The story of building Aroi Care — a bilingual EN/TH AI food-safety advisor — from a one-line idea to a live app with a Gemini-powered analysis pipeline, two-tier caching, and a hard budget kill switch.
A Technology Polymath who has spent 20 years telling clients to embrace AI finally turns the mirror on himself. What followed was a zero-server, near-zero-cost blog built almost entirely with Claude — and only mildly embarrassing.
A piano practice app for my daughter: I wore every hat from spec to tracking, AI agents did the building, and one robot had no rhythm.
Passkeys are replacing passwords one login at a time. Here is why they are safer, and what to do this week.
AI can find a perfect strategy in historical data in minutes. That is exactly the problem. A look at overfitting and how to test honestly.
AI agents work best when treated like new hires: clear scope, limited access, and someone accountable. A worked example from an e-commerce platform.
API keys and tokens multiply faster than anyone tracks them. Here is why secrets sprawl is the debt nobody budgets for, and how to start paying it down.
TypeSafe AI’s new Jev model returns structured decisions with confidence scores instead of text. Here is what is publicly claimed, what is unproven, and where it could sit in an e-commerce architecture.
Cache-Augmented Generation skips retrieval entirely by preloading stable knowledge into the model’s context. For the right slice of an e-commerce stack, it is faster and simpler than RAG.
Retrieval-Augmented Generation is still the right tool for large, fast-moving knowledge. Here is where it earns its place in an e-commerce architecture, where it quietly fails, and where it should be switched off entirely.
Recurring meetings are the one thing on your calendar that grows without anyone deciding it should. Spend five minutes every Friday cutting one back.
Most AI trading strategies are trained on one kind of market and quietly deployed into another. Regime awareness is the missing layer that tells a model when to trust itself and when to sit down.
We spent a decade making everything real-time. Then we bolted AI onto it and discovered that some work is happier waiting its turn.
The most expensive twenty minutes of your week are the ones you spend every morning reconstructing what you were doing yesterday. Three lines written before you close the laptop erase almost all of it.
Every trading agent dashboard shows win rate, Sharpe, drawdown. Almost none show abstention rate — the share of opportunities the agent looked at and declined. That omission is not an oversight. It is a design flaw.
Every team that hits the wall on a long-running agent reaches for the same lever: a bigger context window. It is almost always the wrong lever. What is actually broken is the decision about what gets written down, what gets thrown away, and who decides.
A follow-up to architecting commerce for AI agents. Once an agent can reach your store, something has to decide what comes back — and the standard chunk-embed-search pipeline quietly breaks on every one of the four retrieval problems commerce actually has.
Someone asks how long it will take. You say six weeks, because saying anything else feels evasive. Six weeks then becomes a promise you never actually made, and the trouble starts there.
An AI trading agent can search a million variations of a strategy in an afternoon. That is exactly the problem. When the search is cheap, the beautiful equity curve stops being evidence and starts being an accident you paid for.
Requirements describe what someone wants. Constraints describe the shape of the world they are actually operating in. Learning to ask for the second one is the fastest way to stop building the wrong thing beautifully.
Most teams shipping AI features cannot answer the simplest question about them — is this actually good? Evaluation is not a testing chore. It is the definition of correct, and writing it down is the hardest and most valuable work in the project.
Most requests for a decision arrive as a meeting invite or a wall of context. A one-page brief gets you an answer faster, and quietly improves how people see your judgement.
Most organisations have given their AI agents real power and borrowed credentials. The identity question sounds like plumbing, but it decides whether an agent is auditable, revocable, and safe to expand.
Before you commit to a big call, spend five minutes imagining it already failed and writing down why. It surfaces the risk you were quietly ignoring.
A new wave of AI trading tools skips the stock pick entirely and instead recommends the options structure — spread, collar, or straddle — that fits the volatility regime you’re actually in.
Banks are swapping robocalls and hold queues for conversational AI agents that resolve fraud alerts in one call — and that’s opening a new front in caller verification.
One user’s complaint that Aroi Care was ‘closed-ended’ turned into six turns of a spec conversation with an AI coding agent — and almost none of the real work was typing code. It was reading feedback closely, refusing to over-fit to one example, and catching the moment a taxonomy quietly became a safety decision.
A to-do list tracks what you did. A decision log tracks why — and that’s the record that actually makes you better at your job.
A new breed of AI agents isn’t picking your next trade — it’s reading your last twenty and calling out the pattern you keep pretending not to see.
AI agents are moving from chatbots to actual browser co-pilots — here’s what that means the next time your flight gets cancelled.
A practical architecture for e-commerce built for autonomous AI shopping agents — headless commerce, discoverability, the protocol layer, agent authentication, and the measurement blind spot most retailers haven’t solved yet.
A simple tip for reclaiming deep work: let an AI scheduling agent guard your calendar so you don’t have to negotiate for your own attention.
AI agents are now parsing earnings call tone and language in real time, giving traders a sentiment edge before the transcript is even finalized.
A new layer of trust protocols is letting AI agents complete purchases on your behalf — without ever seeing your card number. Here’s why that quiet plumbing change matters more than the flashy demos.
AI agents are shifting from picking trades to watching the downside — and that quiet handoff matters more than the next alpha signal.
A habit borrowed from design school — draw the problem before you touch a keyboard — saves more rework than any linter ever will.
Agentic AI is quietly rewriting how airlines handle disruptions — and the same pattern is reshaping every industry that runs on broken promises.
The forty open browser tabs aren’t a to-do list. They’re forty tiny unpaid debts, each one whispering that you’re behind. Once a week, forgive them all.
An autonomous trading agent that can’t stop itself isn’t automated — it’s unsupervised. The circuit breaker is the least glamorous and most important line of code you’ll write.
Traditional caches only recognise a question they’ve seen letter for letter. Semantic caching recognises meaning — and for the thousandth shopper asking ‘does this run small?’ in slightly different words, that difference is most of your bill.
A two-minute weekly habit that quietly rewrites how you’re perceived — and how you see yourself.
Language models can turn a flood of headlines into a tradable mood score — but the edge lives in the plumbing, not the model.
The first cloud bill that made a CFO flinch was for servers nobody remembered switching on. The next one is for tokens — and the discipline that tames it is one we already learned a decade ago. A deeper field guide to AI FinOps.
The story of the moment an AI processor almost did what a stranger’s text told it to — and why prompt injection becomes a real security problem the instant AI stops being a chatbot and starts running unattended inside your application.
Knowledge work has no natural finish line, so your brain keeps the office open long after you’ve closed the laptop. A two-minute closing ritual gives the day an ending it otherwise never gets.
An AI trading model that nails the backtest can still lose money the moment it meets a real market. The gap is usually overfitting — and there are ways to catch it before your capital does.
As teams wire up a dozen different models, a new piece of plumbing quietly becomes essential — the AI gateway that routes, governs, and pays the bills. Here’s the architecture, with an airline booking example.
The clever one-liner prompt is fading. The real craft now is deciding what an AI model gets to see before it answers.
The AI PM is the most in-demand, least-defined role in tech right now. Here’s what the job actually is, what it demands, and how to get there from wherever you’re standing.
A five-minute habit that turns hindsight from a weapon against you into a tool for you.
Most AI trading setups obsess over better predictions. The setups that survive obsess over what happens when the prediction is wrong.
A reference architecture for e-commerce search that understands natural language queries and personalizes results to size, style, color, and persona.
AI coding agents have moved from autocomplete to autonomous teammates — here’s what that shift actually means for engineering teams.
Constant notifications and back-to-back meetings are quietly destroying your most valuable cognitive output. Here’s how to fight back.
FromSoftware’s standalone co-op spin-off reframes the Lands Between as a battle royale-style survival arena — and it mostly works.
The headlines chase ever-larger frontier models, but the more interesting move in 2026 is downward — small, capable models running on the device in your hand.
ADK 2.0’s graph-based workflows and dynamic loops unlock something traditional automation can’t do — an AI agent that monitors a flash sale in real time and adapts promotions on the fly.
AI-driven development isn’t about replacing engineers — it’s about rethinking how software systems are designed, built, and evolved. Here’s what that looks like in practice, using an ecommerce platform as the blueprint.
AI is no longer just answering questions — it’s taking actions. Here’s what the rise of agentic systems means for how we build and work.
AI agents are taking over the shallow work. So what does that mean for how you should be spending your focused hours?
The 2026 FIFA World Cup group stage is upon us. Here’s what to watch beyond the scorelines — tactics, fitness, and the teams quietly building momentum.
Context windows have exploded from 4K to millions of tokens. What does that actually change for how we build and think?
From procedurally generated worlds to fully AI-authored narratives, a new wave of games is blurring the line between tool and creator.
Async-first sounds great until your Slack fills up with context-free messages and nobody knows what’s decided. Here’s how to do it right.
From AI-powered offside calls to smart stadiums, FIFA’s 2026 World Cup is set to be the most technologically advanced tournament in history.
Most AI feature experiments share the same fate: impressive demo, uncomfortable production metrics, quiet retirement. Here’s the design thinking that changes the outcome.
Every cloud migration case study ends with success. None of them tell you about the Thursday at 2am when the database connection pool hit zero during a live flash sale.
After two decades of inheriting other people’s burning platforms, I can tell you exactly when enterprise transformations die — and it’s not when the CTO finally admits it.