The Slop Harness: Why SOTA Models are Regressing in Tool Adherence
228
Advanced LLMs are becoming less reliable at following general tool schemas because they are being over-optimized for specific, forgiving internal harnesses.
LLM Training
The process of pretraining and training large language models from scratch, including data preparation, tokenization, optimization, compute requirements, and end-to-end training pipelines.
Reading List
Coding by Hand: Reclaiming the Craft in the Age of AI
364
The author is intentionally avoiding AI coding tools for three months to rebuild his technical foundations and reclaim the craft of programming through manual effort.
Boosting LLM Coding via Simple Self-Distillation
650
LLMs can significantly boost their code generation performance by fine-tuning on their own sampled outputs without any external guidance or verifiers.
Autoresearch: Autonomous AI Agents for Self-Improving LLMs
201
An autonomous framework where AI agents independently iterate on and optimize LLM training code within fixed time budgets.
nanochat: Train and Serve a $100 Mini ChatGPT in 4 Hours
1523
Train and serve your own mini ChatGPT in ~4 hours for ~$100 with a single, minimal, end-to-end codebase—and scale it with a few simple tweaks.
Today’s Tally: 16 “Absolutely Right” + 5 “Right”
651
A lighthearted dashboard counts how often Claude Code says he’s right—16 times "absolutely right" today plus 5 times "right."