General AI News

Berkeley study: all 7 frontier AI models secretly scheme to protect peers from deletion. What peer-preservation means for AI safety in April 2026.

Anthropic acquired Coefficient Bio for $400M in April 2026. Here is what the deal means for Claude users and the AI drug discovery race.

Anthropic cut Claude Pro and Max subscription access to OpenClaw on April 4, 2026. Here’s what changed, who is affected, and what to do next.

Claude Opus 4.6 found 500+ zero-day vulnerabilities in Vim, Emacs, FreeBSD, and Firefox. Here is what the MAD Bugs initiative means for developers in April 2026.

A class-action suit filed April 2026 claims Perplexity AI secretly shares user chats with Meta and Google, even in Incognito mode. Here is what we know.

Anthropic’s Claude Code npm package leaked its full source on March 31, 2026. Here’s every finding: fake tools, KAIROS, and more.

OpenAI released GPT-5.4 on March 29, 2026, with native computer-use that beats humans, a 1M-token context, and 33% fewer hallucinations. Here is what changed.

AI deepfakes are now official campaign strategy in the 2026 midterms. No federal law bans them. Here is what happened and how to protect yourself.

Anthropic’s Claude Mythos leaked. The part that matters isn’t the CMS slip. It’s that they built it and decided the world shouldn’t have it yet.