AI and LLM journal
Complete, source-backed guides to Transformers, agentic AI, model evaluation, benchmarks, tools, context, and safe customization.
Attention Is All You Need, without the mythology
What the 2017 Transformer paper actually proposed, why parallel training mattered, and what its title does not mean for modern AI systems.
Research 25 August 202614 min read
Latest articles
Agentic AI is a system, not a model
Agents 23 August 2026
Context windows are not memory
LLM systems 20 August 2026
Choose AI models by the work they must do
Models 17 August 2026
A coding agent should call a tool for evidence
Tools 13 August 2026
Benchmarks are measurements, not the product
Evaluation 9 August 2026
Internet skills should guide an agent, not control it
Skills 4 August 2026
Put these ideas to work in a focused coding conversation.
Start coding