OpenAI, the artificial intelligence research company, has issued a statement addressing claims made by its co-founder Elon Musk and reaffirming its commitment to ensuring advanced AI benefits all of humanity.
There were some significant disagreements and challenges in the relationship between Elon Musk and the OpenAI team over the organization’s mission, structure, and funding path.
What?
In the statement, OpenAI co-founders Greg Brockman, Ilya Sutskever, John Schulman, Sam Altman, and Wojciech Zaremba reveal details about their split from Musk and plans to “dismiss all of Elon’s claims” against the company.
According to OpenAI, as they realized building artificial general intelligence (AGI) would require far more resources than initially anticipated, Musk wanted OpenAI to merge with Tesla or be under his full control. When OpenAI refused those terms in 2018, wanting to avoid any individual having “absolute control,” Musk departed and said he would build an AGI competitor himself.
The statement claims Musk told them their “probability of success was 0” and that he would need “billions per year immediately or forget it” to build safe AGI. However, OpenAI says it is advancing its mission by creating “widely-available beneficial tools” like GPT-4 that are empowering people globally.
OpenAI expresses being “sad” about the rift with Musk, someone they “admired,” but states they are “focused on advancing our mission and have a long way to go” in developing safe and beneficial AGI systems.
OpenAI’s Mission and Early Collaboration with Elon Musk
Musk initially wanted to announce a $1 billion funding commitment for the non-profit OpenAI, which far exceeded what was actually raised from donors.
As OpenAI realized the immense compute resources needed for AGI, they determined a for-profit entity would be required to raise sufficient capital. However, Musk wanted majority control and to be CEO, which OpenAI declined.
Musk suggested merging OpenAI into Tesla as a “cash cow”, but ultimately left OpenAI in early 2018, expressing skepticism about their probability of success versus Google/DeepMind without a “dramatic change” in resources.
OpenAI states their mission is to build safe, beneficial AGI while creating broadly distributed benefits through tools like GPT-4 available today. They aim to be more open earlier but can be “less open” as they get closer to AGI.
There seems to be lingering tension, with OpenAI being “sad” it came to this with someone they admired, and dismissing Musk’s claims against them in court.