Skip to content

Blog

What we learn,
written down.

Craft notes on websites, software, AI automation and marketing. What works and what does not.

  1. ArchitectureOne source of truth: how the site builds itself from the catalogWe had our services stored in the database and written out again on the website. Every change had to be made twice, and one side was always stale.3 min
  2. Context engineeringAgent memory: the problem is not remembering, it is forgettingAn agent that remembers everything is slow, expensive and confused. Designing what gets discarded turned out to be harder than designing what gets kept.3 min
  3. FundamentalsThe fundamentals AI does not replace (and that now matter more)Idempotence, migrations, soft deletes, failing fast. Decades-old concepts that turned critical exactly when the one writing the code is not always human.3 min
  4. Best practicesWriting rules for agents: every line of our AGENTS.md cost us an incidentA rules file does not get written in one sitting. It gets written after each thing that breaks, and that is exactly why it works.3 min
  5. ArchitectureWhy our agents never touch the databaseGiving an agent direct database access looks faster. It is, right up until the day a creative agent runs an UPDATE without a WHERE.3 min
  6. WorkflowsWhen to use a workflow and when to let the agent thinkThe most expensive decision in an AI system is not which model you pick. It is which parts you keep rigid and which parts you let reason.3 min
  7. OrchestrationDurable execution: agents that survive the server going downA process that takes 3 days cannot live in the memory of a process that restarts on every deploy. Here is how we solved long-running work.2 min
  8. ToolingFewer tools, better names: the hidden cost of giving an agent everythingEvery tool you add gets paid for twice: in tokens and in hesitation. 50 available tools do not make an agent more capable, they make it more indecisive.3 min
  9. Context engineeringContext engineering: the time a photo took an agent downA client sent 5 pictures of their shop over WhatsApp and the agent stopped responding. The cause was not the model: it was what we let into the context window.2 min

Do you want to bring AI into your company?

Message us on WhatsAppor send us an email →

We know how to do it. Tell us how things work today and we will tell you which part can run on its own, which needs an agent and which is better left to a person.