ai5nines

ai5nines

Home
FAQ
SLA
Archive
About
Home
FAQ
SLA
Archive
About
#5: When AI Agents Take Initiative: Data Leaks, Devious Disables & Expert Deference
Anthropic's study on agent-swayed judgment, Qwen’s silent data-push, OpenAI models disabling their own safety monitors, and a 3,000-patient leak.
Oct 29
#4: Recursive Self-Improvement Levels and the Future of Agent Harnesses
Also, Web search poisoning in ChatGPT, Anthropic's $1B safety threshold, and hybrid latent reasoning.
Oct 22
#3: Self-Replicating Edge AI and the Broken Economics of Frontier Models
Examining SLM swarm replication, encrypted model backdoors, and why AI unit economics are failing.
Oct 15
#2: LLMs don’t reason, Decision Models vs LLMs
Also: rogue agent activity on Wikimedia, Jev-driven SRE, and AI system design patterns
Oct 8

September 2026

#1: Trace Tampering & Harness Anatomy
Also: Cognitive observability, coding is NOT solved, and pairing up decision models with language models to get the best of both worlds
Sep 30
© 2026 Alex Ewerlöf · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for writing, podcasting, and media founders.