“When we created ChatGpt, we didn’t think it would be this great.” She admitted it – “with a certain embarrassment” – Ilya Sutskeverco-founder and scientific director of OpenAIthe San Francisco company that unveiled its own to the world a year ago ChatGptthe generative artificial intelligence capable of expressing itself like a human being.
Sutskever, formerly Google, is one of the keys to OpenAI’s success. His name appears in the documents that led to the birth of both ChatGpt and Give herOpenAI’s generative AI that is used to produce any type of image, from realistic photos to illustrations.
Artificial intelligence Dall-E 3, the guide: how the ChatGpt tool works to create images by Pier Luigi Pisa 16 October 2023
But at first Sutskever didn’t believe ChatGpt would do any good. “When we asked her a question based on facts, she gave you the wrong answer – the computer scientist told the MIT Technology Review magazine -. And so I thought it would be an insignificant technology, that people using it would say to me: ‘Why do you work on it, it’s so boring’”.
Instead ChatGpt amazed the world, conquering one hundred million active users in just two months. “There is a before and an after ChatGpt” said Sutskever himself, who now works in a company valued at around 86 billion dollars (esteem of the Financial Times).
Part of this success, at least at the beginning, is due – according to Sutskever – “to a spiritual experience”. “It’s the moment people tried ChatGpt for the first time and said, ‘Oh my God, it looks like this computer can understand me.’
The reality, of course, is different. ChatGpt’s creativity is only apparent. In fact, generative AI does not truly understand the intrinsic meaning of what it generates. In fact, when ChatGpt writes a text, it chooses each time which word is most likely to appear after another in a given context.
In the eyes of many, what ChatGpt can do is pure magic. But Sutskever wasn’t the only one at OpenAI who thought this technology wasn’t all that extraordinary. Also Greg Brockman, one of the company’s co-founders, told Forbes that “none of us were that enamored with it at first. Nobody thought it would be really useful.”
So much so that Brockman and his colleagues, in the months preceding the launch of ChatGpt (which took place on November 30, 2022), they were still working on possible alternatives.
Brockman, speaking to Fortune magazine last January, revealed that the decision to make ChatGpt public represented the real turning point for OpenAI. Because real users have contributed, with their questions and feedback, to solving numerous problems, including that of the ChatGpt beta testers that “they didn’t know what to ask her”. “Before the official launch – said Brockman – I was among the ranks of those who weren’t sure it would work”.
A year later, the OpenAI team is aware that they have it in their hands a creature straight out of a science fiction film. So effective that it triggered fears for the future of humanity. The CEO of OpenAI, Sam Altmanis among those who fuel the apocalyptic vision according to which artificial intelligence – in the absence of strict rules – could lead to the extinction of man.
Artificial intelligence ITW 2023, Sam Altman speaks to Italy for the first time: “ChatGpt will make us better” by Pier Luigi Pisa 28 September 2023
For his part, Ilya Sutskever believes that in the future humans they could choose to “be part of AI” to compete with super intelligent machines.
“Will there also be room for humans in a world with more intelligent AI?” Sutskever told the magazine MIT Technolgy Review.
“One possibility, something that might be crazy by today’s standards, but won’t be crazy by future standards, is that many people choose to become part of artificial intelligence,” added the OpenAI computer scientist.
“At first only the most daring people will try to do it,” Sutskever said. Perhaps others will follow. Or not”.