Elon Musk and Sam Altman are about to have a very public, very expensive argument in a courtroom this week. The case has huge implications for OpenAI’s future, the global AI race, and whether a company that started as a non-profit can just flip a switch and become a for-profit giant.
Musk, who was an early co-founder and donor, claims he was tricked into funding OpenAI under false pretenses. He’s asking for $134 billion in damages, wants Altman and president Greg Brockman removed, and demands the company go back to being a non-profit. That’s a tall order, especially with OpenAI’s IPO looming. The court could decide not just the company’s structure, but who gets to run it.
This isn’t just a grudge match between two billionaires. The outcome could reshape how AI companies are allowed to operate, especially if the court rules that non-profit origins don’t give you a free pass to become a for-profit juggernaut. If Musk wins, it might set a precedent that makes other AI startups think twice before pivoting from charity to capitalism.
Meanwhile, the broader AI industry is facing a different kind of crisis: nobody knows how to actually make money from this stuff. Will Douglas Heaven at MIT Technology Review put it perfectly by referencing that old South Park episode where gnomes steal underpants as a business plan. Phase 1: Build the tech. Phase 2: ??? Phase 3: Profit. We’re stuck at Phase 2.
Companies have spent billions on models, infrastructure, and talent. They’ve promised the world that AI will transform everything. But the path from hype to profit is still a giant question mark. Subscription models, enterprise licensing, API fees—none of it has proven sustainable at scale. It’s a problem that’s been around for a while, but it’s getting more urgent as investors start asking hard questions.
And then there’s the dark side of the coin: weaponized deepfakes. Eileen Guo’s piece on this is sobering. Cheap, accessible models are now generating deepfakes that look startlingly real. We’re talking sexually explicit images, political propaganda, and content designed to incite violence. Women and marginalized groups are taking the brunt of it. Experts are alarmed, and they should be. We’ve been warned about this for years, but now it’s actually happening.
The trust erosion is real. People are already questioning whether any video or audio clip can be trusted. That’s a dangerous place to be, and it’s not going to get better as the tech improves.
A few other things caught my eye today. OpenAI ended its exclusive partnership with Microsoft, which means they can now sell to Amazon and other rivals. That’s a big shift, but it also signals that OpenAI is scrambling for revenue ahead of its IPO. They’re missing key growth targets, according to the WSJ. Not a great look.
Google signed a classified AI deal with the Pentagon, allowing AI use for “any lawful government purpose.” Over 600 Google workers had called for a block on the deal. That tension between corporate ambitions and employee ethics isn’t going away.
The full newsletter also mentions that AI firms are training military versions of their models on classified data. That’s a whole other can of worms.
This is a lot to digest in one day. Musk vs. Altman, AI’s profit problem, weaponized deepfakes, and the military-industrial complex getting cozy with AI. It’s a reminder that the technology is moving faster than our ability to regulate it, monetize it, or even understand its consequences.
Comments (0)
Login Log in to comment.
Be the first to comment!