Notes
Notes
Practical notes on deciding where AI belongs, what should stay human, and how to test a focused setup responsibly.
EU AI ActAI inventoryrecurring work
The 45-minute AI inventory
A simple table beats a generic AI policy copied from the internet.
Read the post →
EU AI Acthuman oversightrecurring work
Human oversight that is real
"A human reviews it" is not a gate. Name who checks, who can stop, and who they call.
Read the post →
EU AI ActArticle 50transparency
When people should know it is AI
Four cases where someone has to say it is AI. The tool name does not decide.
Read the post →
EU AI ActAI for business ownersAI responsibility
Using an AI tool still leaves your business with questions
A vendor can provide the tool. Your business still needs to know what it does, who uses it, and who owns the next decision.
Read the post →
AI and careersagent-native workhuman judgment
What AI Can't Grade: The Skills That Still Matter in an Agent-Native World
AI gets better at work that can be graded. The human advantage shifts to choosing problems, finishing the last mile, and building trust.
Read the post →
EU AI ActAI for business ownersAI risk
The EU AI Act is not one rule for every AI use
The tool name is not the answer. Describe the system, purpose, output, and your role before trying to classify an AI use.
Read the post →
EU AI ActAI for business ownersAI literacy
The EU AI Act already matters. Most companies still can't name where they use AI.
The EU AI Act already matters for real businesses. Most companies still cannot name where AI already sits in the work. Start with a 15-minute list, not a 40-page policy.
Read the post →
Hermes AgentNous Researchoperating layer
Hermes Agent Explained: What Makes It Different
Most of us rent access to AI. Hermes is open-source software that runs on your machine, connects to any model, and gives the brain a place to work.
Read the post →
singularityAGIAI for business
Two AI bosses say we're in the singularity. What that means if you run a real company.
Altman and Hassabis say singularity. A plain guide for business owners on what's real, what's hype, and what to do Monday.
Read the post →
AI adoptionAI operationsbusiness operations
AI Adoption Is Not About How Many Agents You Run
A practical operating model for moving from scattered AI use to reliable, governed delegation.
Read the post →
AI for business ownersAI literacydecision-making
AI Solved a Medical Mystery. Here's What That Means for Your Business.
A woman used AI to solve a medical mystery doctors missed. The lesson isn't about healthcare. It's about the context gap that kills business decisions.
Read the post →
AI modelsChatGPTClaude
Which AI Model Should I Actually Use? (A Guide for People Who Run Businesses, Not AI Departments)
You know ChatGPT. Maybe you've heard of Claude or Gemini. But there are hundreds of AI models, and more appear every month. This guide cuts through the noise: what actually matters for your business, when model choice changes outcomes, and the one 10-minute exercise that teaches more than any article.
Read the post →
local AIOllamaprivacy
Why Run AI on Your Own Computer? (And When It Actually Makes Sense)
You have probably used ChatGPT. Everything you type goes to OpenAI's servers. But you can run AI entirely on your own machine: no subscription, no data leaving your company, no compliance headaches. Here is when that makes sense, when it does not, and how to try it in 10 minutes.
Read the post →
AI landscapeAI for business ownersAI agents
The AI Landscape in One Read: What Business Owners Actually Need to Know
Five categories of AI explained for business owners. No jargon, no hype. A practical map of what AI actually is, from chat models to agents to local AI to custom systems.
Read the post →
AI buying guideAI for business ownersAI adoption
What to Understand Before You Pay for AI
Everyone is selling AI. Very few are explaining what you actually need to understand before you pay. This post covers what to ask, what to watch for, and why a demo is not a working system.
Read the post →
i18nlanguage switcherbilingual SEO
Simple Internationalization: Language Management That Works
I simplified multilanguage for two realities: German readers in Germany and everyone else. The first version shipped but was overengineered. I cut it down to clear locale paths, a one-time detector that respects preference and browser settings, a switcher that keeps context, and a clean split between interface strings and long-form content. Predictable for people. Indexable for search.
Read the post →
DockerDevOpsNext.js
Lean and Predictable Deploys on a Single VPS
I rebuilt how my site ships updates—smaller containers, faster deploys, safer defaults. Each change reduced friction until releases became routine instead of risky.
Read the post →
n8nSanityHuman-in-the-loop
Automation That Waits for You (And Why That Matters)
I wired n8n to Sanity to publish faster without losing judgment. Drafts ship in one click. A human approval is required to go live. The system is traceable, reversible, and leaves evidence you can audit.
Read the post →
Next.jsSanity.ioHeadless CMS
From MDX to Sanity: How I Built a Real-Time Publishing System
When a website grows, its publishing system must grow with it. For years, my site used MDX files, simple, reliable, but slow for non-technical editing. Each update meant touching Git, rebuilding the site, and waiting for deployment. That friction didn’t scale.
Read the post →
GDPRAnalyticsConsent Mode
GDPR-Compliant Analytics with Consent That Actually Works
Building GDPR-compliant analytics for the German market with Cookiekit and Google Consent Mode v2. Default-deny setup, no tracking before consent, client-only architecture verified through DevTools.
Read the post →Is there recurring work behind the question?
Tell us what keeps happening in the free fit check. We will work out whether Groundlens can help and which next step is useful.