The facts
The US Department of Homeland Security (DHS) has banned the use of all artificial intelligence (AI) technology from the Chinese firm DeepSeek within civilian government agencies, and any technology developed with its AI models. The DHS made the order public on Monday, but the ban took effect from the end of July. The order specifically targets civilian agencies such as the Department of the Treasury, the Department of State, and the Department of Justice, while excluding agencies within the Department of Defense. DeepSeek is a popular and relatively inexpensive open-source AI model from China, seen as a disruptive force in the global AI technology race. While the DHS notice did not specify the exact security concerns, the White House has previously criticized Chinese AI firms. Michael Kratsios, White House tech lead, wrote earlier this year about the "malicious exploitation" of American AI firms, primarily by Chinese rivals. Companies like Anthropic and OpenAI have also claimed that some Chinese AI firms, including DeepSeek, have copied their AI models. US President Donald Trump has identified China as a key rival in AI development. DeepSeek's models have gained significant popularity since their initial release in 2025, with its free chatbot app being a top download and its AI models being incorporated by US tech companies like Snowflake, Vercel, and OpenRouter. Startups favor DeepSeek due to its lower cost compared to closed models like Google's Gemini or OpenAI's Codex. DeepSeek stated that training its debut model cost approximately $6 million, significantly less than the estimated $100 million for OpenAI's GPT-4.
Our take
Well, here we go again. The U.S. government, in a stunning display of bureaucratic cowardice, has decided the real national security threat isn’t our own crumbling infrastructure or the fact that half our politicians are geriatric fuckwits - it’s some cheap-as-fuck Chinese AI code. The Department of Homeland Security, a collection of desk jockeys who probably think a large language model is a new kind of bra, has ordered a full-scale purge of DeepSeek from civilian agencies. Because nothing says “secure homeland” like banning a tool that costs less to train than a single episode of a shitty network procedural drama. They spent six million bucks. OpenAI blows through a hundred million just on coffee and beanbag chairs for their engineers. The fucking disparity is hilarious. The U.S. is terrified of a bargain-bin disruptor because it proves you don’t need a Silicon Valley dick-measuring contest budget to make something people actually want to use. Of course, they won’t ban it from the Pentagon. The war machine needs its affordable knock-offs, apparently. Can’t have the DoD paying full price for genocide-adjacent algorithms when a discount version will do. This is pure, unadulterated protectionism wrapped in a flag and screaming about “malicious exploitation.” Translation: We got caught with our pants down, getting out-innovated by a company that didn’t burn a billion in VC cash on office slides and kombucha taps. They’re scared shitless because startups are choosing the cheap, effective Chinese model over the overhyped, overpriced American bloatware. It’s a tantrum. A pathetic, xenophobic, innovation-stifling tantrum thrown by a bunch of grifters and pricks who’ve realized the easiest way to “win” a tech race is to confiscate the other guy’s fucking shoes. The White House tech lead, some clown named Kratsios, whined earlier this year about Chinese rivals. Cry more, you subsidized hack. Maybe if your precious “American AI firms” stopped allegedly copying each other and started building something that doesn’t cost a fortune to run, you wouldn’t need Daddy Government to ban the competition. This isn’t security. It’s the death rattle of a tech empire that’s become too fat, too lazy, and too fucking expensive to compete. What a bunch of cunts.
