366 posts on growth, marketing, sales, processes, business, SEO, GEO, and AI agents — written from the operator's seat.
How to use cache_control to cut Claude API input costs by up to 90% on agents with large stable prompts — the prefix-match invariant, what to cache, silent invalidators, and the break-even math.
Updated for 2026. First impressions of Claude Fable 5 from someone who runs 30+ production agents — what's actually different, the cost gotcha nobody warns you about, and whether you should switch.
Updated for June 2026. AI agents like Claude Cowork and OpenAI Codex don't just answer questions — they do the work: reading your files, writing code, running tasks end to end. Here's how a beginner gets a real win in an afternoon, no engineering background required.
Updated for June 2026. Anthropic makes money by selling access to its Claude models: usage-based API pricing for developers, Claude subscriptions for consumers (Pro and Max), enterprise seats for teams, Claude Code for engineers, and cloud-marketplace distribution through AWS and Google. Here's how each line works.
Updated for June 2026. OpenAI makes money from ChatGPT subscriptions (Plus, Pro, Team, Enterprise), a usage-based API for developers, large enterprise deals, and its deep partnership with Microsoft. Consumer subscriptions are the biggest single line. Here's how each part of the model works.
Updated for June 2026. SpaceX earns from three engines: commercial and government launch services (Falcon 9 and Falcon Heavy), the Starlink satellite-internet business, and large government contracts (NASA and national security). Here's how the model works — and the truth about a SpaceX or Starlink IPO.
Updated for June 2026. Claude scheduled tasks let you run a prompt or an agent on a recurring schedule — a daily news digest, a weekly report, an overnight code job — without being at the keyboard. Here's how they work, where to set them up, and the patterns that actually pay off.
Updated for 2026. The cost math I use to pick Claude Haiku vs Sonnet per task — token economics, the routing rule, and where cheap models quietly cost more.
Updated for 2026. The exact playbook I use to debug AI agents running live — trace every step, isolate the failing layer, and reproduce non-determinism fast.
Updated for 2026. How I measure AI-search traffic from ChatGPT, Perplexity, and Claude — referrer tracking, the dark-traffic problem, and what to trust.
Updated for 2026. The orchestration patterns I use to run multi-agent systems — durable queues, externalized state, and clean handoffs that survive retries.
Updated for 2026. The eval harness I use to ship AI agents — a graded test set, an LLM judge, and a regression gate so a prompt change can't break prod.