A satirical tweet from @beffjezos, the parody account riffing on Jeff Bezos, made a straightforward claim: everyone needs to own their own machine that extends their intelligence. The actual tweet sits behind X's JavaScript wall, so we can't quote it directly. But the premise kicked off a grounded Hacker News thread about who controls your cognitive tools.
Someone pointed to US telephone number portability, where regulators forced carriers to give users control over their own phone numbers. The same logic applies here. If your thinking extends into a cloud service you don't control, you're renting your own mind from a tech company.
The hardware to do this already exists. Apple's M-series chips ship with Neural Engines. Intel's Core Ultra and AMD's Ryzen AI processors include dedicated NPUs. Even the Raspberry Pi AI Kit, with its Hailo-8L accelerator, can run computer vision models offline. On the software side, tools like llama.cpp and Ollama let you run quantized models, think Llama 3 or Mistral compressed to 4-bit, on standard hardware. WebLLM takes it further, running models directly in your browser via WebGPU, no server required.
The HN thread split into two camps: people who want to own their intelligence infrastructure, and people talking about disconnecting entirely. That tension isn't going away. With a Raspberry Pi AI Kit and an open source model, you can run your own AI today, on your desk, without asking anyone's permission. That's what cognitive sovereignty looks like.