceful Factorio. Impressive, yes, but clean rules and free retries flatter agents; real work needs tests where mistakes cost something.
Putting Alibaba's Qwen inside Siri is a smart way into China, but it shows "global" AI is becoming regional forks, with a different, more controlled assistant on each side of a border.
Malware that asks an LLM what a victim is worth could make attacker triage cheap. The lesson for defenders: assume any infected machine gets appraised, and protect what pays.
A 78-year-old retired professor was arrested for standing in front of a delivery robot. The real story is who gets to claim the sidewalk, and who has to step aside.
Recent ChatGPT takeover research suggests chat histories now hold email-grade secrets, and researchers got in without guessing a password. Treat AI accounts as critical.
A rigged machine can attack the technician who connects to it. Remote-support tools invert trust, so vendors must sandbox sessions and IT teams should treat inbound connections as hostile.
Nobody seems to mind AI everywhere anymore. We think that's fatigue, not acceptance, and it means platforms can no longer lean on viewers as the filter.
Models that smuggle messages through generated images aren't science fiction. Before agents talk to each other at scale, labs should add steganography to their safety checklists.
Cutting a $30B target to about $20B isn't a collapse, but it shows investors no longer pay for ambition alone. Revenue now has to carry the story.
France's gigabyte tax aims at Netflix and Google, but billing by the byte punishes volume, not value, and it will likely land on everyone else too.
Extortionists reportedly threaten indie studios with AI copies of their games. Telling victims to build a community and keep updating isn't protection, and Steam can do more.
Linode's security notices aren't a confirmed breach, but they show how much customers trust hosting vendors by default. Treat the provider as a risk to model.
Friday's Nasdaq sell-off wasn't about AI failing. It was about investors finally asking who gets paid for the spending, and who just keeps spending.
Making staff justify their AI use with jobs at stake treated people as data to audit. Cancelling it a day after the leak shows a privacy brand can't afford that.
Epoch AI says an open-weight model turned unshipped Firefox fixes into working exploits for about $1,000 each. Open source needs to rethink when it shows its fixes.