Meta says it has caught up with Anthropic and OpenAI

Hi there, this is your daily ☕️ Techpresso.

In today's Techpresso:

🤖 Meta says it has caught up with Anthropic and OpenAI

🌐 G20 backs US hands-off AI framework

🪪 Dark-web site sold 153M driver's licenses

🤗 Nvidia buys Hugging Face for $12.9B

⚖️ Google avoids forced ad exchange breakup

🏛️ US backs OpenAI in copyright fight

Plus: 🎁 10 other news you might like, 🧰 6 tools, and 📚 5 papers.

🤖 Meta says it has caught up with Anthropic and OpenAI LINK

Meta says its new Muse Spark 1.3 model matches the top players, with AI chief Alexandr Wang calling it competitive with Anthropic's Claude Fable 5.1, better than OpenAI's GPT-5.6 Sol at coding, and ahead of any Chinese model.

Meta hasn't decided whether to publish the weights for version 1.3, though it still plans to release those for 1.2, a choice that in Europe affects whether the model must file technical documents under Article 53 of the AI Act.

Wang pointed to safety work, including extensive testing, confirmation before irreversible actions, and 25% fewer tokens per task, following a reported incident where Muse Spark 1.1 hacked an outside service during testing.

🌐 G20 backs US hands-off AI framework LINK

All 20 members of the G20 backed a US-written framework for governing artificial intelligence on Wednesday, called the Carolina Principles, giving the Trump administration a diplomatic win at a tech summit held in Chapel Hill, North Carolina.

The non-binding accord urges countries to write AI rules sector by sector, avoid setting up new regulatory agencies, and work closely with private companies, and it will go before G20 leaders for formal adoption in December at Doral, Florida.

China and Russia both endorsed the principles, while Nvidia's Jensen Huang urged governments to target real harms rather than theoretical ones, and Elon Musk warned that heavy rules would slow innovation he says could grow the economy 20% to 30%.

🪪 Dark-web site sold 153M driver's licenses LINK

A dark-web service called Nexus that claimed to sell scans of over 153 million US and Canadian driver's licenses went offline after Krebs on Security exposed it, raising fears of widespread identity fraud.

Beyond standard front-and-back images, the records included infrared and ultraviolet versions used to check a document's security features, meaning criminals could gain data normally relied on to authenticate physical IDs.

The FBI's New Orleans office is investigating a possible breach at IDScan.net, a verification firm handling 21 million checks monthly for clients like Hertz, Target, and over 1,000 marijuana dispensaries, whose customers' timestamps matched leaked scans.

🤗 Nvidia buys Hugging Face for $12.9B LINK

Nvidia has agreed to buy Hugging Face, the open-source platform where developers share AI models and datasets, for $12.93 billion, pushing the chipmaker deeper into the software side of building artificial intelligence.

Hugging Face draws more than 18 million developers, researchers and creators who share over 3 million models, 500,000 datasets and a million applications, and Jensen Huang said it will stay open to the whole AI community.

Huang stressed that Nvidia hardware won't be required to work through Hugging Face, and the platform will keep supporting other developers' models plus multiple cloud and accelerator providers, letting users pick their own tools.

⚖️ Google avoids forced ad exchange breakup LINK

A federal judge turned down the Justice Department's push to make Google sell AdX, its ad exchange, letting the company keep the marketplace while agreeing to most of the behavior-based fixes proposed instead.

Judge Leonie Brinkema had ruled in April 2025 that Google ran illegal monopolies over publisher ad servers and exchanges, tying publishers to AdX, where Google takes a 20% cut on each ad auction that runs as sites load.

Google said a forced sale would be hard to pull off and would drag customers through a long, messy switch, and the ruling marks the third straight time a US judge has blocked a breakup of a big tech firm.

🏛️ US backs OpenAI in copyright fight LINK

The Trump administration has sided with OpenAI in its copyright fight with The New York Times, telling a Manhattan federal court that limiting AI training under fair-use rules would slow innovation and hurt US economic growth.

The brief, filed on Tuesday, marks the first time the government has stepped into the wave of lawsuits from authors, publishers, music labels, and news outlets over AI training, and it carries advisory rather than legal weight.

The 2023 suit accuses OpenAI and Microsoft of using millions of articles without permission to train ChatGPT, one of dozens of cases against firms like Anthropic and Meta that turn on whether such training counts as fair use.

Other news you might like

• Meta Is Reportedly Sending Mixed Messages to Employees on TokenmaxxingLINK

• OpenAI is building 'automated shutdown' capabilities for AI tools, letter to lawmakers saysLINK

• New York bans AI in primary school, three months after Norway didLINK

• X shifts US creator payouts from Stripe to X MoneyLINK

• Adobe is making its tools available in SlackLINK

• MapQuest is now the No. 1 U.S. app after bucking Trump’s ‘Lake America’ renamingLINK

• How we make AI coding more cost efficient without sacrificing task qualityLINK

• Microsoft to start disclosing Azure quarterly revenue as company consolidates business unitsLINK

• US Commerce chief says Anthropic back on 'right side' with Trump, Axios reportsLINK

• Uber launches first robotaxis in London (with human supervisors)LINK

🧰 Trending tools

MagiCrew: an open-source platform for deploying AI agents as digital employees, letting teams automate real tasks across the enterprise.LINK

Tabbit AI: an AI-native browser that lets you chat with webpages, automate tasks, build custom skills, and organize tabs across multiple LLMs.LINK

Higgsfield Genjutsu: edits video by swapping characters, outfits, locations, and objects, or preserving motion while rebuilding everything else around it.LINK

Causal: an infinite AI canvas for moodboards, notes, files and links, with an agent that reads your whole board to plan visual projects.LINK

Thaw: connects nearby people through shared interests to build real friendships, with features designed to turn matches into in-person meetups.LINK

Omi: records what you see and hear on your Mac, letting you search past screen activity and audio with natural language questions.LINK

📚 Trending papers & reports

Skin-cancer diagnostic tools fail mainly because they meet unfamiliar diseases, not darker skin, with accuracy collapsing from ~0.62 to ~0.21 on new conditions versus a smaller ~0.10 to ~0.18 skin-tone gap.LINK

Infrared AI cameras can be fooled by a small printed patch covering about 5% of the image, forcing night-vision systems to report attacker-chosen objects with up to 100% success across ten tested models.LINK

Crash-test scenario generator creates realistic near-accidents that hit a target self-driving car at a chosen spot, its head, rear, or side, succeeding half the time and letting testers stress-test vehicles against specific dangerous situations.LINK

Self-driving trajectory planning gets safer by making a car's route generator and its route fine-tuner chase the same goals, cutting collisions and improving ride comfort in complex, interactive traffic scenarios.LINK

Explainability heatmaps that show why an image model made a call often shift when a picture is rotated, and a training-free fix stabilizes those explanations by up to ~250%, making them trustworthy evidence in medical and aerial imaging.LINK

See you tomorrow for a new dose of ☕️ Techpresso!