AI - a tool

From Shallow Autocomplete to Deep World Modelling

 This section explores the core mechanism of Large Language Models (LLMs): autoregressive next-token prediction. It contrasts the shallow, immediate context of a smartphone keyboard with the deep, global context of an LLM. It explains how the singular objective of guessing the next word forces the model to build an internal simulation of the world to resolve ambiguities (e.g., tracking suspects in a murder mystery or applying the laws of physics).

Belongs to the series: The Architecture of Approximation: How Next-Word Prediction Unlocks the Structure of Human Knowledge

The Keyboard in Your Pocket

You are typing a message to a friend. You thumb out "I'll meet you at the" and your phone, without being asked, floats three suggestions above the keyboard: pub, station, library. You tap pub. The phone did not think about the pub. It did not picture the sticky tables, the half-pint of lager, the friend already halfway through a packet of crisps at the corner table. It looked at the two or three words behind your cursor, scanned a small statistical table of what usually follows "meet you at the", and handed you the most frequent candidate. That is the whole trick. No world. No meaning. Just a very short memory and a frequency chart.

Now scale that trick up. Not by a factor of ten or a hundred, but by a factor of hundreds of billions. Replace the two-word lookback window with a context that can swallow an entire novel. Replace the frequency chart with a dense web of relationships that encodes, in compressed mathematical form, a sizeable fraction of everything humanity has ever written down. The result is a large language model. And here is the strange part: the task has not changed. The model is still doing the one thing your phone keyboard does. It is guessing which token comes next.

That sentence ought to feel underwhelming. After all the breathless coverage, after the think-pieces about artificial general intelligence and the warnings about existential risk, the core operation of every large language model on Earth is a glorified autocomplete. It never graduated to something harder. Nobody added a reasoning module on top, or a logic engine, or a truth-checker. The architects set a single, almost embarrassingly simple objective—minimise the probability of getting the next word wrong—and then let the mathematics run for months on thousands of graphics processors. What came out the other end was something that can pass a bar exam, write a working Python function, and explain quantum entanglement to a twelve-year-old. The simplicity of the task is the illusion. The complexity of what the task demands is the reality.

There comes a moment when a tool becomes so complex that you forget you invented it. It does not disappear — it is present, massive, functioning. But you no longer know exactly what it does, why it does it, or for whom. And, above all, you are no longer sure who is commanding whom.

We are, in 2026, precisely at that moment. And it is not a metaphor.

If you are between 20 and 25 and reading this, it means you grew up alongside the internet, witnessed the birth of the smartphone as a pocket object, entered university whilst ChatGPT was barely out of beta, and now work or study alongside AI systems that write, code, analyse, consult, and make decisions in place of people who once earned good salaries for exactly those things. Your generation does not have the luxury of distance. You are in the thick of the experiment, not in the stands.

That is why I want to speak directly, without paternalism and without the useless nostalgia of someone lamenting that things were better "in my day". They were not. They were different. But different does not mean today's direction is correct.

The thesis of this text is simple and, I hope, disturbing: we have built extraordinary machines without answering an elementary question — for what? And the absence of that answer is not an administrative oversight. It is a foundational philosophical error, with consequences that will be felt in the body and mind of your generation more than in any other before it.

How a Conference Lanyard Exposed the Trust Crisis at the Heart of Artificial Intelligence

There is a photograph. Two men at a technology conference, sitting close enough that their shoulders nearly touch. One of them is Marc Benioff, the chief executive of Salesforce, a company worth more than the gross domestic product of most countries you could name without pausing. The other is Sam Altman, who runs OpenAI — the organisation that built the software Benioff is about to argue with. Around both men's necks hang lanyards: those strips of nylon webbing that every conference in the world distributes, that nobody keeps, that end up in desk drawers or rubbish bins within forty-eight hours of being issued. Benioff wants his removed from the photograph. He asks ChatGPT to do it.

What follows is not a story about a string but a violation of rights, legal semantics that cannot be fully understood by AI but fills that void with legal hallucinations...pam, pam, pamm.

The Verdict Before the Question

The model responds with something that has become, in the language of AI product development, a "refusal." It tells Benioff that it cannot edit the image because doing so would constitute copyright infringement. Six consecutive refusals, by his account. Six times the word "no," dressed in the vocabulary of legal restraint.

This is worth pausing on, not because the refusal was rude — it was not — but because of what it reveals about the architecture of assumption that sits beneath it. Benioff owns the photograph. Or, at minimum, he is in it, he commissioned it, and he controls its use. The lanyard belongs to a conference. The conference is over. The lanyard is fabric. There is no copyright case here by any reasonable reading of intellectual property law, and yet the model — trained on more legal text than any lawyer will read in a lifetime — invented one.

This is the particular quality of the error that matters. The model did not fail to understand a request. It understood the request perfectly and responded to it with a fabricated obstacle, presented with the confidence of established precedent. It did not say I am uncertain whether this is permissible. It said this is infringement, with the assurance of an institution that has already considered the matter and reached a conclusion.

In the philosophy of law, there is a principle so elementary that it functions as something closer to an instinct: you do not convict a person before hearing the case. Benioff had not finished explaining what he wanted before the system had already placed him on the wrong side of a legal line that does not exist. The presumption of guilt arrived ahead of the evidence, which is to say it arrived in place of it.

A true story about office furniture, professional fear, and a prompt that changed a Tuesday morning

Author's note: The characters are composite fictions. The town hall, the school, and the county are real as types, not as identities. Any resemblance to living civil servants is, in equal measure, unavoidable and deliberate.

The Tuesday with Furniture

At 8:17 on a Tuesday morning, Gheorghe Ciubotariu opens the window of his office at the town hall in Hliboca — not the Ukrainian Hliboca, but the Romanian one: a market town of a few thousand souls separated from the border by twelve kilometres of flat farmland and a time zone that nobody chose. Through the window comes the smell of wet linden blossom and the sound of a tractor. On his desk: a mug of instant coffee, a blue Schneider biro with a chewed cap, and the procurement requisition that the mayor flung on his table on Friday with the words: "Gheorghe, sort it. By Thursday."

The requisition has three lines. Office furniture — chairs, desks, cupboards, shelving units. The council chamber and the secretariat. Estimated value somewhere below the equivalent of twenty-six thousand euros — which means a direct purchase or, if anyone has been adding items, a simplified tendering procedure. The mayor is a newcomer from the private sector who has suddenly noticed that the council chamber looks like a 1994 dental surgery and that, on the 15th of July, a county delegation is coming to visit.

Gheorghe is fifty-four years old. He has been a public procurement adviser for nineteen years — that is to say, before SEAP (Romania's national e-procurement platform), before e-licitatie.ro, before anyone had heard of Law 98/2016 on public procurement. He has survived three iterations of procurement legislation, two reorganisations of ANAP (the National Agency for Public Procurement), the transition from paper files to scanned files to uploaded files, and every platform update that arrived with usage instructions dispatched on Friday evenings at 17:58.

Gheorghe does not hate technology. He respects it the way a mountain person respects an unfamiliar path: with attention, with scepticism, and with the awareness that one wrong step is enough.