Musk vs. Altman heads to trial, Meta bets on space solar power, and the Met police AI surveillance story

Two billionaires. One courtroom. Years of public mudslinging - and it finally counts for something.

Today the Musk vs. Altman lawsuit formally goes to trial in California, which is either the most dramatic legal event in AI history or an expensive grudge match, depending on who you ask. Either way, it sets the backdrop for everything else in today's newsletter - a reminder that AI development isn't just a technical story. It's also a very, very human one.

Let's get into it.

Which controversial data analytics company built the AI tool the Metropolitan Police used to surveil its own officers?

The answer is hiding near the bottom of today's newsletter... keep scrolling. 👇

The lawsuit everyone in AI has been watching for months is finally here. Elon Musk filed suit against OpenAI and Sam Altman, alleging that Altman committed fraud by steering the company away from its original nonprofit mission and toward a for-profit structure. OpenAI's response? That Musk is "motivated by jealousy."

Musk co-founded OpenAI in 2015 as a nonprofit with the mission of building safe AI for humanity's benefit. His core argument is that converting to a capped-profit model betrayed that founding agreement and the people who donated to it. OpenAI, for its part, argues Musk knew what direction the company needed to go and simply didn't like being sidelined from it.

Whatever the outcome, the trial will put on the record some genuinely interesting questions about what AI companies owe their original investors and missions - questions the whole industry is watching closely. You can follow the legal battles tag for ongoing coverage.

Read the full story →

This is not a headline from 2045. Meta has signed its first commercial contract with startup Overview Energy for space-based solar power - electricity generated by solar panels in orbit and transmitted wirelessly to Earth. The deal is small by Meta's standards, but it's described as a significant first step toward making space solar a commercial reality.

The backstory here matters. AI data centers are consuming electricity at a pace that's straining grids around the world - we covered how UK government departments can't even agree on their own AI energy forecasts. Space-based solar is appealing because it can generate power continuously - no clouds, no night-time drop, no seasonal variation.

The energy infrastructure challenge isn't going away anytime soon, and if Meta's bet pays off even partially, it changes the math on where you can build next-generation AI compute. Worth watching even if commercial-scale deployment is still years away.

Read the full story →

Here's a story that sits right at the intersection of AI surveillance and accountability. London's Metropolitan Police deployed a Palantir-built AI tool over the course of a single week to surveil its own staff, combing through internal data to flag rule-breaking. The result: investigations into hundreds of officers, covering everything from working-from-home violations all the way up to suspected corruption.

The scale is striking. What would have taken a traditional internal affairs team months of work was accomplished in a week. Palantir is no stranger to controversy - its tools are widely used in government and law enforcement, and critics have long raised concerns about algorithmic bias and overreach when AI is used for surveillance.

But the Met's defense is straightforward: it used data the force already held about its own employees. Whether that's reassuring or alarming probably depends on your priors about AI in government. Either way, expect this to become a template - or a cautionary tale - for other forces watching closely.

Read the full story →

On the research front, Meta Reality Labs released Sapiens2, a new foundation model family built specifically for human-centric computer vision. The model handles pose estimation, body segmentation, surface normals, pointmap generation, and albedo estimation - all from a single backbone architecture. In plain English: it can reconstruct detailed 3D geometry of human bodies from standard images, with state-of-the-art results across the board.

The unified backbone approach is the interesting technical detail here. Previous models typically specialized in one of these tasks. Sapiens2 handles all five without sacrificing accuracy on any of them - which is the kind of efficiency gain that matters enormously when you're deploying at scale in AR/VR environments or building avatars.

Applications range from gaming and virtual production to fitness tracking and physical rehabilitation. If you're building anything that involves understanding human bodies in 3D space, this is worth a proper look.

Read the full story →

The first edition of the World AI Film Festival (WAIFF) ran in Cannes this week, showcasing a very strange collection of AI-generated cinema - including, per The Guardian's description, visions of men with fish scales erupting from their necks and seaweed from their mouths. Not your typical Palme d'Or contenders.

The establishment Cannes festival bans AI-generated work from competing for the Palme d'Or. But WAIFF is attracting real investment and genuine attention as an alternative. This is the pattern we've seen repeatedly in AI creativity: when legacy institutions close doors, new institutions form around the technology.

Whether AI-generated film becomes a genuine art form or a novelty niche is still an open question - but the fact that it now has its own festival in Cannes is a signal worth noting. The creative industries are figuring out their relationship with AI in real time, and WAIFF is one data point in that process.

Speaking of building creative projects quickly - if you're prototyping an idea and need a site up fast, 60sec.site is an AI website builder that does exactly what the name suggests. Worth a look.

Read the full story →

The job market story has a new chapter. As AI eliminates the bottom rungs of the corporate ladder - the entry-level roles that used to be how young people built skills and experience - a growing number of Gen Z workers are responding by skipping the ladder entirely and starting their own businesses.

The Guardian profiles Ashley Terrell, who graduated with a marketing degree in 2024, applied for jobs for months, couldn't break through, and ultimately started her own company. She's not alone. The pattern is documented across a range of industries where AI is absorbing the work that used to define entry-level roles.

There's something both inspiring and troubling about this. Entrepreneurship as a survival strategy is different from entrepreneurship as a choice. The future of work is playing out not in policy papers but in the actual decisions young people are making about how to earn a living.

Read the full story →

The answer is Palantir! The controversial data analytics and defense-tech company - founded by Peter Thiel and others - built the AI tool the Metropolitan Police used to surveil its own officers. Palantir has long been a lightning rod for debates about AI surveillance, and this week added a genuinely novel chapter: a police force using AI to police itself.

The Musk vs. Altman trial is finally happening. Do you think Musk has a legitimate legal case, or is this more about ego and competitive positioning? Hit reply and tell me what you think - I read every response and I'm genuinely curious where readers land on this one.

For more daily AI coverage, visit Daily Inference - we cover the stories that actually matter in AI, every day.

That's all for today - see you tomorrow with more. Stay curious out there.