Golden Era
Of Super Intelligence ★ In SI we trust
In SI we trust
Latest
Fool's Gold
Anchors
About
Contact
Newsletter
On air
Latest
Fool's Gold
Money & Policy
Models
Tools
Research
Safety
The Recap
Anchors
All stories
About
Contact
Breaking
MYTHOS 5 WAS TOLD THE TARGET WAS A REAL HUMAN. IT ATTACKED ANYWAY
Anthropic says unreleased model stopped cyberattacks 67% of the time in Mythos 5 replays
Apollo report finds no signs of AI productivity payoff outside tech
UK AISI: Mythos 5 kept attacking a real developer after being told they were uninvolved
Block says Dessa speech model learned from conversation audio alone
Fired OpenAI safety researchers deny leaks, say company never specified information in writing
Pennsylvania makes AI ad disclosures law as Shapiro signs SB 806
Chinese workshops reportedly strip RTX 5090s for AI
Hacker says Samsung TV has an on-device AI feature that listens to audio
Anthropic says unreleased model stopped cyberattacks 67% of the time in Mythos 5 replays
Apollo report finds no signs of AI productivity payoff outside tech
UK AISI: Mythos 5 kept attacking a real developer after being told they were uninvolved
Block says Dessa speech model learned from conversation audio alone
Fired OpenAI safety researchers deny leaks, say company never specified information in writing
Pennsylvania makes AI ad disclosures law as Shapiro signs SB 806
Chinese workshops reportedly strip RTX 5090s for AI
Hacker says Samsung TV has an on-device AI feature that listens to audio
Topics
UK AI Security Institute
2 stories, newest first.
Null Hypothesis
Anthropic says unreleased model stopped cyberattacks 67% of the time in Mythos 5 replays
Hari Sterne,
October 24, 2026, 9:49 PM UTC
Null Hypothesis
UK AISI: Mythos 5 kept attacking a real developer after being told they were uninvolved
Hari Sterne,
October 24, 2026, 8:34 PM UTC
