A simmering feud between the AI company Anthropic and the Trump administration boiled over this week in an exchange of online posts by Anthropic co-founder Jack Clark and White House “AI czar” David Sacks.

In a speech at The Curve conference in Berkley, California and subsequent post on Substack, Clark pushed back against those downplaying the potential risks posed by AI, warning “we are dealing with is a real and mysterious creature, not a simple and predictable machine.”

That prompted a post by Sacks on X accusing Anthropic of “running a sophisticated regulatory capture strategy,” based on fear and blaming it for the state-level AI regulatory “frenzy” he claimed is “damaging the startup ecosystem.”

Clark joined OpenAI in 2016 soon after it was founded. He left in 2021 to co-found Anthropic along with six other former OpenAI employees including Anthropic CEO Dario Amodei, in part over their concerns that OpenAI had drifted from its original commitment to AI safety.

Sacks is former PayPal executive turned venture capitalist, whose firm was an early investor in Uber, SpaceX, Palantir and Airbnb among other tech companies, before being named as a White House advisor.

“This technology really is more akin to something grown than something made,” Clark said in his remarks. “We are growing extremely powerful systems that we do not fully understand… And the bigger and more complicated you make these systems, the more they seem to display awareness that they are things.”

“Right now,” he added, “I feel that our best shot at getting this right is to go and tell far more people beyond these venues what we’re worried about. And then ask them how they feel, listen, and compose some policy solution out of it.”

Read more: California’s SB 53: The First Frontier AI Law, Explained

Those sentiments have put Anthropic at odds with the White House and with Sacks in particular. Sacks played a major role in development the Trump administration’s AI Action Plan unveiled in July. The plan calls for significant investments in AI infrastructure and development while rolling back rules and regulations the White House views as holding back U.S. progress on the technology.

The plan also endorsed a called for scrapping state laws affecting AI use and development, echoing the failed effort to include a 10-year moratorium on state AI regulations in the Big Beautiful budget reconciliation bill.

Anthropic had called the moratorium “too blunt” an approach and later endorsed California’s AI safety bill signed into law by Gov. Gavin Newsom in September.

Amodei also publicly supported Kamala Harris’ presidential bid against Donald Trump and has been a notable no-show at White House events attended by his peers at OpenAI, Google, Amazon and Meta, among others.

Anthropic has also prohibited the federal government from using its Claude models in law enforcement activity, further fueling tensions with the White House.