New: my book on artificial intelligence, from its history to agentsDiscover it →
AI trainingAdvertise
Mattia Panatta
IT

Opinions

Where it's going. And what I think.

The guides explain facts; here I take a position. Every opinion is labelled as such, argued with data and sources, and written to be challenged.

  1. A humanoid robot sitting in the lotus position
    FutureAGI is already here. Now the question is superintelligenceWith a concrete definition, systems that handle most cognitive work and use tools to reach a goal, I think we have already reached AGI. The direction now is superintelligence. What makes me think so, what remains open and what changes for us.22 Oct · 7 min
  2. Close-up of a white robotic hand
    TechnologyFrom questions to delegation, the era of always-on agentsGrok Bot, Claude Duet, ChatGPT dots, Muse, OpenClaw: agents have a computer of their own, work while you sleep and organise into teams. The bottleneck is no longer technology: it's trust.22 Oct · 6 min
  3. A row of Venetian masks on a black background
    SocietyEvery chatbot has a character (and none is neutral)Claude has a "constitution", ChatGPT lets you dial warmth and enthusiasm, Grok publishes its instructions after several scandals, Meta had to fix its bots' rules. Assistant personality is designed: how it is made, why it matters and how to guard against flattery.22 Oct · 7 min
  4. Stock charts on a monitor during a trading session
    EconomyIs AI a bubble? It depends where you lookRevenue growing like never before, but also debt-financed data centres and money going round in circles. Why I wouldn't talk about a single bubble: some parts of the industry are solid, others fragile. And what it means for people who use AI every day.22 Oct · 7 min
  5. A condenser microphone with a pop filter in a recording studio
    CultureI made seven songs with AI. Who is the author?For AI Radio I explained to the AI what I wanted to achieve and Suno wrote, composed, arranged and sang. The result is surprising and a little unsettling. What I learned about creativity, rights and transparency.16 Oct · 6 min
  6. A person holding a smartphone with their face in their hands
    SocietyArtificial friends: the AI that keeps us companyMillions of people talk to an AI every day as a friend, a confidant, sometimes a partner. Why it happens, where it can help and where it becomes a risk, especially for young people.15 Oct · 6 min
  7. Coins falling onto a table
    EconomyThe price of intelligence is collapsing. And that is not (only) good newsA task the best model solves for 6 dollars, another does for 72 cents and an open model for 13. What it means when intelligence becomes a cheap commodity, for those who use it and those who sell it.14 Oct · 6 min
  8. A person working on a laptop
    WorkAlone with AI: can one person be worth an agency?I built this site on my own, with artificial intelligence: copy, design, chatbot, music, payments. What I learned about what one person can do today, and what they still cannot.13 Oct · 6 min
  9. A small agent of light pushing a shopping cart between digital shelves
    EconomyWhen the agent does the shopping, who really chooses?ChatGPT dots, Grok Bot and Meta Muse can already search, compare and buy for us. Amazon blocked Muse after two weeks. What changes for shoppers, shops and small businesses.11 Oct · 7 min
  10. Glowing sound waves coming from a person and replacing the keys of a keyboard
    TechnologyVoice will be the new keyboard (and that is not good news for everyone)New voice models listen and speak at the same time, let you interrupt, pick up tone. Talking to AI will become as normal as typing to it. What changes, who gains and where the risks are.10 Oct · 6 min
  11. High-voltage pylons at sunset
    EnvironmentData-centre water is a distraction. Energy isn't.A Gemini answer uses less energy than nine seconds of TV. Yet data centres will double their consumption by 2030. How to hold both facts together, without guilt and without naivety.9 Oct · 6 min
  12. An open door letting light into a dark room
    TechnologyOpen vs closed models: the game that matters more than the leaderboardThe best models are closed, but open ones trail by a few months and cost a fraction. Why, for businesses and professionals, this choice matters more than who tops Arena.8 Oct · 6 min
  13. Two colleagues at a computer in an open-plan office
    WorkAI won't steal your job. It takes away the first rung.The most serious studies say AI automates tasks, not jobs. But the tasks that disappear first are the ones young people used to learn the trade. I say this as someone who has just started working.6 Oct · 7 min
  14. A person reading a book in the sunlight
    SocietyIs AI making us dumber? It depends on what we let it doFrom Socrates against writing to GPS against our sense of direction: every technology "takes something away". With AI the risk is more serious, because it touches thinking itself. But it isn't inevitable.5 Oct · 6 min
  15. EU flags in front of the Berlaymont building in Brussels
    GeopoliticsEurope can't just write the rules of AIThe US and China build the models, Europe writes the laws. Rules matter, but alone they aren't enough: where I think Europe (and Italy) can really play its game.4 Oct · 6 min
  16. Hands using a calculator at a desk
    ProfessionsThe accountant of 2030 won't log invoicesInvoices will log themselves, tax returns will be pre-filled, research will be done by an agent. What will be left for accountants? In my view the best part of the job. But only for those who prepare now.1 Oct · 6 min
  17. A person in the dark scrolling a smartphone
    SocietyIn the age of AI slop, judgement becomes the scarcest commodityThe web is filling with generated content, all similar and all plausible. Why I think this makes signatures, verification and a recognisable voice more valuable, not less.30 Sept · 6 min
  18. Students writing in a classroom
    EducationIn schools, copying with AI isn't the real problemStudents hand in AI-written homework, teachers mark it with AI. Bans don't work and detectors get it wrong. The real problem is what we assess, and how.29 Sept · 6 min
  19. Waves crashing on a red lighthouse in a storm
    RisksHow worried should we be about AI, without panic and without naivetyBetween talk of extinction and talk of overblown alarmism, I try to separate today's concrete risks from tomorrow's speculative ones. Both deserve attention, in different ways.28 Sept · 7 min