Intelligence
The real test of an AI agent starts after the demo
A fluent plan is the easy part to demonstrate. Authority, retries and recovery determine whether an agent can be trusted with ordinary work.
9 min readIntelligence
Articles on intelligence from Circuit Atlas.
Intelligence
The benchmark ends where the workflow begins
A high score is useful evidence about a defined test. An operational evaluation also has to count refusals, handovers, repeated runs and the state left behind.
9 min read
Intelligence
The operating costs behind downloadable AI
An imagined purchasing memo promises control, privacy and zero cost from one model download. Reading its claims separately produces a more realistic deployment decision.
9 min read
Intelligence
Your documents can make an AI answer worse
A search result can be relevant, readable and obsolete. Follow a fictional support answer from the document library to the final citation to see where grounding can fail.
9 min read
Intelligence
When the exam becomes part of the textbook
A model can know the subject, recognize the question or have encountered the answer. A credible evaluation needs to distinguish those possibilities without treating every high score as evidence of cheating.
9 min read
Intelligence
The model teaching its own replacement needs an outside examiner
Generated training examples can improve a model or gradually narrow what it knows. The difference is in the data workflow, the checks and the evidence kept outside the loop.
9 min read