Lead essay
The real cost of a token
What API pricing is actually paying for — hardware, electricity, and a falling curve — and why this changes which apps make sense.
Defragging the
noise
around machine
intelligence.
A slow, cited publication of essays on what is actually happening in AI — and what it means for the people building, buying, and being affected by it. No announcements. No churn. Just careful writing.
Positioning
Defragging.ai is the field guide and assessment platform for organizations and builders reorganizing fragmented AI systems — across hardware memory, model context, agent state, and tool sprawl.
We publish the canonical reference, run the diagnostic, and offer productized engagements for teams that want help executing.
Current issue
- ~3 min read
Evals vs. vibes
Why every serious AI project ends up writing its own evaluation set, and what a good one looks like.
- ~2 min read
Why benchmarks lie
Contamination, saturation, and the incentive to teach to the test — a field guide to reading model launch charts skeptically.
- ~2 min read
Attention, in plain English
The one trick that makes transformers work, explained without matrices — and what changes once you understand it.
- ~2 min read
What is a token, really?
Models don't see words. They see tokens — chunks of bytes carved up by a learned vocabulary. Here is what that means in practice.
- ~6 min read
Cost defragmentation: where AI money is actually wasted
AI bills are not high because tokens are expensive. They are high because the memory is sitting half-empty and the context is not being attended to.
The pillars
Every essay slots into one of a few long-form spines. Start with the foundations, or follow a thread.
