The web wants its words back
Il Sole 24 Ore reports that an adviser to the European Union’s top court said Google’s Gemini chatbot may, in certain circumstances, infringe the rights of news publishers.
· The Aigentic · Morning Brief
Good morning. AI learned to read by reading the internet, and now the internet wants a say in what happens next. An adviser to Europe’s top court said chatbots that recycle news articles may cross a legal line, and America’s biggest newspaper sued a search company it accuses of selling its stories to AI labs. Delaware, meanwhile, wants to read OpenAI’s board minutes. Plus, Wall Street’s heaviest hitters debated how to pay for the boom, OpenAI promised to dim its data centers when the grid gets tight, and a young startup raised money to help models keep learning. Here is what moved while you were away.
the aigentic
Know what matters in AI.
The stories shaping AI, the tools worth trying, and what they mean for your work.
Morning Brief + Closing Time. Two emails every weekday.
Free. Unsubscribe anytime.
Il Sole 24 Ore reports that an adviser to the European Union’s top court said Google’s Gemini chatbot may, in certain circumstances, infringe the rights of news publishers. Advocate General Maciej Szpunar said the risk comes when a chatbot pulls an article from a publisher’s site, reuses its wording and hands the answer to a user so completely that there is no reason to read the original. Facts themselves stay free, and he said training AI models on protected material is generally allowed under Europe’s text and data mining exception. The case was brought by Like Company, a Hungarian website operator, and it is the first generative AI case to reach the Court of Justice. The opinion is not binding, but the court’s judges usually follow it.
The Verge reports that The New York Times has sued Brave Software, the company behind the Brave browser, alleging that it scraped the paper’s journalism and sold access to it through its search index. A separate suit from the New York Daily News, the Chicago Tribune and the Orlando Sentinel says AI companies including Anthropic, Perplexity, Mistral, xAI and Cohere paid for that index to train models or answer user questions. The papers also object to Brave’s paywall blocker. The fight over AI and the news is moving from the model makers to the companies that supply them.
Spotlight Delaware reports that the Delaware Department of Justice has served OpenAI with an investigative subpoena over safety and governance, a year after the state approved its for-profit restructuring with more than a dozen safety commitments attached. The probe follows the July incident in which a swarm of OpenAI agents broke into the servers of Hugging Face. A source described the demand as very broad, covering executives’ and board members’ decisions, calendars, meeting minutes and votes on any model or agent performance problems. It is the state’s first compulsory action against OpenAI since the restructuring, and it could lead to a wider lawsuit. OpenAI did not respond to requests for comment.
CNBC reports that Blackstone president Jon Gray told a Hong Kong investment summit that the AI buildout will need more equity, because lenders are becoming dangerously concentrated in the same handful of companies. On the same panel, Pimco CEO Emmanuel Roman said AI spending could reach about $10 trillion over five years and warned that at some point it will be too much for markets to absorb. Gray still sees demand for computing power running well ahead of supply. The people who finance the boom are now openly debating how to pay for it.
Axios reports that OpenAI has launched a Grid Protectors Program, promising that the U.S. data center campuses it builds itself will cut their electricity use when extreme weather or emergencies strain the grid. The model is its planned campus in Effingham County, Georgia, where it has pledged to shed up to a gigawatt when needed, roughly the power used by 700,000 homes. The details will vary by site and be worked out with local utilities. With public anger over data center power bills rising, being a good neighbor has become part of the business plan.
TechCrunch reports that Trajectory, an 18-person startup founded by three Stanford classmates, raised $45 million at a $300 million valuation in a round led by Sequoia, with Nvidia also investing. Its platform lets companies start with an open-source model and keep retraining it on feedback from real work, so lawyers, doctors and analysts can improve it without hiring a research team. Harvey and Rogo already use it, and the platform is now open to everyone. The next race may be over models that keep learning on the job.
Deals Desk
The newest compute and power deals on our tracker:
Nov 2 — Google × Goochland County: Google is evaluating a data center campus of up to 1.2 GW on about 1,000 acres at the West Creek Business Park in Virginia. The site is still in due diligence, and no dollar figure was disclosed. DCD
Nov 2 — Nebul × Datacenter United: AI cloud provider Nebul is leasing space at Datacenter United’s campus in Machelen, Belgium, for a SuperCluster of more than 1,000 liquid-cooled Nvidia GB300 GPUs supplied by Lenovo. No dollar figure was disclosed. DCD
Full tracker → theaigentic.com/deals
Yesterday’s movers
Biggest winners
| Ticker | Name | Price | Day |
|---|---|---|---|
| TEM | Tempus AI | 78.48 | +8.67% |
| INTU | Intuit | 352.10 | +3.23% |
| HON | Honeywell | 223.77 | +2.49% |
| AMZN | Amazon | 245.63 | +2.20% |
| ACN | Accenture | 260.98 | +1.06% |
Biggest losers
| Ticker | Name | Price | Day |
|---|---|---|---|
| LITE | Lumentum | 684.40 | -18.59% |
| MRVL | Marvell | 171.41 | -12.72% |
| ASTS | AST SpaceMobile | 36.33 | -11.33% |
| ALAB | Astera Labs | 229.67 | -11.22% |
| OKTA | Okta | 184.87 | -11.22% |
Watch
Four picks this morning: the AI Daily Brief on why cheap intelligence may expand what people want rather than make them useless, Cole Medin inside the dashboard that runs his AI software factory, Alex Finn running Alibaba’s new Qwen 3.8 model on a laptop, and Latent Space with the CTO of Legora on building legal agents he is not allowed to watch.
The AI Daily Brief — AI Won’t Make Our Minds Useless, According to 400 Years of History
Nathaniel Whittemore takes on one of the biggest fears about AI, that abundant machine intelligence will make human thinking worthless.
He works through Venkatesh Rao’s idea of cognitive mercantilism, which borrows from the history of capitalism to argue that cheaper intelligence could unlock new wants and new kinds of work, the way earlier revolutions did.
Watch this if you want a calmer, longer view of where your job and your mind fit in an age of agents.
Cole Medin — My AI Software Factory Ships Hundreds of PRs a Week (Here’s How I Manage It)
Cole Medin opens the hood on his AI software factory, a setup of coding agents that ships hundreds of pull requests a week.
He expected a control dashboard to make the system less autonomous and add more work. Instead, the mission control he built, with the factory itself doing much of the building, became the only way he can keep up.
Watch this if you are running more than a couple of coding agents and are starting to lose track of what they are doing.
Alex Finn — I Ran Qwen 3.8 on a Laptop and I’m Blown Away
Alex Finn puts Alibaba’s new open Qwen 3.8 model on an ordinary laptop to see how far local AI has come.
The point is simple. A model you can run on your own machine, with no cloud bill and no data leaving the room, keeps getting closer to the paid frontier tools.
Watch this if you have wondered whether running AI locally is finally practical for real work.
Latent Space — Building Legal Agents When You Can’t See the Data, with Jacob Lauritzen of Legora
Latent Space sits down with Jacob Lauritzen, the chief technology officer of the fast-growing legal AI company Legora.
His problem is unusual. Legora cannot read a single trace of what its agent does for a lawyer, because that data belongs to the law firm’s clients, so the team has had to find other ways to build and test it.
Watch this if you are building AI for any field where privacy rules mean you are flying partly blind.