About Capital & Compute: Coverage & Method
What Capital & Compute covers, how it researches AI infrastructure economics and agent harnesses, who writes it, and the principles behind every post.
Capital & Compute is an independent publication about the seam between two questions every modern technical builder eventually asks: what does this actually cost to do with compute? and where does the money live?
What we cover
- AI infrastructure economics: honest, source-checked pricing comparisons across coding agents, hosted LLM APIs, and locally-run open-weight models, and how those numbers move quarter to quarter.
- Agent harness deep-dives: what tools like Claude Code and other agent systems actually do, where they fail, and what they cost per task.
- Crypto for builders: wallets, custody, stablecoin settlement, and on-chain finance, covered as infrastructure rather than speculation.
- Jurisdictional finance: tax residency, structures, and the compliance plumbing of moving capital across borders.
Who writes it
Capital & Compute is published by a small editorial team under the Capital & Compute brand rather than under individual bylines. We care about the durability of an argument, not the personality behind it. Every post is researched, sourced, and reviewed against our editorial standards, and we disclose how and where AI tools are used in our process.
How we fund the work
The publication is reader-focused and intends to support its work through display advertising and occasional affiliate relationships. Advertising never influences editorial judgement, and any affiliate relationship that affects a recommendation is disclosed inline. See our disclaimer and disclosures for the full policy.
Get in touch
Corrections, questions, and tips are welcome at contact@capitalandcompute.net, or via our contact page. You can also follow new posts over RSS.