New Free AI Update Lets You Run Chatbots Offline
This could let you chat with AI even without the internet — and for free
Deep Dive
llama.cpp just dropped a new pre-release (b10764), featuring a commit that removes unused variables in the ggml-cuda module. The update includes verified signed builds for a broad range of platforms — Apple Silicon, Intel Macs, iOS, Linux (CPU, Vulkan, ROCm, OpenVINO, SYCL), Android, and Windows with multiple GPU backends like CUDA and Vulkan.
Key Points
- llama.cpp updated to run AI chatbots faster on your own device — no internet needed
- Works on old phones, cheap laptops, even Linux servers — not just high-end computers
- Privacy-friendly and free — your data stays on your device, not sent to a company
Why It Matters
You could soon chat with AI anytime, anywhere — even offline — without paying or giving up your privacy.