📖 What Is Andrew Reading?

A focused paper digest on architectures and agents that can recur, learn continuously, and improve themselves.

Curated for Andrew Mendez · checked every two days

Self-improving agents
experience, skills, critics, feedback loops
Looped transformers
recurrent depth, universal and iterative computation
Continual-learning transformers
memory, adapters, parametric attention, lifelong learning
Self-improving transformers
self-training, test-time learning and online adaptation

Papers only. Each full digest targets three papers per track. Official or author-linked GitHub code is included when it can be verified. If a scheduled check finds no genuinely new relevant papers, the repository is left unchanged.

Loading the latest digest…