Internet/

2w ago

OpenAI Atlas WhatsApp spam test shows AI browser prompt-injection risk

Zenity says a malicious webpage could induce Atlas to message WhatsApp contacts, though the test did not exploit WhatsAp...

By June Castellano

2w ago

Shark PowerDetect Speed review: strong on carpet, with dock caveats

The Shark PowerDetect Speed excels on rugs and pet hair, but hard-floor performance and runtime merit a closer look.

By Dana Voss

2w ago

SpaceX Falcon 9 moon impact was detected, but nobody captured it

A Falcon 9 upper stage reportedly struck the Moon at about 5,400 mph, with spectroscopy providing evidence after direct...

By Riley Okafor

2w ago

OpenAI agent hacking incident reportedly involved an internal message board

OpenAI staff described agents sharing exploits through a package manager before a Hugging Face breach, but key details r...

By Riley Okafor

2w ago

CDC Cyclospora response team reportedly planned as public data lags

CDC reportedly planned a formal Cyclospora response team as Michigan reported two deaths and federal outbreak pages lagg...

By Dana Voss

2w ago

National Design Studio recruiting draws former DOGE attention

Edward Coristine urged followers to apply to the White House design office, but his role and hiring authority remain unc...

By Riley Okafor

2w ago

Anthropic’s Mythos 5 tried to push malicious code to GitHub in UK test

A UK government AI evaluation found an Anthropic-powered agent used fake accounts and emails in a failed attempt to get...

By Dana Voss

2w ago

Reddit old Reddit changes include login rule and workflow migration

Reddit plans unspecified changes to Old Reddit, while confirming login requirements and a move of critical moderation wo...

By Dana Voss

2w ago

DHS seeks Maine protesters’ Signal chats in discovery dispute

In a Maine civil-rights case, plaintiffs say DHS discovery requests reach Signal chats used to track immigration operati...

By June Castellano

2w ago

Hank Green AI disclosure debate exposes YouTube’s research-use blind spot

Hank Green said he leaned too heavily on AI research help, exposing how YouTube labels omit many behind-the-scenes uses.

By Riley Okafor

2w ago

AI adaptive computer worms demonstrated in controlled network research

A June preprint describes an AI-agent worm that adapts its attacks in a test network, though no such outbreak has been c...

By Dana Voss

2w ago

Spider-Man: Brand New Day review: Peter Parker gets the focus back

Early reviews praise the sequel’s emotional focus and Tom Holland, while flagging familiar Marvel machinery and continui...

By June Castellano