Field notes from
the audit floor.
Essays on crawl budgets, observability, agent design, and the unglamorous plumbing behind every fast, discoverable, resilient site.
55% of service-business websites aren't visible to ChatGPT, Perplexity, or AI Overviews. We checked 4,688.
TetsuKod.ai audited 4,688 service-business websites against a 17-signal Answer Engine Optimization (AEO) scorecard. 55% aren't clearly visible to ChatGPT, Perplexity, or Google AI Overviews. The 45% that are pull roughly 14× the median organic traffic of the 55% that aren't — a gap that shows up in every industry we tested, and at every business size from single-operator shops to large enterprises.
1fpf: Composable Software For Agentic Systems
1fpf — the system that lets a human review AI-generated PRs in minutes without sacrificing accuracy for speed. Ruthless decomposition layer by layer: schema-derived types, one function per file, folder names that narrate control flow.
AIDEMD - Cascading Intent Into Code
Your AI coding agent can write a full-stack feature in minutes. So why does the output keep drifting? It's not a capability problem — it's a context problem. Your agent can read your code. It can't read your mind.
Why Your Website Should Join the Fediverse
The web is shifting away from centralized platforms and toward open, federated communication. By integrating your website with the Fediverse, your content becomes part of a decentralized distribution network powered by open protocols like ActivityPub. Instead of relying on closed algorithms for visibility, your site becomes the canonical source — distributing posts across interoperable platforms while maintaining full ownership, control, and long-term resilience.
Structured Data in the Age of AI Search: Why It’s No Longer Optional
Search is no longer just about ranking for keywords — it’s about being understood by AI systems that summarize, compare, and recommend content directly inside generative interfaces. In this environment, structured data is essential. By explicitly defining entities like your organization, authors, articles, and services using Schema.org markup, you remove ambiguity and give AI-driven search engines machine-readable context. That clarity increases your chances of proper attribution, rich results eligibility, and visibility inside AI-generated answers.
AI Agents and Self-Orchestrating Workflows in Production Systems
Modern AI is moving beyond chat interfaces and into autonomous execution. Agentic systems represent the next evolution: software agents that reason, select tools, and orchestrate multi-step workflows dynamically — without rigid, pre-defined paths. Unlike traditional automation scripts, these systems adapt in real time, spawning parallel tasks, retrying intelligently, and escalating when confidence drops. In this article, we break down how AI agents work, how self-orchestrating workflows differ from static pipelines, and what it takes to implement them safely and reliably in production.
New essays,
every other Tuesday.
No marketing. No funnel. Just field notes from real audits, written by the engineer who ran them.
Subscribe