AI Agents: The Budget is The Design
Recently, a handful of our agents' analyses came back wrong, and the trace didn't lead to errors in the models. It led to the limits around them, most we had set and one we hadn't. Every AI agent lives inside a set of budgets, and tuning them turned out to be some of the most consequential engineering we've done.
Show Your Work: What it Takes to Trust a Machine
A partner is about to lean on an AI-drafted memo to decide whether to put millions into a company. That moment rests on a question the industry has trained us not to ask: how did the machine get here?
AI Model Temperature: What it is and What it means for Judgment
Every AI agent runs at a temperature. Here is what that means, why it matters, and how each Vantage agent gets its own.
What Separates a Real AI Agent From a Chat Bot or Pipeline With a Name
The word "agent" has been stretched to cover almost anything that calls a model twice. The useful test is whether a system holds a position it can be caught being wrong about.
Building the Learning Loop: The Architecture of an Agent That Actually Improves
Most of the argument about AI agents is still about whether they can act. Inside our fleet that question is settled. The harder one is whether an agent gets better, and the answer turns out to be an argument about where a lesson should live.
Why Our Agents Have No Reward Function: Score Everything, Optimize Nothing
We score every claim our agents make, and we never let a score change an agent's behavior without a person in between. That second rule is the most consequential design decision we've made.
The Attention Ledger: Why Building Less Became the Hard Part
For 30 years the engineering budget did the editing. It is gone, and the discipline it enforced went with it. What remains is the one input AI did not make cheaper, which is the attention of the person on the other side of the screen.
The Future of Interactive: Why the Best Judgment Comes From Argument
The most interactive software most professionals have ever used was not built for work. It was a game. The tools we make our most consequential decisions with have stayed inert, and the lineage that fixes that runs through forty years of teaching software to hold a point of view.
Where the Human Stands: Engineering the Boundary Between People and Machines
When implementing AI, every firm claims a human in the loop, but the phrase has been stretched until it means almost nothing. The real question is architectural, and it has four answers.
Intelligence Was Never One Thing: The Paradox That Shaped How We Built Vantage
The most useful idea for deciding what machines should do in your firm is almost 40 years old. Most of the industry is reading it backwards.