7 min read
The Case for One Developer and Nine Agents
Our coding agents averaged $5.79 per run-hour. Add a developer's supervision, and there is a financial case for trying agents before expanding the team.
Read post
Agents
Practical posts on turning recurring work into controlled Task Machine workflows with approvals, checks, agents on your own machines, and a history you can read.
8 min read
Running AI Agent Client Work Without Losing the Margin
Real agent-task data shows agencies how model cost, review, approvals, and reusable delivery workflows affect capacity and margin.
Read post
8 min read
The Real Economics of AI Agents for Indie Hackers
A 30-day dataset turns agent cost, founder attention, review cycles, and completed tasks into a practical solo-company budget.
Read post
8 min read
How Small Software Teams Should Operate AI Agent Work
Real task data shows why small teams need shared ownership, focused decision queues, clear checks, and task-level agent costs.
Read post
7 min read
What AI Agent Tasks Actually Cost Developers
Real task data shows why agent cost, active time, review cycles, and developer attention need separate measurements.
Read post
9 min read
What 500 Real AI Tasks Taught Us
Thirty days of real agent work show what AI tasks cost, where human review helps, and which numbers matter when the work must ship.
Read post
5 min read
Can AI 2x Human Productivity?
One person with AI matched a two-person team, worked faster, and crossed expertise boundaries in a large field experiment.
Read post
11 min read
The economics of your next hire are changing
How small software teams can compare a Task Machine workflow with a hire and decide where human ownership still belongs.
Read post
14 min read
The Hard Part of Multi-Agent Systems Is Everything Between the Agents
Anthropic's research shows why agent coordination depends on scoped context, auditable memory, authority, and verification.
Read post
6 min read
Cybernetics for AI Agents: A Practical Introduction
A practical guide to feedback, control, disturbances, and evidence in AI agent operations.
Read post
8 min read
From Feedback to Organizational Learning: How AI Systems Should Improve Their Own Workflows
Safe organizational learning turns execution evidence into versioned, human-approved workflow changes with post-change evaluation.
Read post
7 min read
The Good Regulator: What an AI Agent Must Know Before It Can Act
Reliable agent action depends on a current internal model, versioned policy, scoped memory, and evidence after execution.
Read post
Page 1 of 3
Next