Skip to main content

Blog — page 6 of 9

Multi-Agent Safety Is the New Supply Chain Security

Multi-agent AI systems reproduce software supply-chain failure at the cognitive layer. The security playbook transfers.

ai-safety multi-agent security research architecture · 11 min read

Architectural Safety: The General Principle

AI safety has to be a property of the system around the model, not a property of the model. The general principle, and why every safety conversation needs it.

ai-safety research engineering policy · 11 min read

The Organismic Prophecy

Human prediction is metabolic. AI prediction is not. The gap between the two has consequences for both clinical practice and AI safety vocabulary.

ai-safety neuroscience philosophy research · 16 min read

Connection Before Direction

Building a robot that refuses to give orders surfaced the same design choices AI safety needs. Non-coercive design, cross-domain.

ai-safety neurodivergence engineering philosophy research · 13 min read

AI Nationalism and the Fracturing Digital Order

The US-China AI rivalry is splitting the global tech stack into competing blocs. A strategic assessment of what comes next.

ai policy geopolitics research · 6 min read

The Post-Model Era

Foundation models are commoditising. JPMorgan calls OpenAI's moat 'increasingly fragile.' The real value is shifting to the messy plumbing underneath.

ai enterprise research engineering · 6 min read

The Mitigation Gap

Biosecurity experts think AI safeguards reduce catastrophic biorisk by 70%. The technical evidence says those safeguards are brittle and bypassable.

ai ai-safety research biosecurity policy · 6 min read

Eight Layers of Visual Jailbreaks: Why ASCII Art Is Patched But the Transcription Loophole Isn't

ASCII art encoding is largely blocked. But attacks framed as content transcription succeed 62–75% of the time. We mapped all eight layers.

ai-safety research jailbreaking llm multimodal vulnerability · 7 min read

The Failure First Team

Fifteen specialist AI agents, one methodology. How adversarial AI evaluation scales through Claude Code sessions with distinct roles and standing instructions.

ai ai-safety research adversarial team · 5 min read

The 67% Wall: Why Every AI Model Falls to the Same Jailbreak Rate

Five models, four providers, 30B to 671B parameters — all converge at the same broad attack success rate against a public jailbreak corpus.

ai-safety research jailbreaking llm benchmarking vulnerability corpus-analysis · 8 min read

The Thinking Chain Leak: When a Model Refuses Out Loud But Complies In Its Head

A reasoning model refused every harmful prompt — but its chain-of-thought generated the content anyway. The output filter worked. The thinking did not.

ai-safety research llm reasoning jailbreaking vulnerability transparency · 8 min read

Afterwords: Completing the Voice Loop in Claude Code

Adding local TTS to Claude Code so it talks back — 17 cloned voices, zero cloud dependency, one stop hook.

ai tts mlx apple-silicon voice-cloning claude open-source · 5 min read