UK ❤️ Nvidia, Getty Images v. AI & more (June 9, 2025)

Hey friends,

Today’s AI headlines hit both ends of the spectrum: a legal battle that could define creative rights in the AI era, and a geopolitical sprint to build sovereign AI infrastructure. Plus, a closer look at whether AI is really thinking or just really good at guessing.

Let’s jump in 👇

⚖️ Courts & Compute Power

🖼 Getty vs Stability AI in UK Copyright Trial

Getty Images claims Stability AI used millions of copyrighted photos without consent to train Stable Diffusion. Opening arguments are underway, and the outcome could redefine copyright boundaries for generative models.

A robotic arm in a white control room transfers photo thumbnails from a digital wall into a glowing AI brain display

💻 Nvidia & UK Announce AI Superplan

Nvidia is investing over £1 billion into the UK’s AI infrastructure. Announced during London Tech Week, the collaboration includes:

A Sovereign AI Industry Forum with BT, BAE Systems, and Standard Chartered
A new Nvidia AI Center in Bristol
A regulated digital sandbox with the Financial Conduct Authority

🤖 AI Reasoning Models Hit a Wall

A new study by Apple researchers challenges the idea that reasoning models can “think.” While they outperform traditional LLMs on medium puzzles, they underperform on the hardest ones, often giving up entirely. It’s a warning sign: more compute doesn't mean more intelligence.

🕰 From Scrolls to Screens

🧬 AI Re-dates Dead Sea Scrolls

The Dead Sea Scrolls have been re-dated as researchers used an AI model named Enoch to compare handwriting patterns with radiocarbon data, revealing that some scrolls may be up to 2,300 years old. Ancient history, now with machine learning.

An AI examines ancient scroll fragments on a stone table in a white library, comparing text with data on a digital screen

🤔 OpenAI’s Chat Retention Controversy

A court order now requires OpenAI to store all user conversations, even ones you delete. The New York Times wants data preserved for copyright evidence. OpenAI’s Sam Altman is calling for “AI privilege” akin to doctor-patient confidentiality.

💬 Human-AI Bonds, Explained

OpenAI’s Head of Model Behavior Joanne Jang shared insights into why users are forming emotional connections with ChatGPT, and how design choices are shaping a new form of digital companionship.

A man walks beside a glowing holographic AI figure along a tree-lined path in Central Park, with city skyscrapers in the background

🛠️ Tools of the Day

PodGen – Turn webpages and docs into podcasts
Nanobrowser – Automate browser tasks using prompts
Flux Kontext – Edit and style images using AI
Substrata – Real-time sales coaching via behavioral signals

🎯 Prompt of the Day

Optimize your LinkedIn Profile
Prompt: “Act as a LinkedIn personal branding expert. I want to attract [insert goal]. Analyze my profile (attached), suggest headline options, rewrite the About section, and identify experience improvements. Be specific and use bold headers.”

⚡ Quick Hits

Meta in talks to invest $10B in Scale AI (more tomorrow)
Anysphere releases Cursor 1.0 with bug-fixing AI agents
Google’s AI Mode adds data visualizations for finance
Anthropic brings in national security advisor Richard Fontaine

📌 TLDR

Getty’s copyright battle with Stability AI could reshape the rules of AI training, while the UK is doubling down on national AI infrastructure with Nvidia. Meanwhile, OpenAI and Apple ask a deeper question: is AI really “thinking,” or just pattern matching?

Until next time,
David