Posted inBlog
What It Takes to Know If Your AI Agent Actually Works
For a while, the standard way to check whether an AI agent was working was to run it a few times, watch the outputs, and decide it looked reasonable. That…
Posted inBlog
Agent Activity Is Not Agent Work
For the first wave of AI agent deployments, the question organisations asked was whether agents could execute a task at all. That bar has been cleared widely enough that the…
Posted inBlog
The Floor Is the Frontier: Why Agent Reliability Beats Benchmark Chasing
For the first couple of years of serious agent development, most of the energy went into the ceiling: what's the most impressive thing this agent can do? Can it beat…
Posted inBlog
The Performance Curve Inverts Exactly Where You Need It Most
Shaping the flow of ideas at the frontier of cognitive systems. --- A single sentence from a family member can shift an AI's medical recommendation by a factor of twelve.…
Posted inBlog
The Five Questions I Ask Before Every Agent Launch
A few months ago, the habit of reading every agent announcement as soon as it dropped was abandoned. That sounds like neglect, but it was the opposite — the goal…
Posted inBlog
That Single Sentence Is Skewing Your Agent’s Decisions
A few years ago, the conventional wisdom held that the biggest risk with AI agents was that they'd give a wrong answer because they lacked information. Feed them better data…
Posted inBlog
The Floor Test: Three Questions I Ask Before Evaluating Any AI Agent
For a long time, the first question when evaluating a new AI tool was about the ceiling: what's the most impressive thing it can do? The benchmark, the demo, the…