Musk v. Altman closing arguments
Today was the final arguments Musk vs. Altman trial, and I almost feel bad writing about the incredible demolition derby I just witnessed. Steven Molo, Musk’s lawyer, stumbled over his remarks. At one point, he called Greg Brockman – a co-defendant – Greg Altman. He falsely claimed that Musk was not asking for money and should be corrected by the judge. He made it clear that we’ve heard from many liars in recent weeks, but provided little evidence of Musk’s true legal claims.
OpenAI lawyer Sarah Eddy responded by simply arranging the mountain of evidence presented by the company in chronological order. She spent no time trying to pretend that anyone in this trial was particularly reliable. She did, however, have the sting of the day about Musk: “Even the mother of his children can’t support his story. » William Savitt, who took over for the defendant after his presentation, demonstrated the number of times Musk “couldn’t remember” a critical detail – and wondered how a sophisticated businessman couldn’t understand or read a four-page term sheet that OpenAI sent him.
I still wondered why we were all wasting our time here. So let’s talk about gossip, which is the real point of this trial. Was it good? Here are my favorite nuggets.
Even though this trial was supposed to punish Altman and arguably already has, I’d like to focus here on what I really take away: Elon Musk sucks at AI.
Look, Musk has repeatedly said that OpenAI will not succeed. He repeatedly tried to bring him down and steal from his researchers and in one case – that of Andrej Karpathy, a founding member of the team drawn to Tesla – succeeded. But how is xAI doing? Well, it’s a black hole for the money acquired by SpaceX. It’s a hemorrhage of researchers. One of its huge data centers will not be occupied by xAI – there is a deal with Anthropic instead. It could buy Cursor, aiming to match the programming-focused products offered by Anthropic and OpenAI. Professional users of xAI, whether the US government or private companies, have been incentivized to use it. To the extent that his custom CSAM machine Grok, aka MechaHitler, works, it seems to work because Musk has distilled other people’s designs.
Zilis wrote in 2018 that Brockman and Sutskever believed Musk “really didn’t do his homework.” [on] AI/AGI and this concerns them about working with him. I’m leaving this trial thinking that all these fucking liars deserve each other, but to be fair to Brockman and Sutskever, they were absolutely right on that one. The question now is whether anyone considering investing in SpaceX’s upcoming IPO has noticed or cares.
