Latest
The full stream from the desk. Each piece is sourced and reviewed by an editor before it appears here.
Anthropic now sells four current tiers at four price points. What the official pages actually document, what the benchmark tables do and do not let you compare, and how to choose.
Guide7 min read
A single benchmark number hides who ran the test, on what, and what the model had already seen. Seven questions that make a claim checkable.
Guide3 min read
An agent is a model that directs its own tool use. That flexibility is useful and risky. Here is how agents differ from workflows, when to use one, and the controls that matter.
Explainer4 min read
A model you can download is not necessarily a model you are free to use. How the Open Source AI Definition differs from a weights release, with the Llama 4 licence as a worked example.
Explainer3 min read
A model that accepts a very long prompt has not necessarily learned to use all of it. What the research shows, and a test you can run on your own task.
Analysis3 min read
Retrieval-augmented generation and fine-tuning solve different problems. A guide to choosing between them, and to the cheaper option people overlook.
Guide3 min read
MCP is an open standard for connecting AI applications to tools and data. Here is what it defines, what it leaves to you, and where it can go wrong.
Explainer3 min read