ChatGPT has taken the world by storm since its launch last November. STOP: Sam Altman has over a hundred million people using Chat GPT, but now development is to be paused. Photo: AP A few months later, the service had over 100 million active users, making Chat GPT the fastest growing application in the history of the Internet. On March 14 this year, the newest edition, Chat GPT-4, arrived. It’s so impressive that it prompted 500 of the most influential tech experts in the world to send a letter of concern, asking the company behind it to pause development of Chat GPT. In the letter, Elon Musk, Steve Wozniak, Stuart Russell and the others write that the development of technology is going too fast, creating “great risks for society and humanity”. In a new lecture at MIT, Open Ai CEO Sam Altman comes with surprising news: the company is stopping development on the popular service. – In an early version of the letter, it was stated that we are training GPT-5 right now. We don’t do that, and we won’t do that for quite some time, says Altman. Chat GPT artificial intelligence assistant, programmed to help answer questions and solve problems developed by Open Ai, founded by Sam Altman and Elon Musk in 2015 uses machine learning of language models to produce text has no information after 2021 speaks Norwegian Bigger is not necessarily better HAPPY: Ki researcher Inga Strümke has just released a book about artificial intelligence, in which she warns, among other things, about where technology can take us. Photo: Ine Julia Rojahn Schwebs / news – I was very pleased with the action from Sam Altman. As a researcher on artificial intelligence, it is extremely unsatisfying that a system gets better just because it gets bigger. That’s what Ki researcher Inga Strümke says. She says it is enormously resource-intensive to create a better version of Chat GPT than the one that already exists. – Altman probably thinks that we cannot expect a GPT-5 that is even bigger. They think that there is little to be gained by simply making the models bigger. There are a lot of AI developers working on this, it requires a lot of resources. In addition to humans, Ki development costs enormous amounts of power. – The climate footprint of training a large language model is completely perverse. We are talking about so many human lives lived, with all the meat we eat and the flights we take. Atomic bombs for all Gigantic leaps in technology have occurred several times in the course of human history. But the pace we are seeing now is something completely new, believes Strümke. – Imagine if the same had happened with the Manhattan Project. The US does some research on nuclear weapons, and three weeks later everyone can get their own nuclear bomb. The short path from research to publication means that we don’t have time to concentrate and think a bit. She is not alone in being worried. On Monday, Geoffrey Hilton left Google, the man who is referred to as the “godfather of artificial intelligence”. – I am leaving Google to be able to speak freely about the risk Ki poses, Hinton told the New York Times. Open Ai, for its part, has no plans to stop leading the world’s research into artificial intelligence. Although Sam Altman will not develop a new Chat GPT, the research does not stop. – We will not make larger models. We will make them better, says Altman.
ttn-69