Writing · Nº 041065 essays
The writing
Ideas from inside the work. Explore technology, entrepreneurship, and the practice of building things that last.
93 essays found · Page 4 of 8
Clear filtersThesisTrust Is a Feature, Not a SloganTrust is a feature you engineer, not a slogan you print. Here is how to build trust into an AI product with logs, limits, and proof buyers can verify.OperatingHow to Run a Membership Organization ElectionThe hard part of a membership organization election is not the ballot. It is member roll integrity. Freeze a clean roster and most disputes disappear.EngineeringHow to Build Guardrails Into an AI ProductEngineering guardrails as a first-class layer: input validation, output checks, refusal behavior, and audit trails that block damage instead of embarrassment.OperatingHow to Sample AI Outputs for Quality ReviewYou cannot review every AI output, so sample them. How to build a sampling and QA process that catches drift and gives you real numbers to show buyers.EngineeringHow to Write an AI Incident Response Plan Before You Need ItAn AI incident response plan you write after the failure is a postmortem. Here is how to build the runbook that contains the damage while it is happening.OperatingDoes Governance Slow Down AI Shipping? No.The objection that governance slows down AI shipping is backwards. Real governance makes you ship faster by removing the fear that makes teams hesitate.EngineeringContent Versioning: Why It Matters More Than You ThinkContent versioning is not a nice-to-have. Here is why tracking versions of your content prevents stale claims, lost edits, and approvals nobody can prove.EngineeringSnapshot vs On-Chain Voting: Which to UseSnapshot voting is gasless and off-chain; on-chain voting is binding and costly. Here is how to choose between them for governance without fooling yourself.AIWhen Your AI Should Refuse to AnswerKnowing when your AI should abstain instead of answering is a governance feature. How to build confident refusal, and why it beats a confident wrong answer.OperatingApproval Gates in Automation: When to Require a HumanNot every step in an AI workflow should run unattended. Here is how to place approval gates so high-stakes actions pause for a human without killing throughput.OperatingRanked Choice vs Approval Voting for Small ElectionsRanked choice vs approval voting for a board or committee seat. For small multi-candidate elections, approval usually wins on simplicity and fewer disputes.StrategyThe EU AI Act: What You Actually Have to DocumentThe EU AI Act turns governance into paperwork obligations. Here is what a high-risk AI system must document, and why most of it you should keep anyway.