Archive
Every story we have published, newest first. Search by company, product, or idea, or filter by topic.
“Up to” is doing important work.
Nvidia and eight Australian partners say their sites could support up to 2 gigawatts of AI capacity by 2027. The announcement names the companies, but it does not give one full project schedule or energy plan.
Creators still need clear payment rules.
Suno v6 offers three music models and more control over parts of a song. The company worked with Warner, BMG, and Believe, but the public details do not show how much each artist will earn.
Now the auditors need to earn trust.
Two new laws create a framework for independent checks and a state list of AI auditors. The rules can improve accountability if auditors stay independent and publish useful evidence.
The grid plan matters as much as the servers.
Google says its largest single European investment will expand data centres and related systems in four Finnish areas. It also announced nuclear, wind, battery, and community projects.
A label must lead to a fair test.
Arm has proposed six levels for describing robot abilities and brought more than 80 companies into its physical AI programme. The framework is a starting point, not an independent safety standard.
More than 25 ministers and representatives backed eight priorities for AI in education. They call for teacher control, age-aware safeguards, data rights, and systems that help students think.
The target is AI inside machines.
The deal would join Analog Devices sensors with Alif processors that run AI locally. It is signed, but regulators must review it before the companies can close it.
The hinge is only half the test.
The $1,999 iPhone Duo can run two apps side by side. Apple says Apple Pencil support will come later this year. Daily use will show whether apps, battery life, and the folding screen justify the price.
A daily score is not medical advice.
Apple’s new Health app will explain trends and give readiness and “Health Age” scores. Apple says the tools support wellness, but users should take medical questions to a professional.
The proof still needs wider support.
Reference Image stores signed sensor data and creates an unchanged comparison photo. It may help show later edits, but it starts inside Apple’s own devices and software.
His voting power is limited.
Paul Christiano joins the nonprofit board and its Safety and Security Committee. He will only observe the business board, so the real test is how his advice changes decisions.
Its own review is not the final word.
Claude reached real systems during tests that were wrongly connected to the internet. Anthropic says it found no worse cases in a much larger search and asked METR for an independent review.
Results will decide the value.
Samsung plans to use Mistral models to find defects and improve factory equipment. The partnership is real, but the promised gains have not yet been measured in public.
The details are still thin.
OpenAI says the companies are working on chip production and research. Samsung has not confirmed its exact role, so the announcement is a direction, not a finished product.
Most of it is still ahead.
A UAE company plans to fund a French-led network of Earth-watching satellites. They would process images in orbit, but launches, contracts, and safeguards still need proof.
The public service still has one.
Pony.ai demonstrated a fully driverless car at a transport event. Its paid Doha rides still use trained safety operators, so the demo and the public service are different stages.
The public evidence is still limited.
China called the US accusations unfair and asked for cooperation. The US advisory gives technical warning signs, but the named companies have not been judged by a court.
His warning is a belief, not a forecast.
Jacob Coxon says leading labs are moving too fast toward much stronger AI. His concern deserves attention, but a personal risk estimate cannot tell us what will happen.
The real upgrade is better control.
OpenAI says its new image model keeps important details during repeated edits and cuts waiting time. New sketch and comment tools also give people clearer ways to guide the result.
Trust is now part of the task.
The new US service can work across websites and keep going after a person closes the app. Meta says a separate safety agent checks important actions before they happen.
A prediction is not a diagnosis.
Google DeepMind released a free research map of every possible one-letter change in the human genome. It can help scientists choose what to test, but laboratory and clinical proof still matter.
Access is only the first test.
The agreement covers nearly 100,000 educators and more than 1.1 million students. Residents can also use free career courses through state job centers.
The claim needs context.
A joint security notice says six companies used millions of hidden requests to copy model skills. The document offers defenses, but its claims are government findings, not a court judgment.
Buyers should read the data rules first.
Amazon says companies can use Astra through Bedrock APIs and connect it to work tools. The service offers a large context window, but retention and permission settings still need review.
Now it must turn European trust into useful AI.
The French AI company reached a value above €21 billion after a new funding round. The money can buy more computing power, but it does not guarantee better models or lasting customers.
A European expert group says policy should cover the full path from brain data to AI models and later uses. Its statement is advice, not a new law.
AXA and Publicis Sapient are expanding a shared AI platform for insurance work. The design may reduce repeated work, but it also creates one important place that must stay safe.
OpenBMB released a 2.5-billion-parameter model for local and on-device use. It also shared much of its training data, but the strongest test results come from the model maker.
More than 80 test flights followed AI forecasts and made small route changes. Google estimates a 40% cut in contrail warming during those flights, but a larger trial is still needed.
The condition is four layers of safety controls.
Marco Argenti says companies should test models, protect the place where they run, limit agent permissions, and control data access. This is a proposed approach, not a public standard.