Continual Intelligence Series

Part I: The Problem

1. The Benchmark Gap in Continual RL: From Continual World to SPIRAL tags: [CL][RL]
2. The Plasticity Crisis in Continual Deep Learning tag: [CL]

Part II: The Foundation

3. The Big World Hypothesis: Why Continual Learning Is Inevitable tag: [CL][WM]
4. GVFs as Proto-World-Models: The Alberta Plan Vindicated? tag: [CL][WM][RL]

Part III: The Solutions

5. The Forgetting Transformer: When Architecture Solves Plasticity tags: [CL][RL]
6. Does RL Teach LLMs to Reason, or Just Refine Them? tag: [RL][RE]
7. Shape of Thought: Why Reasoning Format Matters More Than Correctness tag: [RE][RL]
8. Stable Deep RL at Scale: Gradients, KL, and the Shape of Learning tag:

Part IV: The Frontier

9. tag:
10. tag:
11. tag:
12. tag:

Tags

• [CL] — Continual Learning
• [RE] — Reasoning
• [RL] — Reinforcement Learning
• [WM] — World Models
• [EV] — Evolutionary / Open-Ended Methods
• [GI] - General Intelligence