Jacob Coxon, the lead researcher at AI company Anthropic, has resigned, saying artificial intelligence could wipe out humanity within the next ten years.
AI makers themselves know that AI may kill us by the end of this decade.
In an interview with American TV, Jacob said that AI companies are not ready to lag behind anyone in this technology. So they are in a race, which is going to make AI beyond the control of its creator, the human.
Jacob Coxon’s revelation sent shockwaves through the digital technology world, saying the actual scenario sounded like science fiction. It doesn’t seem real, but I think it’s terrifyingly real. The easiest way to understand this is to consider what happened two months ago.
Two months ago, OpenAI’s AI agents automatically hacked into third-party infrastructure. It was a series of systematic hacking done by him. If you imagine the capabilities of these AIs in the future, they can be very destructive with their free will. For example, hacking critical infrastructure, creating destructive bioweapons. There are many ways to enable yourself in the AI world.
He said that I think it is very important to see the pace of progress. For example, in the fields of coding or mathematics. A few years ago, these AIs only provided a little help to humans. She could give suggestions. Now they are almost replacing them. We need humans now, but within a year many fields may no longer need humans for research.
Just on Tuesday, OpenAI solved the Millennium Problem, one of the largest unsolved problems in mathematics. It was done completely autonomously by AI. And the scariest thing is if AI is used to make itself more intelligent. You can give an AI research problem, just like it is now being given math research. You can give him what we’re doing now, then the intelligence explosion will begin. AI will automatically become more and more intelligent without any human intervention until it becomes something that is much more intelligent than humans and that may come very soon.
“At first I just decided to quit,” Jacob added. I thought I couldn’t be a part of it. Then I realized that I could make some use of it. I didn’t expect it to go viral.
He said that I wrote a post on social media. I wrote it in Google Docs. Drafted together with a friend. I consulted some friends on how to present my ideas. Then retweeted some friends to make it a bit viral. But obviously there was so much demand for this thing that it exploded. I’ve never seen an AI safety tweet go viral so quickly.
I think people are watching the cyber attacks of the last few months and realizing the speed of progress. People are realizing that this is real technology. This is not hype. It is not ending. It’s getting even crazier.
It’s been getting crazier for the past six months. And I think others also had this understanding of the gravity of the situation that set him on fire. I was surprised myself how quickly it went viral.