Travis Kalanick Wants a Second Chance at Self-Driving Cars — This Time Without Uber

Money, power, and unfinished scores are shaping the AI news today. A tech founder is circling back to an old dream. A massive AI infrastructure project is raising questions about who’s actually in charge. And writers are learning that winning a legal battle doesn’t mean winning the war.

Kalanick’s New Company Is Reportedly Eyeing the Robotaxi Market

Robotaxis — fully autonomous vehicles that pick up passengers without a human driver — are no longer just a Silicon Valley fantasy. According to TechCrunch, Travis Kalanick, the co-founder who built Uber into a household name before being pushed out in 2017, may be steering his current company Atoms toward that exact business. He’s described it as completing “unfinished business” — a reference to Uber’s self-driving car program, which the company eventually sold off in 2020.

The history here matters. Uber spent billions trying to build autonomous vehicles and walked away. Kalanick is reportedly framing Atoms as the chance to finish what he started. Think of it like a chef who left a restaurant before the signature dish was ready, then opened a new kitchen to finally cook it.

No major tech publication other than TechCrunch has confirmed the story yet, so treat it as early-stage reporting. But if it’s accurate, Kalanick would be entering a field already crowded with Waymo, Tesla, and a handful of well-funded competitors. The real question for everyday people: more competition in robotaxis could eventually mean cheaper, more available rides in cities where these services operate.

Why this matters: Kalanick helped reshape how people get around once already. Whether he can do it again in autonomous vehicles — with a much smaller company — is a story worth tracking closely.

“Travis Kalanick moving Atoms into robotaxi business as ‘unfinished business.’”

A $3.2 Billion AI Data Center Has a Complicated Answer to “Who’s Responsible?”

Data centers — the massive warehouses of servers that power AI tools — are getting built at a pace most people don’t see coming. But according to Ars Technica, one $3.2 billion project reportedly involves such a layered web of corporate structures that it’s genuinely unclear who would be held accountable if something went wrong — whether that’s an environmental spill, a safety incident, or a community impact.

Imagine ordering food through an app, which contracts with a courier service, which uses independent drivers — and then trying to figure out who’s responsible when your order goes badly wrong. These data center arrangements can work similarly. Separate entities handle the land, the construction, the operations, and the financing, which can make legal and regulatory accountability genuinely murky.

For people living near these facilities, that ambiguity is not abstract. Data centers consume enormous amounts of electricity and water. Communities that host them sometimes discover, after the fact, that no single company is clearly on the hook for managing those impacts. This story is reportedly still developing, and Ars Technica notes the accountability questions are part of a broader pattern across the industry.

Why this matters: The infrastructure powering your AI tools is growing fast, and the rules governing who answers for its effects haven’t kept up.

“$3.2 billion AI data center with unclear corporate accountability structures.”

Copyright — the legal right authors have over their own published work — has become a central battleground as AI companies trained their models on books, articles, and other written material. Anthropic, the company behind the Claude AI assistant, reportedly reached a settlement with writers whose work was used in that training. Now, according to TechCrunch, the authors themselves are pushing back — because publishers and literary agents are claiming a share of that money too.

The dispute is essentially about who owns the harm. Authors wrote the books. Publishers helped distribute them. Agents brokered the deals. Each party has some contractual relationship to the work, which means each is now arguing they deserve a cut of whatever Anthropic paid out. It’s similar to a band winning a royalty dispute and then arguing with their label and manager about who keeps how much.

For working writers, this is discouraging news. They fought for compensation — in many cases waiting years — only to face a secondary fight over the payout itself. The story reflects a broader pattern: legal victories in AI copyright cases don’t automatically translate into money in authors’ pockets.

Why this matters: How this dispute settles will shape whether future AI copyright agreements actually benefit the creators at the center of them.

“Publishers and literary agents dividing settlement money from Anthropic.”

Also Happening in AI

OpenAI published a behind-the-scenes look at how its researchers are using AI agents — software that can take sequences of actions on its own — to speed up coding and experimentation work internally. Separately, a HuggingFace team demonstrated something genuinely surprising: they used reinforcement learning (a training method where a model learns by trial, error, and reward) to teach a coding model to generate watercolor-style images, showing how techniques travel unexpectedly across domains. On the research front, a new benchmark called WearableQA is testing whether AI can reason meaningfully over real health data from wearable devices and blood tests — a small but meaningful step toward AI that could assist with personal health monitoring. Meanwhile, a team built UniMate, a model that generates realistic 3D character animations from text descriptions alone, and another group created Diffusion TV, a physical installation that lets visitors manually adjust how an AI image-generation process unfolds in real time.

What to Watch

The Anthropic settlement dispute is likely a preview of fights we’ll see repeatedly as more AI copyright cases resolve. Watch for whether courts or contracts start drawing clearer lines about how creators — not just rights-holders — get compensated. On the infrastructure side, the accountability questions raised by the $3.2 billion data center story point to a regulatory gap that local governments are only beginning to notice. The next few months will show whether any jurisdiction moves to close it.