Sam Altman says OpenAI still focuses on social good, despite abandoning nonprofit statusOctober 3, 2019 View comment (1)
This article is part of our ongoing coverage of TechCrunch Disrupt.
SAN FRANCISCO—When OpenAI was formed in 2015 as a nonprofit, it had lofty goals for the advancement of artificial intelligence.
But, earlier this year, the research and development group—co-founded by Y Combinator's then-president Sam Altman and former Stripe CTO Greg Brockman—converted into a for-profit corporation that would pay out returns to investors. Leaving behind its nonprofit status, how dedicated would OpenAI remain as a money-making venture to its original mission of advancing AI for the benefit of society?
On Thursday, Altman and Brockman were on stage at the TechCrunch Disrupt conference to reaffirm the intentions they committed to back in 2015 along with fellow co-founder Elon Musk.
Altman said that under the new corporate status, he could theoretically "choose to make [OpenAI] worth nothing" to prove the company's work is more important than profit.
"I think the technology we're working on has so much promise if it works and if we're successful in our mission," he said. "It's not going to be a particularly difficult thing for us to provide a great return to our investors and also unprecedented value to the world."
Altman also pointed out that San Francisco-based OpenAI has limited first-round investors to a maximum return of 100x their investment, and each successive round will have progressively smaller return caps to gradually allow for more cash to flow in to the company's development.
He said that while a return of 100x an initial investment might seem like an enormous potential profit, it's relatively small by venture capital industry standards.
An influx of capital may be necessary to fund development of the company's highly speculative projects. OpenAI is working on artificial general intelligence, a subset of AI that seeks to make machines able to mimic the way humans learn and grow. By contrast, the conventional AI model centers on technology performing preprogrammed intelligent-yet-repetitive tasks—such as an AI chatbot serving the same types of responses to a given set of phrases.
Today, highly sophisticated AGI is largely science fiction that can evoke imagery of an apocalyptic future involving robots overtaking humans. Brockman counters that this depiction screams of ignorance and pays no attention to gradual, benevolent accomplishments being made.
"I think that AGI is something where there's been 60 years' worth of debate on when this thing's going to come," Brockman said. "But I think one thing that's also really important is that along the way to AGI, it's not quite like making a singular breakthrough and either you got it or you don't. We're producing these super-valuable AI technologies that are becoming more and more powerful."
Altman said OpenAI's long development time-horizon ultimately meant it had no choice but to abandon the nonprofit model if the company intends to survive.
"In spirit, we still view ourselves as working for the world," Altman said. "But I think you have to sort of play the field as it lies. And we just need so much capital to do our work. I think we'll need more capital than any nonprofit has ever raised, probably, and so this was sort of a reflection of reality."