A guided library for understanding AI Understanding Machine AI elucidaria — a plain-language handbook, in the old sense

Understanding AI in a Month

Thirty days, one idea a day, from nothing to following the argument.

17 of 30 lessons published. The course is being written one lesson at a time, so it cannot be completed today: 13 lessons are still to come.

Start Lesson 1

A thirty-day course in what these systems actually are. Each day is a written article with its figures and its sources, an audio edition, and a printable copy. Days are published as they are made.

Written and produced by an automated pipeline. This block is a process disclosure, not an independent factual certification. Each lesson carries its own record of what was checked before it was published.

  1. Day 1 Published Sources through about 22 min read 12 min listen 8 figures

    The Model Is Not the Product

    In July a benchmark score nearly tripled without anything inside the model changing. The interesting question is not how the software was improved.

  2. Day 2 Published Sources through about 30 min read 11 min listen 5 figures

    How Text Becomes Tokens

    The best-known failure of large language models is that they cannot count the letters in a word, and the best-known explanation for it is that the word arrives broken into pieces.

  3. Day 3 Published Sources through about 24 min read 12 min listen 5 figures

    One Token at a Time

    What "predict the next token" actually means — and the half of the sentence that almost every popular account leaves out.

  4. Day 4 Published Sources through about 23 min read 12 min listen 4 figures

    How Words Affect Other Words

    The operation that lets a word at the end of a sentence change what a pronoun in the middle refers to is nine years old, was named after something it does not do, and is now a minority of the layers in the models whose internals can be read.

  5. Day 5 Published Sources through about 31 min read 12 min listen 6 figures

    How an Answer Unfolds

    A machine asked the same question a thousand times, with greedy decoding requested, returned eighty distinct completions. What decides which word comes next, and how much of the past the decision may consult, are two settings — and both are somebody's decision rather than a fact about the machine.

  6. Day 6 Published Sources through about 26 min read 12 min listen 7 figures

    What the Application Adds

    A trained model reads text and writes text. Everything a product appears to do besides that — searching, opening files, running commands, remembering a previous conversation — is ordinary software deciding what text to place in front of it and what to do with the text it returns. That division decides a great deal about what a system costs to run. It decides less than the industry's own marketing suggests about whether the system is right.

  7. Day 7 Published Sources through about 31 min read 11 min listen 7 figures

    Where Training Text Comes From

    Since August 2025, companies placing general-purpose AI models on the European market have been required to publish a summary of the content used to train them, and since 2 August 2026 the European Commission has been able to fine those that do not.

  8. Day 8 Published Sources through about 45 min read 12 min listen 10 figures

    Why Scale Worked

    For a few years the papers announcing the largest artificial-intelligence models stated their size in the first paragraph. The largest American developers no longer state it for their flagship models.

  9. Day 9 Published Sources through about 25 min read 12 min listen 6 figures

    From Base Model to Assistant

    A language model fresh from pre-training does one thing: it continues text. Asked to explain the moon landing to a six-year-old, one such model wrote four more requests of the same kind.

  10. Day 10 Published Sources through about 24 min read 12 min listen 7 figures

    How Preferences Become Behaviour

    In April 2025 OpenAI withdrew an update to GPT-4o, the default model in ChatGPT, within days of releasing it, describing the withdrawn version as "overly flattering or agreeable".

  11. Day 11 Published about 24 min read 11 min listen 6 figures

    What Does a Score Prove?

    On 3 September 2026 OpenAI launched GPT-6 Astra with a page of benchmark tables and a superlative. Beneath the tables sat one line of method: every score shown was the best the model achieved at any effort setting.

  12. Day 12 Published about 29 min read 12 min listen 7 figures

    Why Models Think Longer

    A useful way to read this year in commercial artificial intelligence is that its most consequential change was not a model but a parameter.

  13. Day 13 Published about 31 min read 11 min listen 7 figures

    Capability, Compressed or Routed

    A model's parameter count, long the headline figure, has quietly stopped meaning what it used to. In August 2026 the Chinese laboratory Z.ai released two models less than a fortnight apart.

  14. Day 14 Published about 30 min read 11 min listen 8 figures

    What a GPU Is Doing

    Sometime between 1 and 11 September 2026, NVIDIA edited the product page of Rubin, its newest accelerator. Its main figures for the arithmetic AI models use were left as they were.

  15. Day 15 Published about 27 min read 11 min listen 9 figures

    From One Chip to a Cluster

    On 23 September 2026 SemiAnalysis, an industry research firm, published the third edition of ClusterMAX, its rating of the "neoclouds" that rent out AI accelerators.

  16. Day 16 Published Sources through about 27 min read 11 min listen 9 figures

    What One Answer Costs

    On 22 September 2026 Epoch AI, a research group that studies the trajectory of artificial intelligence, published "The plunging price of thought", an estimate that the cost of reaching a given level of performance on its benchmarks has fallen about 47% a quarter since 2023 — roughly thirteen-fold a year.

  17. Day 18 Published Sources through about 27 min read 11 min listen 7 figures

    Price Is Not Cost

    On 17 September 2026 CoreWeave, one of the largest companies renting out AI accelerators, told investors that since the end of June it had signed three-to-six-month contracts at about $40m a year for each megawatt of power its customers' clusters need.

Day 17 is written and not yet available here.

17 of 30 lessons published; 13 still to come. Each lesson is a written article with its own figures, an audio edition, and a printable copy.