“AI has many results,” he said. “But what I think is automatic research. When we look at human history, many of this are about making new technologies about humans. When computers themselves can develop new technologies, they look like a very important, Um, infection point.
“We already see that these models help scientists. But when they are able to work on the long horizon – when they are able to set up a research program for themselves – the world will feel meaningful.”
Chen’s models, the ability to work itself for models is key. He said, “I mean, I think everyone has their own AGI’s praise.” “But this concept of autonomous time – only the amount of time in which the model can only make productive progress on a difficult issue without targeting a dead end.
This is a bold ideology – and is far from the capabilities of today’s models. But nevertheless, I was exposed to how Chen and Pachuki made the AGI’s voice almost un -serious. Compare how when I talked to him 18 months ago, how did the suitcase responded. He told me, “This monument is about to scatter the earth.” “There will be one before and after.” What he was making faced by the deprivation of what he was making, Suscore changed his career attention by designing better and better models to find out how to control a technology that he believes will soon be more smarter than himself.
Two years ago, Soskoor called him a super -league team as a super -league team, which he would share with John Lake, another researcher at Open AC Safety. The claim was that the team would help to know how to control the fictitious superintendent of a full -fledged part of the open -fired resources. Today, most of the Super Laminat team, including Suscore and Like, have left the company and the team is no longer present.
When the leak left, he said that the reason was that the team was not given help, he felt that he deserved it. That posted Yo on x: “Making more human machines than humanity is a naturally dangerous effort. Openi is giving a great responsibility to the whole of humanity. But in recent years, security culture and process have backed shiny products.” Researchers on the other departure shared similar statements.
I asked Chen and Pachaki what they do to such concerns. Chen said, “Many of these things are very personal decisions.” You know, a researcher can happen in a way, you know. “
He resumed. “They may believe that this field is being developed in a particular way and that their research is about to end and the fruit is about to fall. And, you know, you may not be the way the company you want, it will not restructure it. It is a very dynamic field.”