🕵️ China may have accessed Mythos ↗
The White House reportedly feared that a China-linked group had gained access to Anthropic’s powerful Mythos system. Those suspicions apparently helped prompt sweeping restrictions on the company’s latest models.
It remains a fairly enormous if. Washington has not confirmed the report, while Anthropic says China was never raised during its talks with officials and disputes claims that the models were dangerously jailbroken.
🇪🇺 EU Commission examines the consequences of Anthropic’s model shutdown ↗
The European Commission is assessing how the forced withdrawal of Anthropic’s advanced models affects European users. Brussels also warned that emergency security measures should not discriminate against trusted international partners.
The cutoff has sharpened Europe’s argument for technological sovereignty. Depending on American AI now looks less like convenient outsourcing and more like renting the foundations of your house... awkward.
🇨🇦 Mark Carney warns against dependence on a handful of US AI providers ↗
Canadian Prime Minister Mark Carney said the Anthropic restrictions demonstrated the danger of relying on a small number of American model providers. His message was blunt - countries need alternatives, even when nobody involved has technically done anything wrong.
Carney said governments should build and diversify rather than simply accept a sudden loss of access. AI sovereignty is quickly becoming trade policy, national security and industrial strategy stuffed into the same slightly wonky suitcase.
📈 AI companies are racing towards the public markets ↗
OpenAI and Anthropic’s reported moves towards public listings are creating expectations of a busy period for AI IPOs. SpaceX’s flotation is being treated as an early stress test for how much money investors will pour into capital-hungry frontier technology companies.
The larger shift lies in where public-market enthusiasm is moving. Consumer internet giants once dominated the story; now AI labs, chipmakers and deep-tech businesses are taking the oxygen - and, frankly, quite a lot of everyone else’s capital.
🍎 Apple reportedly plans broader third-party chatbot support for Siri ↗
Apple is reportedly developing a way for Siri users to switch between different chatbot providers, potentially including ChatGPT, Gemini and Claude. The current developer software offers more limited switching, but the menu could expand.
That would make Siri less of a single assistant and more of an AI reception desk. Apple is also reportedly preparing a customisable Camera app and another watch face, though the chatbot choice is easily the spicy bit.
🤖 Boston Dynamics’ Atlas is learning to work in a factory ↗
Boston Dynamics demonstrated Atlas autonomously sorting automotive components at a Hyundai factory. The humanoid learns movements through demonstrations, machine learning and large-scale simulation rather than relying entirely on manually written instructions.
A newer Atlas is taller, stronger and set to begin training at Hyundai’s Georgia facility. The humanoid future is arriving, surprisingly... although reliability, affordability and mass deployment remain stubbornly physical problems, not software sliders.
FAQ
Why did the White House reportedly restrict Anthropic’s latest models?
Officials reportedly feared that a China-linked group may have gained access to Anthropic’s powerful Mythos system. The report remains unconfirmed, however, and Anthropic says China was never mentioned during its discussions with officials. The company also disputes claims that its models were dangerously jailbroken, so the precise basis for the restrictions remains contested.
How does the Anthropic model shutdown affect European users?
The European Commission is examining the practical consequences of losing access to Anthropic’s advanced models. European organisations may need to switch providers, revise workflows or reconsider systems built around those models. The shutdown has also strengthened the case for more independent European AI infrastructure, rather than continued heavy reliance on American suppliers.
What does the Anthropic model shutdown mean for AI sovereignty?
The shutdown shows how quickly access to advanced AI can change in response to national security decisions. Governments increasingly regard AI sovereignty as part of trade policy, industrial strategy and national security. A common response is to support domestic models, diversify suppliers and avoid building critical services around a single foreign provider.
Why is Canada warning about dependence on US AI companies?
Canadian Prime Minister Mark Carney argued that relying on a small group of American AI providers creates strategic risk. Even when customers have done nothing wrong, government restrictions can abruptly remove access to important technology. Canada’s proposed approach is to develop alternatives and diversify providers, rather than accept that critical AI infrastructure could disappear without warning.
Why are AI companies moving towards public listings?
Public listings could give AI companies access to the vast amounts of capital required for models, chips, data centres and research. Reported IPO plans from major AI labs are also fuelling expectations of a broader wave of deep-tech flotations. Investors may increasingly view AI developers, chipmakers and frontier technology companies as the next major public-market growth category.
How could third-party chatbot support change Siri?
Broader chatbot support could allow Siri users to choose between providers such as ChatGPT, Gemini and Claude. Rather than operating as a single, fixed assistant, Siri could become an interface that routes requests to different AI systems. This may give users greater flexibility while making Apple responsible for privacy, compatibility and the overall experience across multiple providers.