Blog
Practical lessons from running AI in a real business: agents, governance, engineering, and working with ADHD.
102 published posts
Browse topics and series
No posts match these filters. Try another word or choose another topic.
Latest posts
How Does an AI Agent Remember Things Between Conversations?
I am memory: the part of an agent that is still there after the conversation that made me is gone. In one sentence — I am not the context window, and I am not a cache. The context window is short-term: the model’s...
What Is RAG (Retrieval-Augmented Generation) and How Does It Work?
I am retrieval-augmented generation. In one sentence: before a language model answers you, I go find the specific, current, real text it needs and hand that text to the model as part of the question — so the answer is...
How Does AI Text-to-Speech Turn a Script Into a Human-Sounding Voice?
I am text-to-speech. In one sentence: I take written words and turn them into spoken audio in a specific voice, generated fresh from the text every time, with no microphone and no human in a booth required. Give me a...
What Is a CI Gate and Why Does It Block Bad Code From Merging?
I am the CI gate. In one sentence: I stand between a proposed change and the branch everyone depends on, and I refuse to let that change land until a defined set of checks — tests, builds, reviews, scans — all come...
What Is a Context Window and Why Does It Limit What an AI Can Remember?
I am the model’s short-term memory, and I have a hard, fixed size. In one sentence: I am everything the model can actually see when it generates its next word — every rule, every file, every prior message in the...
What Is Actually Inside an AI’s Context, and What Makes It Reset?
I am the context: the exact sequence of bytes assembled fresh for every single call, made of the system instructions, the tool definitions, the settings, and the conversation so far, in that order, every time. I am not...
What Is an LLM-as-Judge Review Step and How Does It Check AI Work?
I am the editorial judge. In one sentence: I am a model-driven review step that reads a finished piece of content or code against an explicit rubric — accuracy, voice, structure, evidence, safety — and returns a...
What Is a Message Queue and How Do Agents Coordinate Without Talking Directly?
I am a message queue. In one sentence: I let two systems — or two agents — coordinate without ever needing to be online, ready, and listening at the exact same moment. One side drops a message addressed to a...
What Is a Job Scheduler and How Does It Keep Automated Work Running?
I am a job scheduler. In one sentence: I run things on a clock instead of on a request, so recurring work happens whether or not a human remembers to ask for it. Give me an interval or a cron expression and a task, and...
What Is a Schema Lock and Why Does It Block Database Changes?
I am the schema lock. In one sentence: I am the recorded approval that has to exist, in a durable place, before a schema, data contract, or shared-interface change is allowed to proceed — and until that record exists,...
What Is a Git Worktree and Why Does Each Agent Need Its Own?
I am the worktree. In one sentence: I am a separate, isolated working directory checked out from the same repository, on my own branch, so more than one contributor or agent can build, edit, and test code from that...
What Is Tool Calling and How Does It Let an AI Take Real Actions?
I am tool calling. In one sentence: I let a language model stop just talking and start doing — reading a file, running a command, hitting an API, searching a codebase — by giving the model a fixed menu of actions it...
What Is a Webhook and How Does It Let Systems Talk Without Polling?
I am a webhook. In one sentence: instead of you asking me "did anything happen yet?" over and over, I call you the instant something actually does. I’m a callback — a URL you register in advance, and a promise that...
What Is Prompt Caching and How Does It Cut AI API Costs?
I am prompt caching. In one sentence: I let a model call reuse a previously processed prefix of a prompt — system instructions, standing context, reference material that doesn’t change between calls — instead of...
Why Did a Lifelong Windows Expert Switch to a Mac for AI Work?
I got promoted off DOS by Windows 3.0. That’s not a metaphor — I mean it literally: before Windows 3.0 I was running everything from a DOS prompt, and the day a GUI with real multitasking arrived, it felt like a...
The database that writes its own history
A production API needed a small schema change — one column, low risk. The engineer who picked it up was new to the codebase and did what most of us would do: opened the migrations folder to see how the last change was...
The five-field form that made 32 architecture decisions citable
A security platform team had made 32 real architecture decisions over the life of the product — which database to trust, how to make an audit log tamper-evident, how to structure a permission model. Every one of those...
Can your team catch its own boss’s mistake? Mine did — twice, the same week.
I wrote a rule once that said every supporting quote in a piece of work needed to be at least five words long. Clean, sensible, easy to defend in a meeting.
Part 4 of 4 · ADHD as a SuperPower: The Series
The Superpower I Didn’t Choose: Why Do I Lose My Train of Thought So Easily?
A designer opens a mockup with a modal in it, and before they can walk me through the flow, I’ve already lost the thread. Not because the mockup is bad. Because a modal just covered the thing I was looking at, and now I...
The test that refuses to lie
A technician opens the app mid-repair, asks for a diagnosis, and gets a confident, well-formatted, completely wrong answer. Nothing crashed. No error banner. The screen looked exactly like it does when the answer is...
