Guy Freeman

The last thing I wrote about building software with a coding agent ended on a question I could not answer with more machinery: is this test still capable of telling me I am wrong?

In April I wrote, on this site, that a paper was being prepared for arXiv: Credence: A Bayesian Decision-Theoretic Framework for LLM Agent Tool Selection.

A passing test suite is the thing every engineer wants to see, and it is worth much less than it looks.

The dominant approach to AI alignment involves training.

Ask an engineer how long an agent should deliberate before it commits, and the answer usually arrives as a number.

In 2011, with Jim Smith, I published “A Bayesian approach to event trees” in Bayesian Analysis.

The previous essay in this series ended on an identity and left most of the bill unpaid.

There is a familiar argument for why Rust is preferable to C++, or TypeScript to JavaScript, and it has nothing to do with syntax or speed.

I built a machine that decides whether to answer.

Your OpenClaw agent makes a small decision on every tool call, and right now it makes most of them badly.

Update (June 2026): a lot has shipped since this post.

In the last post I built a governance layer for a coding agent’s tool calls: a body that hooks the agent’s tool_call event, extracts a few features, and dispatches ask, proceed, or block; and a brain, a Julia daemon that holds a belief and maximises expec…

There is a coding agent — pi, the AI tool I use to write code — and it makes tool calls all day.

This is Part 1 of a series on Bayesian decision-theoretic agents.

“Help me keep that coding agent pup on the straight and narrow, as it likes to be ‘pragmatic’.”

I have about eleven million words of personal documents.

The production question about LLM agents, once you have gotten past whether they work at all, is how much they cost.

The accuracy paradox post reported the headline: a Bayesian agent scoring +129.5 against an LLM agent’s +10.8, despite lower accuracy.

The previous post in this series described what I called Tier 1 of the Credence architecture: a DSL for Bayesian decision agents with three types, four axioms, and a constitution forbidding everything else.

The standard way to evaluate a question-answering system is to measure how often it gets the right answer.

This is Part 3 of a series.

This is Part 2 of a series.

I wrote an essay arguing that decision theory had been quietly abandoned by mainstream AI — not because it stopped working, but because deep learning absorbed all the oxygen.

Someone asked a good question.

The companion essay argued that LLM-based “agents” don’t earn the title.

I’ve spent the last few months building agents that maintain actual beliefs and update them from evidence — first a Bayesian learner that teaches itself which foods are safe, then an evolutionary system that discovers its own cognitive architecture.

In Part 1, I built an agent that learns which foods are safe through Bayesian inference.

I step on a bathroom scale and my body fat percentage gets beamed to a server in Shenzhen.

You’re stranded somewhere unfamiliar with twelve types of food scattered around.

Body composition scales are, on their own terms, genuinely useful devices.

Every GTD app I’ve tried commits one of two sins.

The madpersons behind the retro-chic data analysis tool VisiData have created an alternative to the infamous “Advent of Code”.