It looks like Elon Musk is increasingly unhappy with OpenAI, the artificial-intelligence research firm he helped found five years ago.
Microsoft announced on Tuesday that it was exclusively licensing GPT-3, a natural-language AI-powered tool made by OpenAI.
The announcement was met with some dismay on Twitter from users who had thought OpenAI's mission statement was to make technologies like GPT-3 widely available. Musk, who cofounded the company in 2015 as a nonprofit AI research body, was among those who criticized the deal.
"This does seem like the opposite of open. OpenAI is essentially captured by Microsoft," he said.
This does seem like the opposite of open. OpenAI is essentially captured by Microsoft.— Elon Musk (@elonmusk) September 24, 2020
Microsoft Chief Technology Officer Kevin Scott said in a blog post the company was licensing the software to "leverage its technical innovations to develop and deliver advanced AI solutions for our customers, as well as create new solutions that harness the amazing power of advanced natural language generation."
It's not clear what its commercial uses might be, though Scott hinted it could be used to build translation tools and even help people write.
It is also unclear how much exclusivity this license gives Microsoft. In his post, Scott said OpenAI would continue to offer access to GPT-3 via its application programming interface.
OpenAI reiterated this in its own blog post, saying "the deal has no impact on continued access to the GPT-3 model through OpenAI's API, and existing and future users of it will continue building applications with our API as usual."
A Microsoft spokesperson told The Verge the deal gave Microsoft exclusive access to GPT-3's underlying code.
Musk's disdain for Microsoft might be related to his ongoing sparring with its cofounder Bill Gates, though this also isn't the first time he has openly criticized OpenAI.
In February, he tweeted that his confidence in the company's approach to safety around AI was "not high." Though Musk was the company's cochair at its founding in 2015, he said in February that he now had "no control" and "limited insight."
GPT-3 has made headlines because of its ability to generate text that reads as though it were written by a human. When it was unveiled in February 2019, OpenAI said the software wouldn't be open source for fear it could be misused. For example, it could be used to generate large quantities of fake Amazon reviews. It launched the tool in a private beta in June.