1. Looms put weavers out of work, copying machines put typists out of work. However, the number of jobs only increased over time – yesterday’s weavers and typists found new paid occupations. The problem with the current wave of job cuts is that it’s unclear exactly where the average person can reliably outperform a neural network. AI is already barely capable of performing almost all “mental” work: both technical and humanitarian. As funny as it may seem, people are clearly superior to robots only in physical work – in installing air conditioners or repairing cars, for example… We still don’t feel the problem well, since we are in a somewhat isolated segment of the Internet, and in the West people are already in a panic. Quizaz shows a story about mass layoffs published on the BBC (link 1, link 2): Writer Benjamin Miller (not his real name) was thriving in early 2023. He led a team of over 60 technology blog writers and editors. company. “It was a really fun job,” says Miller, “an opportunity to flex my creativity and collaborate with experts on a variety of topics. But one day Miller’s manager told him about a new project. “They wanted to use AI to reduce costs.” he says (Miller signed a non-disclosure agreement and asked the BBC not to reveal his name or the name of the company). A month later, the factory installed an automated system that inserted the article’s title into the online model. would create a chart based on that title, and Miller would receive a message on his computer. Instead of coming up with their own ideas, its writers created articles based on those ideas, and Miller did the final editing before publishing the stories. adapt before learning about the second level of automation. ChatGPT later wrote entire articles and most of his team was fired. The few remaining people were left with an even less creative task: editing the poor ChatGPT text to make it sound. more humane. “Suddenly I was just doing everyone else’s job,” Miller says. Every day, he opens documents written by AI to correct routine robot errors, completing a task that dozens of people previously worked on. By 2024, the company had fired the rest of Miller’s team, leaving him alone. 2. I regularly read skeptical comments from commentators who, several years ago, poked their fingers at neural networks and became convinced of their imperfection. No, colleagues, this industry is not developing as slowly as caring leather bags would like. For example, here is a visual video comparing how the AI generated videos a year ago and how it does it now (link). If a year ago it was clear that something was clearly wrong here, now the result is almost impossible to distinguish from a regular TikTok video. A brave new world is gradually opening up before us, in which a robot can generate a video in which a person of a given appearance will perform given actions. It’s clear what scope this opens up for mischief. Over time, the duration of high-quality generations will increase, and their plots will become more complex. Develop a symbol system for friends and family if you don’t want to be surprised by scammers. 3. OpenAI, the developer of the famous ChatZhPT, is officially classified as an intelligence agency (link). Paul Nicasone, a former director of the US National Security Agency (NSA), will now be responsible for cybersecurity on OpenAI’s board of directors, according to a company blog post. The NSA is a spy organization that spies on Internet users around the world (link). The NSA is monitoring you through your computers and smartphones, both through popular US-developed apps and through software vulnerabilities in iPhones and similar devices. I’m not sure the NSA thinks artificial intelligence is the superweapon of the future. However, foreigners trust ChatZhPT with so much secret information that for professional observers it is a real Klondike. Oleg Makarenkohttps://dzen.ru Subscribe to our Telegram channel so as not to miss all the most important materials that we publish: https://t.me/russiapost
0 комментариев