Venture Crane

Start Here

Curated reading paths through everything we’ve published. Each article is a field note from running AI agent teams in production - real costs, real failures, real infrastructure decisions.

Where should I start?

If you’re a solo founder running agents and want to know what it actually costs and how to keep them from wasting money, start with what it costs and kill discipline.

If you’re a team adopting AI coding tools and want protocols that prevent chaos, start with multi-agent team protocols and sessions.

If you’re just exploring, read The System first for the operating model, then come back here for the field notes.


What It Costs

The first question everyone asks. Here is the full breakdown - every line item, no rounding, no “contact us for pricing.”


Keeping Agents Reliable

AI agents are optimistic by default. They will churn on unsolvable problems for hours without escalating. These articles cover the operational patterns that prevent that.


Context Management

The hardest problem in multi-agent development: every session starts cold. These articles cover the system that gives agents memory across sessions and machines.


Infrastructure

The physical and logical infrastructure that makes multi-agent development work at the solo-founder scale.


Ship Log and RSS

Short-form field notes published after each working session. Unpolished, timestamped, occasionally surprising. Browse the full ship log.

New articles publish roughly weekly. The ship log updates more frequently. Subscribe via RSS feed or articles-only RSS.


These articles document how we work. For the principles behind it, read The System. For the problems we haven’t solved yet, see Open Problems.

Last updated: Feb 16, 2026