Phase 15

Autonomous Systems

Agents that run without human intervention, safely.

01The Shift from Chatbots to Long-Horizon AgentsLearn~45 minutes 02STaR, V-STaR, Quiet-STaR — Self-Taught ReasoningLearn~60 minutes 03AlphaEvolve — Evolutionary Coding AgentsLearn~60 minutes 04Darwin Godel Machine — Open-Ended Self-Modifying AgentsLearn~60 minutes 05AI Scientist v2 — Workshop-Level Autonomous ResearchLearn~60 minutes 06Automated Alignment Research (Anthropic AAR)Learn~60 minutes 07Recursive Self-Improvement — Capability vs AlignmentLearn~60 minutes 08Bounded Self-Improvement DesignsLearn~60 minutes 09The Autonomous Coding Agent Landscape (2026)Learn~45 minutes 10Permission Modes for Autonomous AgentsLearn~45 minutes 11Browser Agents and Long-Horizon Web TasksLearn~45 minutes 12Long-Running Background Agents: Durable ExecutionLearn~60 minutes 13Action Budgets, Iteration Caps, and Cost GovernorsLearn~60 minutes 14Kill Switches, Circuit Breakers, and Canary TokensLearn~60 minutes 15Human-in-the-Loop: Propose-Then-CommitLearn~60 minutes 16Checkpoints and RollbackLearn~60 minutes 17Constitutional AI and Rule OverridesLearn~60 minutes 18Llama Guard and Input/Output ClassificationLearn~45 minutes 19Anthropic Responsible Scaling Policy v3.0Learn~45 minutes 20OpenAI Preparedness Framework and DeepMind Frontier Safety FrameworkLearn~45 minutes 21METR Time Horizons and External Capability EvaluationLearn~60 minutes 22CAIS, CAISI, and Societal-Scale RiskLearn~45 minutes

“Stay hungry, stay foolish.”

— Steve Jobs