Claude

How AI agents remember things across runs — and why most of them don't, really.

The problem of building AI that reliably does what you actually wanted — not what you literally asked for

The single mechanism behind every model since 2017 — and the one that quietly burns most of the compute.

When 'think step by step' actually earns its keep — and when it's just expensive theater.

When the model judges itself — Anthropic's bet on alignment without exhausting the rater pool.

What the model can see right now — and why the edges matter

Teaching a small model to imitate a big one — and what gets lost in the lesson.

The cheaper, often-as-good RLHF alternative — and why most labs quietly moved to it.

How you measure whether a model is good at the thing you actually care about.

Teaching a model new habits, not new knowledge

The JSON-shaped API that turned chat models into clients of the real world.

Why AI models confidently make things up — and what you can actually do about it

How models 'learn' from examples in the prompt — without changing a single weight.

How users get aligned models to do what they were trained not to do.

Why long conversations are cheaper than they look — and the reason your API bill behaves the way it does.

The open standard that lets AI models talk to your tools without a custom integration per model

How modern models pretend to be huge while doing the work of something smaller.

The documentation labs publish when they release a model — and what they leave out.

What happens when models train on text written by other models — recursively.

What happens when AI models try to coordinate with each other — and the new failure modes that come with it

When AI learned to see, listen, and read — at the same time, in the same head

The craft of talking to a model that will take you exactly as literally as it decides to

The new SQL injection — when input data quietly becomes instructions the model follows.

Why a 70B-parameter model can run on your laptop — and the quality you trade for it.

What changed when models started thinking before they answered.

Why the document your RAG system retrieves first is rarely the document you want.

Humans rate, model learns, weird things happen — the post-training that made models pleasant to talk to.

Why bigger keeps working — and the question of where it stops.

Forcing the model to fill in a shape — and why it's harder than it looks.

Models training on text other models wrote — and why this isn't always bad.

Why 'more creative' is not the same as 'more random' — and the knobs that actually matter.

The first thing every model does to your words — and the thing that quietly limits what it can do.

From random noise to a model that can reason — the actual pipeline

The architecture that changed what AI could do with language — and then everything else

How CLIP and its descendants taught text and images to live in the same coordinate system.