Skip to content

Topic · Themes

Understanding AI: all our articles

All 122 Alyriaz articles on the “Understanding AI” topic: news, analysis and explainers, newest first. To stay up to date, you can get an alert as soon as a new article on this topic is published.

Filter in the library

  1. · Society

    The languages the big models ignore

    State-of-the-art models perform poorly in most of the world's languages. We explain why, what it costs hundreds of millions of people, and why local players are stepping in.
  2. · Society

    A court now requires AI use to be declared, after a 40% rise in related cases

    An Australian administrative court will make it mandatory to declare the use of AI from October 2026, after noting a 40% rise in related cases. We analyse the measure and what the figures reveal.
  3. · Explainers

    MIT, Apache, "open": how to read a model's licence

    MIT licence, Apache 2.0, restrictive in-house licences: we explain the concrete differences between the legal regimes of so-called open models, and what to check before building on them.
  4. · Models

    It's happening today: the Sonnet 5 price increase takes effect, and it's steeper than 50%

    The promotional rate for Claude Sonnet 5 ends today. We explain why the real increase exceeds the advertised 50%, and what you should concretely check in your costs.
  5. · Explainers

    Why doesn't the same AI respond differently depending on the app?

    The same model produces very different responses depending on the application using it. We explain the five factors behind this gap, and why it's useful to know.
  6. · Vision

    What will remain of us?

    It is now technically possible to recreate the voice, handwriting and mannerisms of a person who has died. We examine what this possibility entails, for those left behind as well as for the one who is no longer here.
  7. · Explainers

    Scaling laws: why "bigger" worked, and why it's slowing down

    Scaling laws describe a predictable relationship between size, data, compute and performance. We explain what they actually say and why the industry is now looking for other levers.
  8. · Explainers

    What is distillation? Fitting a large model into a small one

    Distillation transfers the capabilities of a large model to a much smaller one. We explain the principle, why it works better than direct training, and its limitations.
  9. · Society

    OpenAI publishes the first results from its in-house chip

    OpenAI has published the first results from its in-house inference accelerator, touted as delivering strong speed and efficiency. We analyse what this move reveals about the balance of power in infrastructure.
  10. · Vision

    Can you delegate your memory?

    Entrusting one's memory to a system that retains, summarises and retrieves raises a different question from that of the notebook or the search engine. We examine what memory does and what is lost in delegating it.
  11. · Explainers

    How can you tell if a text was written by an AI?

    Between the myths about em dashes and unreliable detectors, here's a look at what actually helps spot generated text, and why caution is still warranted.
  12. · Models

    Nvidia is putting into production a chip that does only one thing: respond fast

    Nvidia has announced that its inference accelerator, born from its acquisition of Groq, is entering production. Here’s why a chip specialised in responding rather than training is a significant signal.
  13. · Explainers

    General-purpose or specialised model: which one should you choose for your job?

    Faced with the vertical models emerging in law, healthcare or finance, we explain when a specialised model genuinely adds value and when a well-equipped generalist is enough.
  14. · Explainers

    Dense model or mixture of experts: two philosophies, two use cases

    A dense model uses all its parameters for every request, while a mixture of experts only activates a fraction of them. We explain why this architectural choice determines the possible use cases, particularly on-device.
  15. · Explainers

    How does an AI find a security flaw?

    Understanding how a model identifies a software vulnerability sheds light on both its capabilities and its limitations. The mechanism is explained without providing any exploitable details.
  16. · Society

    A library calls for volunteers to save the books AI is having destroyed

    A major online library is calling for volunteers to digitise and preserve physical books, as AI companies buy and destroy them to build corpora. We analyse this response and what it reveals.
  17. · Society

    One link was enough to leak your emails: the Copilot flaw that makes agentjacking a reality.

    Microsoft has patched a vulnerability that, via a simple malicious link, could automatically trigger instructions in Copilot and exfiltrate connected data. We break down the mechanism and the lessons to be learned.
  18. · Explainers

    AI, machine learning, deep learning: what contains what?

    Artificial intelligence, machine learning, deep learning, generative AI: these terms refer to nested sets, not synonyms. Let's clarify once and for all, with concrete examples.
  19. · Explainers

    What is homomorphic encryption? Computing on data without ever reading it

    Homomorphic encryption makes it possible to perform calculations on encrypted data without ever accessing it in plaintext. We explain the principle, its current limitations, and what it could change for AI and privacy.
  20. · Models

    DeepSeek raises prices by 93%: has the price war hit rock bottom?

    DeepSeek raised the price of its Flash model by 93% on 14 August, while other players are still cutting theirs. We analyse what this reversal reveals about the real economics of inference.

Other topics in “Themes”

All the other topics

Every article, month by month →