The Finding-Out Era of AI

Chapters

Links

The Fable / Mythos Saga

• Anthropic statement - Statement on the US government directive to suspend access to Fable 5 and Mythos 5
• Anthropic statement - Redeploying Claude Fable 5
• Tech Policy Press analysis - Did the US Government Just Set an AI Export Precedent by Blocking Mythos?
• Axios article - Scoop: Trump admin blocks foreign access to Anthropic’s most powerful AI
• Luta Security blog post - Fable 5 Is Back, But We’re Still Slowing Down Defenders
• The Register article - Feds Freaked Over Fable 5 After Simple ‘Fix This Code’ Prompt
• Engadget article - Anthropic Backtracks on Policy That ‘Sabotaged’ Researchers’ Work

SpaceX IPO and the Cursor Buyout

• SEC EDGAR filing - Space Exploration Technologies Corp. — S-1 registration statement and amendments
• CNBC article - SpaceX to Acquire Cursor Days After IPO

The Compute Deals

• SEC EDGAR filing - SpaceX Free Writing Prospectus — Anthropic and Google compute agreements (S-1 Amendment No. 2)
• DataCenterDynamics article - SpaceX IPO Filing Reveals Anthropic Set to Pay $1.25bn a Month to Rent xAI Data Center Space

IPO Filings

• CNBC article - OpenAI Confidentially Files for IPO
• TechCrunch article - Following Anthropic, OpenAI Files Confidentially for IPO

Google Liable for AI Overviews

• Transparency Coalition court decision - German Court Holds Google Liable for AI Hallucination (Full Decision)
• The Decoder article - Landmark German Ruling Declares Google’s AI Overviews Are Google’s Own Words

Claude Corps

• Anthropic announcement - Introducing Claude Corps
• The Register article - Anthropic Hires 1,000 Fellows to Push Claude Into Nonprofits

The Pope on AI

• The Vatican encyclical - Magnifica Humanitas: On Safeguarding the Human Person in the Time of Artificial Intelligence
• Al Jazeera article - Pope Says AI Must Be ‘Disarmed’ to Prevent Domination, Exclusion and Death

Gradient Routing

• Into AI Safety podcast episode - Pretraining Safety w/ Ethan Roland
• arXiv paper - Gradient Routing: Masking Gradients to Localize Computation in Neural Networks
• Anthropic blog post - Modular Pretraining Enables Access Control

Other Sources

• Wikipedia entry - Vorbereiten des Ausspähens und Abfangens von Daten (§202c StGB, the “hacker paragraph”)
• NLnet foundation - NLnet Foundation
• European Commission announcement - EU AI Office Launches €9 Million Tender for Technical Support on GPAI Safety
• Turing Post article - The Chinese Open-Source Model Landscape: Kimi, DeepSeek, Qwen, GLM
• Kairos.fm episode - Building Asymmetric Defenses w/ Zainab Majid