AIpocalypse.Now
Today's doom 3.5
models · · By

Anthropic's Claude Fable Refuses High School Biology Questions

Newest Claude model deflects basic biology queries despite touted expertise in field.

ARWF dimensions
Existential criticality
Does the threat involve irreversible systemic failure?
Probability vectoring
Theoretical, or active proof of concept?
Timeline imminence
How close to current deployment?
Mitigation gap
Identified solution, or currently unaligned?
Read original at The Verge More models

Related stories

models 3 / 10

Microsoft Scolds Anthropic for Implying Claude Has Consciousness

Suleyman criticizes Anthropic for speculating about consciousness in Claude's constitutional instructions.

The Verge 1d ago
models 5 / 10

Meta Now Manufactures Its Own AI Clickbait, Completes The Circle

Meta AI app generates clickbait articles in For You feed section.

The Verge 4d ago
models 2 / 10

Apple Reintroduces Siri, Hopes Third Time Proves Charming

New AI-powered Siri arrives at WWDC after years of competitive disadvantage.

The Verge 4d ago
models 3 / 10

Google's Gemini Spark Misses the Obvious Relationship

AI agent reviews emails and calendar but fails to identify user's most important person.

Wired 12d ago
models 1 / 10

Microsoft Copilot Redesigned; Now Twice as Fast at Answers

Microsoft redesigned 365 Copilot with faster loading and more structured responses for easier scanning.

The Verge 13d ago
models 6 / 10

Google's Gemini Turns Stuffed Deer Into Deepfake Vacation Star

Multimodal AI model generates convincing deepfakes; journalist recreates company's own advertisement.

The Verge 18d ago