Sutskever became an AI superstar, co-authoring a breakthrough paper that showed how AI can learn to recognize images just by being exposed to large amounts of data. He ended up happily as a lead scientist in the Google Brain team.
In mid-2015 Altman cold-emailed Sutskever to invite him to dinner with Musk, Brockman and others at the swank Rosewood Hotel on Sand Hill Road in Palo Alto. Only later did Satskever realize that he was the guest of honor. “That was a common conversation about AI and AGI in the future,” he says. More specifically, he discussed whether “Google and DeepMind were so far ahead that it was impossible to catch them, or whether it was still possible to create a counterbalancing lab, as Elon put it.” No one overtly tried to recruit Sutskever at dinner, but the conversation attracted him.
Sutskever soon wrote an email to Altman saying he was game to lead the project – but the message got stuck in his drafts folder. Altman bounced back, and after months of fending off Google’s counteroffer, Sutskever signed on. He would soon become the soul of the company and its driving force in research.
Sutskever joined Altman and Musk in recruiting people to the project, culminating in Napa Valley where many potential OpenAI researchers encouraged each other. Of course, some targets will resist the bait. John Carmack, the legendary gaming coder behind doom, EarthquakeAnd countless other titles, an Altman pitch rejected.
OpenAI was officially launched in December 2015. At the time, when I interviewed Musk and Altman, they presented the project to me as an effort to make AI safe and accessible by sharing it with the world. In other words, open source. OpenAI, he told me, will not apply for a patent. Everyone can use their success. Some future Dr. Wouldn’t that empower Evil? I was surprised. Musk said that’s a good question. But Altman had an answer: Humans are generally good, and while OpenAI will provide powerful tools for the majority, bad actors will be overwhelmed. He admitted that if Dr. Evil used tools that could not be resisted, “so we’re in a really bad place.” But both Musk and Altman believed that a safer path for AI would be in the hands of a research operation unpolluted by discovery missions, the constant temptation to ignore the needs of humans in pursuit of boffo quarterly results.
Altman cautioned me not to expect results anytime soon. “It’s going to look like a research lab for a long time,” he said.
There was another reason to lower expectations. Google and others have been developing and implementing AI for years. While OpenAI had a multi-billion dollar commitment (mostly by Musk), a great team of researchers and engineers, and a lofty mission, it had no clue how to pursue its goals. Altman remembers a moment when the small team gathered in Brockman’s apartment—they didn’t yet have an office. “I was like, what should we do?”
A little more than a year after OpenAI’s founding, I had breakfast with Brockman in San Francisco. For the CTO of a company with words open up As his name suggests, he was pretty transparent with the details. He confirmed that the nonprofit will be able to draw on its initial billion-dollar donation for some time. The salaries of 25 people on its staff — who were paid well below market rates — ate up the bulk of OpenAI’s costs. “The goal for us, what we’re really pushing,” he said, “is to have systems that can do things that humans couldn’t do before.” But back then, it seems that there was a group of researchers who published research papers. After the interview, I took him to the company’s new office in the Mission District, but he wouldn’t let me past the vestibule. He went into a closet to get me a t-shirt.
If I had gone in and asked around, I would have known exactly how much OpenAI is was Brockman admits now that “nothing was working.” His researchers were throwing algorithmic spaghetti at the ceiling to see what stuck. He invented systems that solved video games and put a lot of effort into robotics. “We knew what We had to,” Altman says. “We knew why We wanted to do it. But we had no idea how.”