In 180 days ChatGPT and his twins passed tests better than Ivy League students (ask the University of Pennsylvania professors), sang better than the stars of the Grammy Awards (Drake and The Weeknd rocked Spotify with their voices) pulled a non-trivial original song), dispensed advice like weind mentors, and wrote articles in seconds that could pass expert editor review. Whoever made it has offered to take a six-month break to see if the situation is really out of control. – Something like when we used to play in childhood and as soon as we started losing we used to find excuses to interrupt the game.
Despite the extraordinary amount of learning produced by the billions of parameters that activate the network, six months ago we were convinced reading or learningHe It is never possible to beat the ability to have a sensitive nature, to have an empathetic outlook, to be able to contextualize information, These qualities are typical of the extraordinary beings that are all kinds of people who have populated the world since the time of sapiens.
But can we be sure yet? Or should we finally surrender to a doomsday scenario in the form of Blade Runner? It should be made clear right away that Generative AI is not limited to a computer and software that does things faster than before. It’s a real technological leap forward for one simple reason: it’s not only faster, but above all it does things differently. In an area hitherto only touched by machines.
in an essay published ineconomist by anthropologists Yuval Harari, author of the best seller «Sapiens», outlines how language has evolved in sapiens compared to other species, The thing that really sets Sapiens apart from its predecessors and other species is its wealth of language. the language gave them gossip (yes, even gossip has isolated a social network from more or less dangerous situations), to create values, to generate culture-like symbols and, perhaps more importantly in the mists of time, of religion. medium that has allowed communities to live in different parts of the world and identify with them.
Harari explains how The New Result of Machines, or the New Paradigm of Generative Intelligence, Based on Neurolinguistic Programming, And machines are learning to use language very well and in just a few months. And it’s not about Siri and Alexa, who, by comparison, seem to be punctual and apprehensive mothers and fathers, telling us things we already know and don’t currently remember. The new AI is “cool”, it’s refreshing. she is creative.
It resembles the pictures that DALL-E produces on command. Therein lies the power of Midjourney’s iconic images, self-produced with some of the input we shared. They are able to simulate an attack on the Pentagon so effectively that, even if only for a few minutes, it knocks down the world’s stock exchanges.
We thought that the strong point of ChatGPT is its ability to learn based on a sophisticated statistical model that is lucky enough to benefit from the ocean of information, the Internet., However, the limit was clear: the information had already been produced and chewed up and was therefore linked to the past. A deep limitation: The car is rational and says interesting things, but is set up like a big rear-view mirror.
Harari got a point. scheduled tribeand the machine learns to speak and communicate and create content and context, eventually it can dodge, explain, stimulate, It infers on the semantic structures of the underlying language. Language and communication: the combination that forms the communities that make up our civic lives. And which therefore makes the technological leap one of the strongest anomalies of the last thirty years.
What if the AI has “hallucinations” as they technically say in the computer science world and starts producing texts and fake news? What happens if generative AI goes unchecked and starts engaging us in compelling and inappropriate ways on topics and issues? That’s why Europe, once again at the forefront of regulations, is setting itself the goal of dynamic regulation. ie based on handrails and not on self-imposed rules,
We May Be Afraid Of The Consequences Of Artificial Intelligence As We’re Learning About It Today, absolutely love team work was one of the triumphs of modern society (2+2 makes 5 or 6 instead of 4 if the team works), harnesses the many minds of the “past” and potentially influences the future and the present by thought Simulates the future that sapiens have created. But it is the sapiens that have so far been able to embrace the challenges and make them their own. Not denying them.
Think about fire and nature in general. For better or worse, every day they confront us with their power. This time we’re dealing with something made by sapiens, It would be strange to be afraid of them. And fear the challenge we have given ourselves. The goal is to be able to dominate him as always. It is happening this time also. We are sure of it. Or rather, at least for now, we think we are.