Let's define some terms. AI, Artificial intelligence (Turing, 1950): When a computer can convince a human that it is human 30% of the time after 5 minutes of questioning by text messages, given an a-priori probability of 50%. That happened in 2023.
AGI, Artificial general intelligence: when a computer can do everything that humans can do. We know AGI is not yet solved because people are still getting paid to do work that machines aren't smart enough to do. ASI, artificial super intelligence: machines that are smarter than humans. That depends on the definition of intelligence. If you mean the Turing test then ASI is impossible because the highest possible score is human level. If you mean Legg and Hutter universal intelligence as goal achievement over a universal distribution of environments, then we need a computable approximation of that. A reasonable measure might be dollars per hour. That happened about 1800 when world GDP doubled from the medieval baseline due to technology. If you mean when computers surpass an IQ of 100 then we don't know because IQ is actually a scalar approximation of a weighted set of capabilities where the weights are set by humans with ulterior motives. For example, we weight math and spatial reasoning (where males score higher) and language skills (where females score higher) so that both sexes average to 100. We don't know how to score a machine that is a billion times faster than humans or has a billion times better recall at certain tasks but is completely unable to do others. Most people assume ASI will come after AGI, but if ASI means doing everything as well as a human and at least one thing better, then both will happen at the same time because computers have been doing some things better since the 1950s. Now to answer your points. 1. Intelligence is potentially a threat because prediction measures intelligence and two agents can't mutually predict each other by Wolpert's law. 2. AI is already automating human labor by modeling human minds, even if you don't think so. You do realize that we live in a system of pervasive surveillance to provide training data for AI? 3. A goal seeking agent by definition seeks to maximize utility. We live in a finite universe. Therefore all possible utility functions have a maximum. The best it can do is stay in that state. That is the equivalent of death. -- Matt Mahoney, [email protected] On Sun, Jun 21, 2026, 5:41 PM twenkid <[email protected]> wrote: > Matt, write your paper as you wish, the style you want and include what > you want, e.g. your version with your abstract. "Just because I quote" - > the quotes are an organic part and a trigger of the rest - I do credit > other researchers' contribution or related work, unlike what they do > regarding my work, preceding theirs with from several years to up to 20-25 > years and still counting. > > > *"The actual threat of AI ..."* > My interpretation, as in the paper, is that AI, thinking machines or the > proper-intelligence (cognition, general intelligence) *is/are not the > threat.* > > On the other hand, the homo sapiens are a historically proven "threat" > without any doubt - many of their leaders and the leaders their > "sub-leaders" supported were monsters, cheaters, liars and sadists who are > a threat both for their own "species" and for the thinking machines (BTW, > one German and his democratic and progressive nation and corporations were > trying to protect the humanity and "species", that is why they exterminated > millions, did you forget. He took lessons from his cousins and recent > compatriots from an island near-by, and a continent over the ocean). > > > > *"Automating human labor requires modeling 8 billion human brains"*It > doesn't and "economically valuable work" is a joke (see the footnote etc.) > > (...) > > *"The actual threat of AI is that if you got everything you wanted, then > you would be dead.* > > * "*Yes, e.g. you could become a cyborg :))), or the machines will be the > better "humans'", and the "cyborg" might "be death", according to some > legal system of the species of the AI-aligners like "John" Connor, who are > protecting the humanity from the terminators. (...) > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + > delivery options <https://agi.topicbox.com/groups/agi/subscription> > Permalink > <https://agi.topicbox.com/groups/agi/Teedaf4a19e56877e-M54366152cc3806d424fec416> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Teedaf4a19e56877e-Mfe0365bee4a78367e90362ec Delivery options: https://agi.topicbox.com/groups/agi/subscription
