Long explanations, when the subject earns one

Each piece answers one question properly and all the way through. No listicles, and no predictions about what happens in five years.

16 articles

Five stacked layers from electricity to business, and an arrow showing that value accumulates from the bottom up.

The AI ecosystem: from electricity to the customer's benefit

The whole chain in one piece: where the electricity comes from, what happens in a data centre, how a model is trained, and where the money finally ends up.

4 August 202614 min

On the left one large glowing block, on the right an endless row of small units passing through three gateways of increasing width. A one-off outlay and endlessly repeated use are different things, and the same row looks different depending on how widely the measurement boundary is drawn.

Article · Infrastructure

What one AI question actually uses

The bottle-of-water claim comes from a real study whose result changed meaning as it travelled. What it measured, what production measurements now show, and why two honest numbers can differ by a factor of a hundred.

30 August 202612 min

One stream branches into two. On one route a large heap piles up past a checkpoint; on the other only a single item gets through the same kind of gate. Volume produced and work delivered are not the same thing.

Article · Applications

Vibe coding and AI-assisted programming

Two different ways of working get grouped under one label, though the distinction that matters is whether the generated code is reviewed. What each approach is good for, what remains after the demo, and what productivity research actually shows.

20 August 202620 min

A grid of everyday subjects from the home to the forest and from the production line to sport, that is, the range of AI use cases.

Article · Applications

AI use cases in plain language

Thirty concrete use cases, from the home to the forest, from the production line to a sports club. What the machine does in each, where it goes quietly wrong, and what happens to a use case after it is chosen.

9 August 202618 min

Four agent structures side by side: a single agent, a chain, a parallel run and a supervisor tree.

Article · Applications

Agents, orchestration and use cases

What a multi-agent system actually is, the four structures you find in one, what each is good for, and when a single agent is enough.

6 August 202619 min

On the left a lifeless list of principles, on the right one of them turned into a decision point with two ways out.

Article · Rules and responsibility

AI ethics without the sermon

Lists of principles settle nothing. Seven places where the harm actually happens, and the questions that turn them into decisions.

6 August 20265 min

Nine things sorted into three baskets. One is still in transit and its basket has not been decided.

Article · Rules and responsibility

An AI playbook somebody actually reads

Nine points, a three-basket model, and a list of what not to put in. The first version takes a day.

6 August 20265 min

A long queue of projects waiting at a narrow gate, with only a few that have made it through on the other side.

Article · Infrastructure

Finland's data centre boom and the queue for the grid

Over three gigawatts under contract and 50,000 megawatts as enquiries. What the numbers actually mean, and where the real bottleneck is.

6 August 20265 min

An enormous column of investment whose top runs off the top of the picture, and beside it a far shorter column of measured results.

Article · Business

Is there an AI bubble

Four companies will invest 760 billion dollars this year. Meanwhile most companies get no measurable effect on their results. Both can be true.

6 August 20265 min

A timeline on which one deadline has moved later. The old date is faded and an arrow points to the new one.

Article · Rules and responsibility

The AI Act: what applies, and when

The high-risk deadline moved in the summer of 2026, but most guides still quote the old dates. What an organisation using AI actually needs to know.

6 August 20268 min

A document in which one line differs from the rest and hijacks the processing towards a different outcome than intended.

Article · Applications

When an agent believes the wrong text

Prompt injection has topped the OWASP list for a third year, and there is no fix. Why not, and what to do instead.

6 August 20265 min

Two shapes of different kinds that partly overlap, but neither contains the other. They answer different questions.

Article · Models

Sovereign AI and base models

Two phrases that turn up in the same sentence but answer different questions. One says what kind of model it is, the other says on whose terms it is used.

5 August 20267 min