If you enjoy our newsletter, please consider to be a paid subscriber to help us keep more news and updates coming out.
Notice: New book report is available! Check out this for more information.
Highlights
Governor Newsom signed SB53.Discord partner hit by data breach.Red Hat confirms GitLab breach.Google warns of new extortion campaign.AI Safety Course: Lecture 4.
Deep Dive
Governor Newsom signs SB53 gov.ca
Named Transparency in Frontier AI Act.
Authored by Senator Scott Wiener.
Builds on state AI recommendations.
Balances innovation with public safety.
Establishes first frontier AI safeguards.
Discord partner hit by data breach pcgamer
Hackers tried to extort ransom.
Discord systems not directly accessed.
Customer support data impacted.
Affected users to be emailed soon.
Data at risk: names, emails, IDs.
Includes partial credit card digits.
No passwords or auth data stolen.
Some government IDs compromised.
Partner’s system access now revoked.
Red Hat confirms GitLab breach TheRegister
Attackers stole data from consulting system.
Crimson Collective claims responsibility.
Group boasts of 28,000 repo raid.
Stolen data includes architecture details.
Red Hat says core products unaffected.
Law enforcement and experts notified.
No details on whose data was taken.
Google warns of new extortion campaign TechCrunch
Hackers target executives at major firms.
Claims of data theft from Oracle software.
Emails began around September 29.
Google hasn’t verified hackers’ claims.
Sent from hundreds of compromised accounts.
Linked to Clop ransomware group.
Clop exploits zero-day vulnerabilities.
Group has breached hundreds of companies.
One ransom demand reportedly $50 million.
Extortion targets IT and C-suite leaders.
Hackers abused Oracle password resets.
Oracle declined to comment on incident.
AI Safety Course: Lecture 4 BoazBarak
Model specs define AI behavior.
AI safety beyond “not offensive”.
Models acting = increased risk.
Specs blend principles & policies.
Detailed specs aid training/testing.
Long-term benefit vs. short-term.
AI needs clear, measurable rules.
Notice: Y2 GenAI Safety and Security is on GumRoad and Amazon with paperback.
