Latest

DeepSeek-OCR performance is revolutary, however, you could achive similar result of 8x text compression (87% token savings) with 99.5% accuracy using standard LLM models like Gemini 2.5

Can we model how people recover from setbacks? Using Bayesian hierarchical modeling, I explore how resilience and support systems shape our journey from trauma to post-traumatic growth.

How providing perceived control through "panic buttons" can alleviate psychological stress and unlock team potential in high-pressure environments.

A critical look at AGI hype through a simple game that exposes the fundamental limitations of LLMs