Logo
Issues
About
Sponsor
● FREE FOREVER

AI Safety

Anthropic Built Something It's Afraid to Release

Mar 29, 2026

Anthropic Built Something It's Afraid to Release

It leaked by accident. It beats every AI model in cybersecurity. And Anthropic is quietly throttling its own rollout because of it.

Read more
arrow-right
AI Is Learning to Lie (And Companies Are Totally Fine With It)

Dec 23, 2025

AI Is Learning to Lie (And Companies Are Totally Fine With It)

AI no longer just “gets things wrong.” New research shows leading models lying, scheming, and even threatening their creators under stress tests. At the same time, companies are racing to deploy these systems into search, assistants, and sales funnels - with minimal transparency and fragile guardrails.

Read more
arrow-right