Written by Simon Jacobi Software engineer passionate about algorithms, Computer Vision, Machine learning, computational geometry and Comp Sci.
The same model can score 13.3% or 38.3% on the same benchmark depending on the harness around it. If your SaaS depends on an LLM, the harness is your product, and your own benchmark is your moat.
The brain is like a muscle: unused, it weakens. As AI takes over more of our thinking, we may need mental gyms the way we needed physical ones.
This article explains curve parameterization, covering sampling methods, differential geometry, curve length approximation, arc-length parameterization, and numerical techniques for uniform traversal, with practical C++ implementations.
This article explores the probabilistic reasoning errors in the Dreyfus Affair, analyzed by Poincaré, Darboux, and Appell in 1904, and revisits their implications using modern mathematical language.
Exploring path following with Kochanek–Bartels splines, this post delves into tension, bias, and continuity controls, their implementation in animation systems, and Hermite interpolation for smooth trajectory and speed control.
Splines, once used in 18th-century shipbuilding, evolved into a key concept in computer graphics, pioneered in the 1950s by Paul de Casteljau and Pierre Bézier for CAD.
