World Labs releases Marble: detailed, high resolution and «infinite» 3D worlds

Dr. Fei-Fei Li, considered «the godmother of AI,» has long argued that «spatial intelligence» is necessary for training robots how to interact with the real world — and her lab has now revealed their new world builder model to the public.

Continue reading “World Labs releases Marble: detailed, high resolution and «infinite» 3D worlds”

Sam Altman on teen use: «Some of our principles are in conflict»

OpenAI will start automatic age checks on its users, and direct teens to clean, "age-appropriate" version.
Happy and clean ChatGPT is coming for teens, and it will call the cops if you cross the line. (Picture: generated).
Trying to balance freedom with safety, OpenAI is going all in on an age-appropriate version of ChatGPT.

Teen use of chatbots and their potential harm is rapidly becoming a hot-button political issue, complete with a Congressional hearing and an FTC probe.

OpenAI is therefore reiterating their new policies on teen use and parental controls, and says they will be rolling out automatic age verification for under-18 users that should default to the teen version when in doubt.

Continue reading “Sam Altman on teen use: «Some of our principles are in conflict»”

OpenAI announces GPT-5 Codex

GPT-5 Codex is slightly better than vanilla GPT-5 in benchmarks.
OpenAI is especially proud of the code review function in the new Codex. (Picture: OpenAI)
Savvy users have been using GPT-5-high with the Codex CLI (Command Line Interface) on their terminals for weeks, and consensus seems to be that it competes well with Claude.

Now, OpenAI is launching a custom, optimized version of GPT-5 for the Codex coding agent that they say is faster, more reliable and more steerable than before.

Continue reading “OpenAI announces GPT-5 Codex”

People are mostly using ChatGPT for personal stuff, OpenAI study finds

Most ChatGPT use is personal, and personal guidance is the top category.
People aren’t using ChatGPT for PhD-level stuff, OpenAI discovers. (Picture: Adobe)
Surprisingly few users are using the ChatGPT app for professional work, finds a new study by OpenAI.

The highest use category for ChatGPT is «non-work-related» messages, according to the study of anonymized ChatGPT usage, having grown from 53% in June 2024 to 73% of all questions posed as of July 2025.

Continue reading “People are mostly using ChatGPT for personal stuff, OpenAI study finds”

Owner of Billboard, Rolling Stone, sues Google over AI Overviews

To stop AI Overviews, you also need to stop appearing in Google's search results, and there is no way of opting out, Penske says.
Speaking for the entire media industry, Penske says AI Overviews are creating havoc on their business model. (Picture: screenshot)
Penske Media Corporation claims Google is siphoning off traffic to their websites and stealing their content with the overviews feature.

This is the first major publisher suing Google for the feature, as research from Pew shows that less than one percent of users click on from links that have an AI Overview on the results.

20% drop in traffic
PMC, which is also the parent company of Variety and The Hollywood Reporter, has seen traffic drop by some 20% and affiliate revenue decline 30% since Google’s overviews started on their stories.

Continue reading “Owner of Billboard, Rolling Stone, sues Google over AI Overviews”

Huge AI investments in Britain? Nvidia, OpenAI chiefs to travel with Trump

Could it be a new Stargate node in the UK? Billions of AI dollars are reported to be incoming.
Jensen Huang and Sam Altman are on the brink of some serious UK investments. (Picture: Adobe)
There is a UK/US summit coming up next Tuesday, where Donald Trump is expected to get a royal, red carpet welcome — but take a look at the guest list, and some real investments might be behind it, CNBC reports

Traveling with Trump is both OpenAI’s Sam Altman and Nvidia chief Jensen Huang, and speculation is rife that they are carrying some serious investment news.

«Sizable deal»
According to CNBC, both OpenAI and Nvidia are in discussions about a «major investment,» and a «sizable deal» for data center development in the UK that «could be worth billions of dollars.»

Continue reading “Huge AI investments in Britain? Nvidia, OpenAI chiefs to travel with Trump”

Friday roundup: OpenAI deals with Microsoft, makes a movie, and Albania gets an AI-generated minister

The first feature length movie made almost entirely by AI is set to debut at next year's Cannes Festival.
Made with «OpenAI resources,» this movie is built from animated uploaded drawings and prompts. (Picture: Screenshot, Critterz)
Microsoft agrees with OpenAI to keep talking
Microsoft is in a complex business relationship with OpenAI, where the early investor gets access to the latest AI tech and OpenAI gets access to computing power. They have just reached a “non-binding memorandum of understanding (MOU) for the next phase of our partnership.” This could allow OpenAI to go for-profit, under the control of a non-profit entity said to retain an ownership stake of more than $100 billion. Many takes on this today, but OpenAI has been moving away from Microsoft for funding, operations and cloud computing lately. The final deal will likely include some kind of a new investment in the now $500 billion company, and may unlock further market opportunities for OpenAI.
More at: OpenAI and Microsoft’s joint statement, x.com announcement, Reuters, Axios.

OpenAI goes to the movies
A new animated a-list movie, «Critterz» is under development using «OpenAI’s resources.» It should be ready for the Cannes Film Festival, meaning production time will be drastically sped up to only nine months. The script is written by part of the team from «Paddington in Peru», and it is spearheaded by Chad Nelson, who is a creative specialist at OpenAI. The technique looks to be to feed drawings to a large language model and have it animate them. The movie therefore streamlines animation, but wont skimp on voice actors, Gizmodo writes.
More at: The Wall Street Journal, Gizmodo and Engadget.

Read on for more news!

Continue reading “Friday roundup: OpenAI deals with Microsoft, makes a movie, and Albania gets an AI-generated minister”

Report: OpenAI signed $300 billion cloud computing deal with Oracle

OpenAI is revealed as one of the largest Oracle customers with outstanding orders of some $300 billion.
The 4.5 GW data center run by Oracle in Abilene, Texas is at the heart of the deal. (Picture: OpenAI)
In what could be one of the largest data center deals ever, OpenAI is identified by The Wall Street Journal as a significant contributor to Oracle’s recent jump in new orders.

The cloud computing giant announced on Tuesday a jump of some 359% in committed deals for $455 billion in outstanding orders — surprising the market and sending its stock up 35.95% on Wednesday, CNBC reports.

Continue reading “Report: OpenAI signed $300 billion cloud computing deal with Oracle”

Anthropic’s Claude can now create advanced files for you

Claude's files support Excel, PowerPoint, Word and PDF formats.
Claude can make sophisticated files out of any data you give it. (Picture: Anthropic)
Supporting Excel, PowerPoint and Word formats as well as PDFs, you can now get detailed financial reports and presentations from the chatbot — downloadable as professional files.

If you supply it with the data, like sales reports or statistics, Claude can now use its own sandboxed computer and coding access to provide you with well presented, clean data in file form.

Continue reading “Anthropic’s Claude can now create advanced files for you”

Veo 3 gets vertical video support, 1080p and a price cut in the Gemini API

Finally tiktok an reels ready, veo 3 has opened up a market for virality
Veo 3 is getting some massive API updates today. (Picutre: Screenshot).
In a big day for video generation at Google, the Veo 3 generator finally gets ready for Tiktok and Reels — while also hitting «general availability» in the API, according to a new blog post.

Those are the exact words that Google uses for Gemini 2.5 Flash to describe that you get as much as you can use, so they might be hinting here that there are no usage limits on the API access.

Previously, Gemini Pro users would only get three generations per day, and Ultra would get five. But if you pay as you go in the API, you might get as much as you can chew.

Continue reading “Veo 3 gets vertical video support, 1080p and a price cut in the Gemini API”

Finally: Google reveals how many queries you get on Gemini

For the first time, Google is sharing its real usage limits on their website.
Gemini is stingy with its free users, now we know exactly by how much. (Picture: Google)
Google has decided to come clean on the usage limits of its Gemini plans — and it’s a mixed bag for the free tier.

Most people have already concluded that Gemini only gives you a handful of queries per day for free, but, until now, Google wouldn’t say precisely how many.

Now it’s official, and on Google’s website.

Continue reading “Finally: Google reveals how many queries you get on Gemini”

OpenAI paper: Why hallucinations happen, and what can be done to fix them

Still confounding the AI industry is what to do with confidently wrong answers, aka hallucinations.
If we reward no answers instead of wild guesses, the industry might move away from hallucinations, OpenAI argues in a new paper. (Picture: generated).

Confidently wrong answers by large language models have been plaguing both users and labs since AIs inception, but a new study from OpenAI seeks to find a solution.

LLMs are trained on finding the next word in huge datasets, they say, focusing solely on finding the correct word in a sequence rather than looking for accuracy.

Continue reading “OpenAI paper: Why hallucinations happen, and what can be done to fix them”

Anthropic’s copyright settlement to cost $1.5 billion or more

Anthropic will pay $3,000 per book for an estimated 500,000 books, and more if further claims surface.
The Anthropic settlement is predicted to push other AI labs into negotiations over similar claims. (Picture: Adobe)
The landmark court settlement will be the largest copyright payout in history, but Anthropic avoids admitting guilt.

The epic class action lawsuit concerned a library of 7 million pirated books used in training, and had Anthropic looking at $150,000 in penalties per instance of copyright theft, but it was settled last week without disclosing terms.

Continue reading “Anthropic’s copyright settlement to cost $1.5 billion or more”

Friday roundup: Unis hiring AI officers, OpenAI on jobs and Nano Banana

Broadcom touts a $10 billion order from a mystery client, believed to be OpenAI.
Not much is known about the custom chips Broadcom will make for OpenAI, scheduled for next year. (Picture: Adobe)

OpenAI will make custom chips with Broadcom
With Nvidia lurking in the background, more companies are working on their custom AI chips. Now OpenAI has entered the fray, said to produce their own chips with Broadcom next year. It will be for internal use, and won’t be released broadly. They have a long history with this, having first entered talks with TSMC last year. Broadcom said on its earnings call this Thursday that it had secured a $10B order for AI chips without naming from whom, and now the Financial Times is reporting that it is, indeed, OpenAI, who has no comment on this.
More at: Financial Times (Paywalled) and Reuters.

Amazon lens lets you shop for anything you can see
The latest feature in the Amazon Shopping app on iOS lets you simply point your camera on anything you like, and shop for the same or similar items in real-time. It partners with Amazon’s AI shopping assistant, Rufus, to also answer questions about the products in the shop. It should «roll out to more customers in the coming weeks,» meaning there’s likely an Android version in the works.
More at: Amazon’s product page, and The Verge.

Read on for more News!

Continue reading “Friday roundup: Unis hiring AI officers, OpenAI on jobs and Nano Banana”

Report: Apple planning an AI search engine for Siri as early as March 2026

The long promised Siri LLM upgrade is slated for March, 2026.
Siri will take on Perplexity and OpenAI with its coming revamp. (Picture: Apple)
World Knowledge Answers, as it is known internally, will be a massive upgrade for their voice search assistant Siri, according to Bloomberg.

Rumor is that Apple will use an underlying, custom model from Google for the brains of the assistant, and Apple has been looking at it for quite a while.

Continue reading “Report: Apple planning an AI search engine for Siri as early as March 2026”