🤝 OpenAI ends Microsoft legal peril over its $50B Amazon deal ↗
OpenAI and Microsoft reworked their agreement again, and the main practical point is that OpenAI’s huge Amazon cloud deal no longer looks like a legal tripwire for Microsoft.
The peculiar part is that some are framing this as OpenAI “winning” against Microsoft. Both sides, though, seem to leave with something in hand - OpenAI gets more flexibility, while Microsoft gets less exposure and sharper clarity. Corporate origami, basically.
⚖️ The Elon Musk vs. OpenAI trial starts today ↗
Elon Musk’s court fight with OpenAI moved into trial mode, with jury selection beginning and Sam Altman reportedly present in the courtroom.
The case centres on Musk’s claim that OpenAI drifted from its founding mission toward a more profit-driven structure. It is part legal drama, part founder-family breakup, part Silicon Valley thunderstorm in a glass jar.
🧾 South Africa withdraws AI policy due to fake AI-generated sources ↗
South Africa pulled its draft national AI policy after fake AI-generated references were found in the document. That is almost too on-the-nose, but here we are.
The policy had proposed major AI governance bodies and incentives, but the citation problem badly undercut its credibility. Officials called it a serious error and promised accountability, which is fair - an AI policy getting tripped up by AI slop is a very modern pothole.
🧠 DeepMind’s David Silver just raised $1.1B to build an AI that learns without human data ↗
David Silver, known for major DeepMind breakthroughs, raised a hefty $1.1 billion for a new AI effort focused on systems that learn without relying on human data.
That is a big philosophical swing. Most frontier AI still feeds heavily on human-made text, images, code and behaviour. This pitch is closer to “let the machine figure it out itself” - elegant, slightly terrifying, and very expensive.
🛰️ Meta inks deal for solar power at night, beamed from space ↗
Meta signed an agreement with Overview Energy that could involve satellites beaming infrared light down to solar farms powering data centres at night.
Yes, space power for AI data centres. It sounds like a sci-fi prop someone left in a boardroom, but the logic is straightforward: AI needs mountains of electricity, and Big Tech is now hunting for power in less familiar places.
📱 Investors back Skye’s AI home screen app for iPhone ahead of launch ↗
Skye, an AI home screen app for iPhone from Signull Labs, picked up investor backing before launch.
The idea fits the current “AI as interface” rush - not just another chatbot box, but something closer to a front door for your phone. Potentially practical, potentially another layer of app-soup. Still, investors clearly think the home screen is up for grabs.
FAQ
What were the biggest stories in this AI news roundup?
The main AI news themes were legal pressure, cloud deals, governance missteps, energy demand and fresh startup funding. OpenAI appeared in two major legal and business stories, while South Africa’s withdrawn AI policy showed how quickly credibility can be damaged by poor sourcing. The roundup also covered AI infrastructure, space-based energy ideas and phone interfaces.
Why does OpenAI’s Amazon cloud deal matter for Microsoft?
OpenAI’s large Amazon cloud deal matters because its agreement with Microsoft had created potential legal uncertainty. The article says OpenAI and Microsoft reworked their arrangement so the Amazon deal no longer appears to be a legal tripwire for Microsoft. OpenAI gains more flexibility, while Microsoft reduces exposure and gets clearer boundaries.
What is Elon Musk’s trial against OpenAI about?
Elon Musk’s trial against OpenAI centres on his claim that the company moved away from its original founding mission and became more profit-driven. According to the article, jury selection began and Sam Altman was reportedly present. The case is framed as both a legal fight and a broader dispute over OpenAI’s direction.
Why did South Africa withdraw its AI policy?
South Africa withdrew its draft AI policy after fake AI-generated references were found in the document. That damaged the credibility of a policy meant to govern AI responsibly. The article notes that the draft included proposed governance bodies and incentives, but the citation problem became a serious error that officials said would require accountability.
What does it mean for AI to learn without human data?
An AI system that learns without human data would rely less on human-created text, images, code or behaviour. The article describes David Silver’s new effort as focused on systems that figure things out more independently. Typically, this points toward approaches where models learn through environments, feedback or self-directed exploration rather than mainly copying human examples.
Why are Big Tech companies looking at unusual energy sources for AI?
AI systems and data centres require large amounts of electricity, so major tech companies are exploring less familiar energy options. The article mentions Meta’s agreement involving possible satellites that beam infrared light to solar farms at night. The logic is practical: as AI demand grows, companies need more reliable ways to power infrastructure.