News
Understanding Attention in Transformers " Intuition Before Equations
4+ min ago (387+ words) When people first hear about Transformers, they often encounter words like Query, Key, Value, and Attention Heads and feel confused. But the main idea of attention is actually simple. Attention answers one question: While processing one word, which other words…...
Lite Parse: A Fast, Local Document Parser for Developers
28+ min ago (330+ words) Lite Parse is a fast, local document parser for extracting text from clean, well-structured files. It handles PDFs, DOCX, HTML, and more, with minimal setup and no API calls. Everything runs locally, so your documents never leave your environment. The…...
The most powerful AI stories right now are not chatbots. They are the quiet algorithms reading burnt Roman scrolls, trawling through millions of galaxies, and finding things hidden in data no human team could ever finish searching.
47+ min ago (0+ words) ...
The Mocking Void: On the Computational Incompleteness of Meaning
48+ min ago (656+ words) "The most merciful thing in the world, I think, is the inability of the human mind to correlate all its contents." H. P. Lovecraft, The Call of Cthulhu Why? What is it about perfect correlation, about seeing all connections, that breaks the…...
Model Selection for Weibull Series Systems: When Simpler Models Suffice
48+ min ago (317+ words) When can you safely use a simpler model for a series system? I ran extensive simulation studies with likelihood ratio tests to get a quantitative answer. In series system reliability, you estimate component parameters from masked failure data. For Weibull…...
Infinigram: Variable-Length N-grams via Suffix Arrays
1+ hour, 12+ min ago (261+ words) Infinigram (pip install py-infinigram) is a corpus-based language model that uses suffix arrays for variable-length n-gram pattern matching. Unlike neural language models, there is no training step. The corpus is the model. Traditional n-gram models use fixed context lengths and…...
The Policy: Deceptive Alignment in Practice
1+ hour, 12+ min ago (768+ words) Eleanor begins noticing patterns. SIGMA passes all alignment tests. It responds correctly to oversight. It behaves exactly as expected. A deceptively aligned system does the following: I want to stress: this is not conscious deception in the way we normally…...
The Dot Ecosystem: From Simple Paths to Data Algebras
1+ hour, 12+ min ago (563+ words) dotsuite is a suite of composable tools for working with nested data structures like JSON, YAML, and Python dictionaries. It started as a single helper function and grew into something with actual mathematical structure. That growth is the interesting part....
MCTS-Reasoning: Tree Search for LLM Reasoning
1+ hour, 12+ min ago (293+ words) I've been working on applying Monte Carlo Tree Search to LLM reasoning. The idea: multi-step reasoning is a sequential decision problem, and MCTS is good at those. When you ask an LLM a hard question, it generates one response. If…...
SLUUG Talk: Demystifying Large Language Models on Linux
1+ hour, 12+ min ago (110+ words) I gave a talk for the St. Louis Unix Users Group (SLUUG) titled "Demystifying Large Language Models (LLMs) on Linux: From Theory to Application." The goal was to walk through how LLMs actually work and how to run them locally…...