Hatvani Péter 💻

Author: P. Hatvani

  • The Unreasonable Efficiency of Underlying Systems

    —

    by

    P. Hatvani
    in

    This is the story of why the current AI boom was inevitable from the very core of engineering and information technology. If you think about it, AI was inevitable, it all boiled down to having an underlying system supporting faster development. It began a long time ago, in a time when engineers were designing physical…

  • Holograms and Washing Machines

    —

    by

    P. Hatvani
    in

    Did you see the white round, shapeless robot that is cleaning, teaching or just thinking? Or the holographic supernova of a brain and head mixed in the supersonic blueness? I was wondering why all AI visualizations were so similar, this is my discovery of the topic. This image is the main visual for the wikipedia…

  • Lazy Jack, the Most Human Character

    —

    by

    P. Hatvani
    in

    You might have heard that monkey see, monkey do. In reality monkey only does if they see the why. A not-so-recent experiment showed that overimitation in humans is quite prevalent. More on this in this article by the New York Times. In this research children were shown a complicated contraption and a long process on…

  • A Library You Can Ask

    —

    by

    P. Hatvani
    in

    The Magyar Elektronikus Könyvtár — MEK, the digital library of the National Széchényi Library — is one of those quietly enormous public goods: thousands upon thousands of Hungarian books, free to anyone. You can browse it, search it, download from it. What you can’t do is ask it anything — point at the whole collection…

  • This is the Softest Software has ever been

    —

    by

    P. Hatvani
    in

    Uncle Bob has this punchline: “Software” is a compound word = “soft” (changeable) + “ware” (product). “If it’s hard to change, you have just reinvented Hardware.” Uncle Bob means soft as a virtue, the kind you earn. Code stays changeable because someone had the discipline to keep it that way. Which is where my train…

  • Benchmarking Visual Language models for OCR

    —

    by

    P. Hatvani
    in

    This is the full write-up on our experiment with Gábor Madarász. Digitizing books is a longstanding problem in the NLP field. It is an even harder problem in low-resource language such as Hungarian. We have tested in a short experiment how far the latest generation of Qwen models (3rd generation) has come. We will first…

  • Four Filters and a Human: An On-Prem GPT-4 Pipeline for Freight Email

    —

    by

    P. Hatvani
    in

    Four Filters and a Human Freight coordination, stripped down, is a matching problem buried under email. An offer lands in someone’s inbox — a load that needs to go from here to there, by a certain day. Somewhere out there is a contractor with a truck the right size, near the right area, willing to…