Anthropic has embedded approximately six engineers at the National Security Agency to deploy its Mythos AI model for offensive cyber operations, the Financial Times reported Thursday. The engineers are customizing the model for specific applications, potentially including infiltrating networks in China and Iran, according to one source cited by the FT. The deployment occurred as Anthropic simultaneously published research calling for a coordinated global pause mechanism on AI development and filed for an IPO that could value the company above $1 trillion.
Anthropic Deploys Engineers at NSA for Mythos AI Operations
Anthropic placed about six engineers inside the National Security Agency to help deploy Mythos—its most capable AI model—for offensive cyber operations, according to the Financial Times report. The engineers are forward-deployed staff customizing the model for specific applications. One source told the FT the model could be useful for infiltrating networks in countries like China and Iran.
Mythos is the same model Anthropic has declined to release publicly, citing misuse risk. The company limited it to vetted partners through Project Glasswing—a restricted coalition that includes Microsoft, Apple, and Amazon. Whether the engineers are involved in active operations is not confirmed.
Anthropic Sues Pentagon Over Supply-Chain Risk Designation
Anthropic is suing the Pentagon after Defense Secretary Pete Hegseth designated the company a supply-chain risk in late February—a label historically reserved for foreign adversaries like Huawei. The designation followed the collapse of a $200 million contract. The sticking point: Anthropic refused to let the Department of Defense use Claude for fully autonomous weapons or domestic mass surveillance. The NSA contract was exempt from that ban.
A California judge blocked the blacklisting as an apparent First Amendment retaliation. A D.C. appeals court denied Anthropic's bid to halt the designation while litigation plays out. The NSA continued using Mythos throughout this period, according to the FT's reporting. The Pentagon's deadline to drop Claude from its systems falls in August.
Claude Writes Over 80% of Anthropic Production Code
On the same day the NSA story broke, Anthropic's internal research institute published "When AI Builds Itself," examining how far Claude has come at automating its own development. Claude now writes more than 80% of the code merged into Anthropic's production codebase—up from low single digits before Claude Code launched in early 2025. Engineers ship roughly eight times as much code per day as they did in 2024.
In April, Claude agents were handed an open AI safety problem—whether a weaker model can reliably supervise a stronger one—and left to run it. Two human researchers over about a week recovered 23% of the performance gap between the models. The agents recovered 97% over 800 cumulative compute hours. Humans set the question while the agents designed every experiment. The report's authors—Anthropic Institute lead Marina Favaro and co-founder Jack Clark—described it as the first published case of Claude exercising research judgment rather than just executing tasks someone else specified.
Anthropic Proposes Verifiable Global AI Development Pause
The report argues for a verifiable global pause—multiple frontier labs halting simultaneously with independent verification that everyone actually stopped. Anthropic stated it would join such a pause. The authors likened the proposed mechanism to Cold War-era nuclear treaties struck between the United States and Russia.
Anthropic warned that the trajectory is heading toward recursive self-improvement: AI systems that autonomously design, build, and train their own successors, with humans playing a diminishing role at every step. Once AI chooses which experiments are worth running—not just runs them—humans lose the last meaningful role in the development loop, according to the report. Small misalignments visible in today's models could compound across self-improving generations until nobody can correct them.
AI Safety Warnings Issued in 2023 Did Not Halt Development
In 2023, over a hundred big names in the AI researching community signed an open letter asking for a global effort to mitigate the risk of extinction that AI development intrinsically has. A few months before that, another open letter demanded OpenAI pause advances on ChatGPT due to its dangerous nature. Nobody stopped after the 2023 open letter. OpenAI didn't. Anthropic didn't.
FAQ
What did Anthropic deploy at the NSA?
Anthropic embedded approximately six engineers at the National Security Agency to deploy its Mythos AI model for offensive cyber operations, according to a Financial Times report published Thursday. The engineers are customizing the model for specific applications, potentially including infiltrating networks in China and Iran.
Why is Anthropic suing the Pentagon?
Anthropic is suing the Pentagon after Defense Secretary Pete Hegseth designated the company a supply-chain risk in late February following the collapse of a $200 million contract. Anthropic refused to let the Department of Defense use Claude for fully autonomous weapons or domestic mass surveillance. A California judge blocked the blacklisting as an apparent First Amendment retaliation, while a D.C. appeals court denied Anthropic's bid to halt the designation during litigation.
How much code does Claude write for Anthropic?
Claude now writes more than 80% of the code merged into Anthropic's production codebase, up from low single digits before Claude Code launched in early 2025. Engineers ship roughly eight times as much code per day as they did in 2024, according to Anthropic's "When AI Builds Itself" report published by the company's internal research institute.
