Language Models
Large language models - the architecture, the capabilities, the limitations, and the foundation everything else is built on.
signals
The Prompt Engineering Ceiling: Why Better Instructions Won't Save You
On frontier models, sophisticated prompting underperforms zero-shot queries. The techniques that made mid-tier models usable are now making frontier models worse.
Features
From Answer to Insight: Why Reasoning Tokens Are a Quiet Revolution in AI
OpenAI's o1 jumped from the 11th to the 83rd percentile on competitive programming. The difference wasn't better data or more parameters; it was reasoning tokens, invisible chains of thought that let models think before they answer.