Alex Ossola

Mandelson hasn't responded to requests for comment, but he's previously said that he had no recollection or record of receiving the funds.

Separately, the Pentagon is raising concerns to President Trump about an extended military campaign against Iran.

Current and former officials say that the warnings have largely been voiced by General Dan Kaine, the chairman of the Joint Chiefs of Staff, within the Defense Department and during meetings of the National Security Council.

But he's not the only one.

Some officials said that such discussions are always part of the contingency planning process before military operations, but that Kaine is seen as an aide who Trump trusts and that the issues will be a factor in the president's decision on whether to attack Iran and how.

A White House spokeswoman said that Kaine is, quote, a talented and highly valued member of President Trump's national security team and that the president listens to others' opinions before making a decision.

Following media reports on Kaine's warnings, Trump posted on social media saying Kaine is not against a war with Iran, but that, quote, And in Iran, many locals are bracing themselves for the chaos and violence that might unfold if the ruling regime is actually brought down.

Some who are worried about their physical safety are already making plans to leave Tehran, which they expect will bear the brunt of any airstrikes.

U.S.

artificial intelligence startup Anthropic put out a blog post today accusing three Chinese AI companies of stealing secrets from its Claude AI model.

Anthropic said that the three companies, DeepSeek, Moonshot AI, and Minimax, prompted Claude more than 16 million times to train and improve their own products.

Representatives from DeepSeek, Moonshot, and Minimax didn't respond to requests for comment.

For more, I'm joined now by Robert McMillan, who covers computer security for The Journal.

Bob, the report says that the Chinese AI companies are using a process called distillation here.

What does that entail and why are they doing it?

Anthropic said the activity by the Chinese developers raised national security concerns for the U.S.

How exactly does that work?

I'm sure copying effectively their model means that these Chinese companies could also be presenting a business risk to Anthropic.

How much of a threat is that?

And is Anthropic the only company that's dealing with this distillation issue?