The Tesla Side Hustle That Nearly Broke OpenAI
Court testimony from Greg Brockman has exposed a messy, little known chapter in AI history: while building what was supposed to be a nonprofit dedicated to safe AGI, Brockman and three other OpenAI engineers were secretly moonlighting for Elon Musk’s Tesla. Musk personally requested their help on self driving technology, and Brockman admitted in court that it felt less like a favor and more like an order. ‘It was pretty clear this was not something we could say no to,’ he testified, revealing a power dynamic that undermines OpenAI’s founding narrative of independence.
This double duty wasn’t a brief detour. The OpenAI crew spent months working on Tesla’s autonomous driving systems, and one of them, Andrej Karpathy, eventually left to run Autopilot full time. Musk even offered a theatrical apology in court: ‘I have an apology and a confession. I made an offer to Andrej to run autopilot and he accepted.’ The confession is revealing: Musk was willing to cannibalize a nonprofit he helped fund for his car company’s gain.
The Hypocrisy of ‘Open’ AI
This revelation shreds any remaining pretense that OpenAI was ever truly independent. The same organization that now preaches about safety and alignment was, at its inception, a talent farm for Musk’s commercial ambitions. Brockman says he worried about this conflict from day one, yet he participated anyway. That’s not principled AI governance; that’s career survival in Musk’s orbit.
And the pattern continues. OpenAI’s current legal battle with Musk is less about the future of AI and more about who gets to cash in. Musk now claims the nonprofit he once bled dry for Tesla talent has betrayed its mission. He’s not wrong, but he’s hardly a clean advocate. This testimony proves Musk saw OpenAI as a resource to be exploited, not a mission to be protected. The entire saga stinks of ego and opportunism, with the public left to wonder if any major AI lab can truly operate without hidden strings.
Source: Theverge
