Writing
Engineering, delivery, and building things which last.
The Greenbox Story · Finding the Fit
Prioritisation: What Changes First
The team knows what's wrong — the value prop, the pricing, the unit economics. They can't fix everything at once. Now they need to decide what changes first.
Read article
Exam Room · AIF-C01
SageMaker, Bedrock, or a Managed API
A platform team has five AI-shaped requests landing in a single sprint: transcribe call centre audio, detect anomalies in sensor data, extract text from scanned forms, summarise customer emails, and detect faces in CCTV. Someone has already typed 'use SageMaker' into three design docs. Someone else insists Bedrock is the answer. A third voice mutters about purpose-built services. AWS has at least three answers to every AI problem, so there's no single platform that wins; what matters is how to tell which layer of the stack each request lands on, and what that choice costs in time, money, and flexibility.
Read article
Under the Hood · The AI Field Guide
The Boring Baseline That Wins
TF-IDF, logistic regression, naive Bayes, k-means, LDA. The fifty lines of scikit-learn that beat your fancy model on the small problem you actually have. Why these baselines still win, and why the correct starting point in 2026 is often the same as it was in 2006.
Read article
Under the Hood
A Gentle Guide to Typography: From Chisels to Character Sets
A warm, thorough walk through the world of typography, from hand-carved letters and Gutenberg's press to fonts, glyphs, kerning, serifs, and Unicode. Everything you wanted to know about how written language gets its shape.
Read article
Under the Hood · Time
The Clock Inside You
Your body has its own clock, and it doesn't care about UTC. It runs on light, adenosine, and a cluster of 20,000 neurons behind your eyes. Jet lag, shift work, larks and owls -- the biology of time is a different machine from the physics, and it keeps its own hours.
Read article
Exam Room · AIF-C01
Agents, Chains, and Retrieval
A product manager wants a 'GenAI assistant' for internal operations -- something that can answer questions, look up customer records, draft emails, and file Jira tickets. Three architectural patterns keep coming up: chains, retrieval, and agents. They sound similar, they all use foundation models, and teams routinely reach for the most elaborate one when a simpler pattern would do. There's no single 'best' here; what matters is which one fits each piece of the assistant's workload, and when elaboration costs more than it earns.
Read articleThe Greenbox Story · Finding the Fit
Pricing Experiments: The Right Box at the Right Price
Greenbox has been charging $25 for the small box and $45 for the large box since day one. Nobody complained. Maya thought that meant the prices were right. A conversation with Lee at the kitchen table suggests otherwise.
Read article
Exam Room · AIF-C01
Guardrails, Watermarks, and Refusals
A fintech ships a customer-facing chatbot on Bedrock. Legal asks: can it give financial advice? Risk asks: can it leak customer account numbers? Compliance asks: if an auditor requests proof a response came from our model, can we demonstrate it? Three questions, three different controls, all of them Bedrock-native. The controls exist; the work is matching the right one to each question and figuring out what the shape of a 'responsible AI' configuration actually looks like when the auditor arrives.
Read article
Under the Hood · The AI Field Guide
Before the Transformer
n-grams. HMMs. CRFs. The language models and sequence taggers that ran the internet before deep learning, and that quietly still do, in autocomplete, spam filters, biomedical NER, speech recognition. What they are, why they still ship, and when they're the correct answer.
Read article
The Workshop
The Workshop: Assumption Mapping
A pattern for surfacing the beliefs hiding underneath a plan, plotting them on an evidence/impact grid, and deciding which ones to test before you commit. Facilitator's playbook, failure modes, and what to do with the grid afterwards.
Read article
Under the Hood · Time
Can You Turn Back Time?
General relativity permits time loops. Quantum mechanics hints that the future can influence the past. Hawking threw a party for time travellers and nobody came. The physics of time travel is stranger, and more serious, than science fiction suggests.
Read article
Exam Room · AIF-C01
Forecasting Without Writing Python
A category manager has 18 months of weekly sales data for 400 SKUs and a deadline to forecast next quarter. She doesn't code. The ML team is booked until Q3. The ask is a tool that lets her build a forecast herself -- importable, reviewable, explainable -- without waiting for engineering. Which AWS box she clicks matters less than what kind of problem this actually is, what features of the data can honestly feed into a model, and what the business user has to understand for the output to be defensible when finance asks ''why this number?''.
Read article
The Greenbox Story · Finding the Fit
Business Model Canvas: Does This Actually Work?
Maya needs to convince investors that Greenbox can reach 1,000 subscribers. But when the team maps the business model, they discover the unit economics don't add up — and Lee admits he's out of his depth.
Read article
Exam Room · AIF-C01
Grounding a Chatbot in Your Own PDFs
A facilities team has 600 PDFs -- equipment manuals, safety procedures, maintenance schedules -- sitting on a SharePoint drive. Engineers want a chatbot that answers 'how do I reset the chiller on floor 4?' in seconds instead of a ten-minute PDF hunt. Retrieval-augmented generation can do this; whether it does it well depends on what the corpus actually looks like, what kinds of questions the engineers really ask, and which configuration knobs decide whether the answers are any good once a managed service is on the table.
Read article
Under the Hood · The AI Field Guide
After the Transformer
Transformers have ruled language modelling for nearly a decade. They have a known weakness, and several research lines are trying to replace them. Mamba, RWKV, RetNet, Hyena, diffusion-for-text -- what they are, what they fix, and which ones are likely to matter.
Read article
The Workshop
The Workshop: Jobs to be Done
A half-day pattern for finding out what your customers actually hired your product to do. Switch interviews, the four forces, shaping the material into job statements, and what to do when the room wants to list features instead.
Read article
Under the Hood · Time
Does Time Even Exist?
The arrow of time isn't in the equations. "Now" isn't a location. The most fundamental theories of physics may contain no time variable at all. A tour of the foundations, from the block universe to the holographic principle.
Read article
Exam Room · AIF-C01
Prompt, Retrieve, or Fine-Tune
A legal-ops team wants a model that answers questions about their 4,000 in-house contract templates. The first prototype, a plain Claude call with the question in the prompt, hallucinates clause numbers. Someone suggests fine-tuning; someone else suggests RAG. They solve different problems, so 'which is better' is the wrong frame; what matters is which problem the team actually has, and what each adaptation technique costs in time, data, and recurring spend.
Read article
The Greenbox Story · Finding the Fit
Assumption Mapping: Testing What You Believe
The team lists everything they believe but haven't validated, ranks by risk, and discovers that their most confident assumption is the one most likely to be wrong.
Read article
Exam Room · AIF-C01
From Raw Model to Production Endpoint
A product team wants a chatbot that summarises support tickets. They have the tickets, a cloud account, and no ML background. Somebody says 'use a foundation model'. Between that sentence and a working endpoint sit roughly seven distinct stages, each with its own AWS service and its own decisions. Picking the model is the easy part; the real work is figuring out which stages this team can skip, which they absolutely cannot, and what AWS gives them at each step.
Read article
Under the Hood · The AI Field Guide
The Reranker You Didn't Know You Needed
RAG explanations stop at 'embed the query, look up the nearest documents, hand them to the LLM.' That's the demo. In production, there's a second pass between the lookup and the LLM, and it's the one that actually makes retrieval work.
Read article
Consulting and Craft · Through the Kitchen
The Knife in My Hand
A sharp knife is safer than a dull one. Pick the right tool for the job, then put in the practice. Speed comes from practice, not pressure. A few focussed, well-maintained, well-practised tools will more predictably take you further than the latest beautiful shiny offering.
Read article
Under the Hood · Time
Time Is Weirder Than You Think
Time doesn't flow at the same rate everywhere. It slows near massive objects, dilates at high speeds, and might not 'flow' at all. From GPS corrections to black holes, the physics that makes time strange.
Read article
Exam Room · AIF-C01
Buy, Borrow, Build
A product manager with no ML background has been told to add AI to a SaaS product, and has heard of Bedrock, SageMaker, Comprehend, Translate, Textract, Rekognition. AWS has three different shapes of AI offering, and the shortest path depends entirely on whether a ready-made service already does the job.
Read article
The Greenbox Story · Finding the Fit
Jobs to Be Done: Why Subscribers Actually Stay
Greenbox discovers that subscribers don't stay for fresh vegetables. They stay because they don't have to think about dinner on Tuesday.
Read article
Exam Room · SAA-C03
The Archive Nobody Reads
Some data exists for compliance, not for use. Tens of terabytes of records sitting untouched until an auditor wants them. S3 has eight storage classes; only one of them is built for that pattern, and getting it wrong can cost an order of magnitude in a year you weren't paying attention to the bill.
Read article
Under the Hood · The AI Field Guide
The Other Transformers
BERT and T5 are transformers too, but they aren't trying to be ChatGPT. They're trying to be the boring layer underneath -- classifiers, embeddings, structured transformations -- and they're often a better answer than an LLM.
Read article
The Workshop
The Workshop: User Story Mapping
A pattern for laying out the whole user experience as a left-to-right narrative and then slicing it into releases, so the team can see both the shape of the thing they're building and the thinnest honest version they can ship first.
Read article
Under the Hood · Time
Ticks or Tocks?
A second used to be a fraction of the day. Now it's defined by the vibrations of a caesium atom -- and even that might not be precise enough. From quartz watches to optical lattice clocks, the story of how we learned to count time.
Read article
Exam Room · SAA-C03
The Closest Healthy Region
A multi-region application needs to route requests to the closest healthy region, failing over automatically when the preferred one drops out -- with no client-side retries and no extra health-check plumbing to maintain. Route 53 can do all of that in a single record set. Finding the correct combination means touring all seven routing policies and the attributes that separate them.
Read articleThe Greenbox Story · Shipping What Matters
Accessibility: A Product Decision, Not a Compliance Tick
A subscriber sends Greenbox an email that changes how they think about the signup flow. Not a complaint -- a question the team didn't know they needed to answer.
Read article
Under the Hood · Time
What Day Is It?
The date next to the time on your phone is its own kind of fragile. Calendars argue with the moon, the sun, and each other; whole days have been deleted by decree; and the year number on your screen depends on which monk's arithmetic your ancestors trusted.
Read article
The Workshop
The Workshop: Impact Mapping
A pattern for tracing a business goal through the people whose behaviour has to change, through the changes themselves, to the things you might build — so the team can tell the difference between work that moves the number and work that just feels productive.
Read article
The Greenbox Story · Shipping What Matters
User Story Mapping: Seeing the Whole
The Greenbox team has a growing backlog of stories but no sense of the whole. User Story Mapping lays out the full user journey, shows the gaps, and makes release planning obvious instead of political.
Read article
The Workshop
The Workshop: Sprint Planning
A pattern for turning a refined backlog into a sprint the team believes in: a goal, a set of stories, task breakdown, and an explicit commitment, in one focused session.
Read article
Under the Hood · Time
What Time Is It?
The hour on your phone is a fragile compromise between the sun and politics. Sundials, shipwrecks, railway time, DST, and the volunteer-maintained database that keeps the world's clocks roughly honest.
Read article
The Workshop
The Workshop: Example Mapping
A pattern for breaking a user story into rules and concrete examples in twenty-five minutes, so the team knows whether it's ready to build before the sprint starts. Facilitator's playbook, failure modes, and what to do with the cards afterwards.
Read article
The Greenbox Story · Shipping What Matters
Impact Mapping: Connecting Work to Goals
The Greenbox team is shipping features, but are they the correct ones? Impact Mapping connects engineering work to business goals -- and reveals that the obvious next feature isn't always the important one.
Read article
The Workshop
The Workshop: Event Storming a Process
The default Event Storming session and the one you'll run most often. One process, one wall, everyone who touches it in the room, three hours. You leave with a precise shared model of the flow and a short list of the questions it raised. Where Big Picture looks for shape across a whole domain, Process Level looks for precision within one flow.
Read article
The Workshop
The Workshop: Event Storming a Domain
The widest zoom, and Brandolini's own entry point to Event Storming. Stand a whole business (or a whole product) in front of one wall with everyone who touches it. The output isn't a design or a plan; it's one picture that every department recognises, and a shortlist of the places worth digging into next.
Read article
The Greenbox Story · Shipping What Matters
Teaching Your LLM the Codebase: CLAUDE.md and AGENTS.md
The Greenbox team's actual CLAUDE.md and AGENTS.md files. What goes in them, how they're structured, and how they shape the code an LLM generates.
Read article
The Greenbox Story · Shipping What Matters
Teaching Your LLM the Codebase
Tom and Priya are both using LLMs to write code. They're getting different results. Not wrong -- different. CLAUDE.md is how the team teaches the LLM to write code like them.
Read article
The Greenbox Story · Shipping What Matters
Behaviour-Driven Development: From Stories to Working Software
Example Maps become stories. Stories become tests. Tests drive code. And in a world where LLMs can write the code, the discovery work matters more than ever, because the bottleneck isn't implementation any more; it's knowing what to build.
Read article
Consulting and Craft · Through the Kitchen
The Quiet Jar in the Fridge
My last sourdough starter died through a quiet chain of postponed feeds. I'm starting a new one today. Most of what I'm learning as I begin again, I wish I'd known a decade earlier about codebases.
Read article
The Greenbox Story · From Chaos to Clarity
Sprint Planning: Turning Sticky Notes into Delivery
Walls of sticky notes, a prioritised backlog, concrete examples, and six weeks until the funding deadline. The team has done the discovery. Now they need a rhythm for delivery.
Read article
Under the Hood · The AI Field Guide
To LLMs… and Beyond!
LLMs are one corner of a much larger field. Diffusion models, reasoning models, multimodal systems, open-weight vs closed — what they are, how they differ, and how to choose.
Read article
The Greenbox Story · From Chaos to Clarity
Example Mapping: Making Stories Concrete
Four colours of card, twenty-five minutes, and a vague story becomes something you can actually build. Example Mapping is the bridge between understanding and implementation -- and the technique you'll use more than any other.
Read article
Under the Hood · The AI Field Guide
How LLMs Actually Work
Tokens, transformers, attention, and the training pipeline: what large language models actually do when they 'predict the next token', why they hallucinate, and why they're so good at code.
Read article
The Greenbox Story · From Chaos to Clarity
Event Storming: Building Shared Understanding
The Greenbox team covers a wall in sticky notes and discovers they've been building on assumptions. A deep dive into Event Storming, the workshop that gets an entire domain out of one person's head and into shared understanding.
Read article
The Greenbox Story · From Chaos to Clarity
Retrospectives: Catching the Wrong Kind of Fast
A small team, a good idea, LLMs generating code at lightning speed, and four weeks of building the wrong thing. The first part of a series on getting from discovery to delivery without wasting everyone's time.
Read article
Consulting and Craft · In Practice
The Value Is in Ideas, Not Code
LLMs have made code implementation almost trivial. The bottleneck has shifted from writing code to knowing what to ask for. Your library of patterns, concepts, and hard-won experience is now your competitive advantage.
Read article
The Greenbox Story · From Chaos to Clarity
Minimum Viable Product: The First Box
Twenty-two people signed up from a flyer at the Margaret River farmers market. The first delivery was a disaster -- wrong addresses, wilted spinach, and a box that arrived on Wednesday instead of Thursday. Maya cried in her car. Then she drove to the next address.
Read article
The Greenbox Story · From Chaos to Clarity
Customer Discovery: Before the First Line of Code
Maya grew up on a farm outside Margaret River. She left to study computer science, spent a decade in consulting, and came back to Perth with an idea that wouldn't let go: what if the best produce in Western Australia could reach people's doors every week?
Read article