Open Weights, Brainwashing, and Hardware-Level Morality
We need safety designs that assume full access to model weights
AI
OpenAI’s Charter
You may be surprised at their stance on the AGI arms race
DeepSeek Foreshadows Recursive Self-Improvement
No fanfare, no papers—just abrupt capability shifts
Charles Darwin: The Godfather of Reinforcement Learning
Evolution is a population-level reinforcement learner
A Brief History of Artificial Intelligence
A meditation on the culmination of the computer revolution
The Origin of the Attention Mechanism in LLMs
Dzmitry Bahdanau explains how he arrived at the idea