Actualités technologiques et économiques

China Mobile, one of the country’s state-backed telecoms giants, has tripled the capacity of its largest intelligent computing hub in northern Inner Mongolia, amid Beijing’s push to boost digital infrastructure to power the country’s artif…

The country’s tech giants are snapping up products never designed for advanced AI to satisfy soaring computing demands

Committee chair rejects criticism that letter was ‘pre-determined’ by the government Follow our Australia politics live blog for latest updates Get our new political email , free app or daily news podcast Anthony Albanese’s office signed o…

For most languages of the world, language model pre-training operates in a data-constrained regime where models must repeat their training data many times, degrading generalization. Two remedies exist: aggressive hyperparameter tuning such…

"No more hand-written code" or AI-free art for delayed Marvel/Star Wars shooter.

Ahold Delhaize kampt met aanhoudende netwerk- en distributieproblemen door een beveiligingsincident dat systemen offline heeft gedwongen, terwijl een gelijktijdige cyberaanval op koelsystemen in Franse supermarkten leidde tot massale voorr…

Meta's new Muse Wonder models rank in the top three on six out of ten benchmarks, with the biggest gains in agentic tasks. But the cost per task has nearly tripled. The article Meta's new frontier model Muse Wonder Full breaks into the AI…

“I may have information regarding this case," Claude AI told Philly police.

If you’ve ever wondered how Claude and similar AI agents can answer complex questions in seconds and where they get the information, you’re not alone. Gurmeet Saran works to find that answer every day as an engineer at Anthropic in San Fra…

Sakana AI's MASS method lets one language model improve itself by solving tasks, judging results, and refining workflows, reaching a win rate of nearly 70 percent against the base model after two rounds. But when the model misses the same…

Where Muse Spark aimed to be a lightweight model that could perform at low latency, the Wonder models are larger and said to be more capable, with higher benchmark scores and better results on complex tasks like coding.

Musk, we find your lack of security disturbing.

Russian bloggers warn Russia’s hyper-centralized digital economy is vulnerable.

Models routinely take shortcuts to pass evaluations. Who gets to decide what incidents are innocuous?

Other disciplines should take heed

“Priming” reuses pretrained Transformer weights to build models with more efficient architectures at less than 1% of the original training cost and with throughputs up to twice as high.

The prime minister tells MPs the next summit with the EU will be held on 20 November Burnham to urge MPs to embrace closer ties with EU Zia Yusuf , Reform UK’s home affairs spokesperson, is speaking now. He started his speech with an accou…

Let’s be clear: the risk to humanity is unprecedented. Without an urgent public debate and robust use of international law, we will lose control of the future In my two years as attorney general, I spent more time working on national secur…

A Silicon Valley startup that uses weather balloons and AI to predict the weather won a U.S. military contract worth almost $50 million.

Meta’s biggest model rivals the world’s best—and includes new features to keep agents from going off the rails.