White House AI and Crypto Czar David Sacks pushed back on Wednesday, 17 June 2026, against what he called a 'misleading narrative' in recent media coverage, asserting that he had consistently treated Anthropic's Mythos Preview AI model as a genuine cyber threat even while questioning the company's messaging tactics.

Context

The controversy centres on coverage of Sacks's public statements made in early April 2026, shortly after Anthropic launched Mythos Preview — a frontier AI model whose claimed offensive cyber capabilities triggered wide industry debate. Critics argued that Sacks had downplayed the threat; Sacks contends those accounts 'egregiously cherry-picked' his remarks by highlighting his scepticism of Anthropic's messaging while omitting his explicit calls for urgent defensive action.

On the All-In Podcast on 10 April 2026, just three days after Mythos Preview launched, Sacks said: 'With cyber, I actually would give them credit in this case and say, this is more on the real side.' He went on to urge 'every company, or IT department, or CISO that is managing code bases' to use the following months to detect dormant bugs and roll out patches.

Policy Backdrop

Sacks's position reflects a broader posture within the Trump administration — prioritising rapid, practical deployment of AI for cyber defence rather than regulatory pauses, while remaining sceptical of what officials see as alarmist narratives from certain AI labs. The administration has since its 2025 executive actions emphasised public-private cooperation on hardening critical systems against AI-enabled attacks.

In a 30 April 2026 post, Sacks described Mythos as 'the first of many models that can automate cyber tasks,' noting that such models 'do not create vulnerabilities; they discover them.' He argued that using AI to find and patch existing bugs would ultimately strengthen defences, but stressed that 'cyber defenders' must gain access to these tools before adversaries do — including those operating non-US or Chinese models.

Sacks also cited tech commentator Ben Thompson as having independently reached a similar assessment of Anthropic's communications record, framing his own critique as neither unique nor particularly controversial within informed technology circles.

Stakeholders and Impact

CISOs, IT departments, and national security agencies are the primary audience for Sacks's guidance. His call for immediate vulnerability scanning and patch deployment carries weight given his dual role as a senior White House official and a prominent Silicon Valley investor through Craft Ventures.

For Anthropic, the episode has sharpened a credibility debate: Sacks argued on 12 April that a growing number of observers were questioning whether the company had become 'the AI industry's boy who cried wolf,' and warned that the threats must materialise or the company faces a serious credibility problem. He added that Anthropic's 'needlessly confrontational posture toward the Administration has distracted from' the shared mission of cyber defence.

What's Next

Sacks framed the situation as a 'shot clock': Mythos-level capabilities will diffuse widely — including to non-US and Chinese models — and defenders must find and patch vulnerabilities before that window closes. He called for 'cooperation between government and industry' and urged all parties to move forward on 'tangible defensive action' rather than prolonged policy debates.

Analysts will watch whether the administration formalises joint vulnerability-disclosure and patching programmes with major tech companies, and how Anthropic responds to the administration's ongoing criticism of its public communications strategy around AI risk.