bootgpt

NIGHTRUN is a local LLM runtime written in Rust that boots directly from a USB stick on ordinary x86_64 PCs or from an SD card on a Raspberry Pi 5, without a conventional operating system underneath. Its UEFI application loads a quantized model into RAM, renders a chat terminal through the framebuffer, and runs inference locally and offline without a network stack, host process, or kernel. The installer builds and flashes a bootable image, guides model selection and verified downloads, and supports GGUF inspection and conversion into NIGHTRUN's .nrm container. During boot, the model is streamed into RAM with checksummed sections; storage is then sealed so generation cannot read from disk. The runtime includes custom framebuffer rendering, USB keyboard input, multi-core inference, serial diagnostics, hand-written quantized kernels for x86_64 and Raspberry Pi architectures, and Q8_0, Q4_K, and Q6_K weight formats. It requires Linux, Rust, and approximately 6 GB of free disk space for the model and image.

gbos-vm runs a real Googlebook OS recovery image in a GPU-accelerated virtual machine on Apple Silicon Macs. Its install script downloads and verifies Google's ARM64 recovery image, replaces hardware-specific components with virtual-device components from Android Cuttlefish, and boots the result under QEMU with Vulkan translated to Metal through MoltenVK; the project also builds patched virglrenderer, Mesa, a macOS viewer, and an Android pointer helper. It packages the VM as a macOS app with clipboard integration, configurable display, memory, CPU, networking, audio, Bluetooth, and pointer modes. The project is intended as a development VM rather than a verified Google or Dell product; it requires an Apple Silicon Mac and substantial local storage for downloads and build artifacts, and is licensed MIT for its own code, with one Apache-2.0 patch.

mini-AGI is an experimental continual-learning, byte-level language model that trains from scratch on a CUDA-capable GPU with at least 8 GB of VRAM. It reads 256 byte values directly rather than using a conventional tokenizer, processes data one chunk at a time, and uses the same forward path for training and generation. Its architecture combines dense prelude blocks with a recurrent block applied up to 24 times, adaptive PonderNet-style halting, and per-application top-8 routing through a dynamically growing and pruning expert pool. Expert weights and Adam moments are stored as files on disk; a working set is paged through RAM and VRAM as needed, allowing the pool size to exceed available VRAM. The project includes Python commands for building corpora, reading files, streaming training, serving a local web interface, and inspecting or replicating continual-learning experiments. The repository describes the current model as toy-level rather than frontier-capable, and its weights are generated in the local weights directory rather than distributed with the source.

WormGPT is a named AI chatbot reported to be designed and marketed for cybercriminal use, offering guidance and code related to malware, phishing, and intrusion techniques. It emerged in reporting from late 2023 and was discussed by cybersecurity firms and in underground forums as an illicitly purposed generative AI service.

GPT with Taste is a behavioral recipe and system-instruction package for GPT-based coding harnesses, developed by NB-Corp. It provides a single English authoring recipe that compiles into host-specific instructions and event steering for OpenAI Codex, Claude Code, Kiki, Pi, Oh My Pi, and DeepSeek Harness. The instructions promote sensible defaults, continued execution, synthesis of delegated-agent results, and attention to engineering and visual details. For Codex, the installer writes managed prompt files and can add hooks that refresh scoped reminders when users submit work, child agents start, or supported wait results arrive. It supports a prompt-only installation or a Node.js-based installation with hooks, requires no npm packages, and leaves the model, provider, authentication, permissions, MCP servers, roles, and AGENTS.md unchanged. The repository includes harness porting, authoring, recovery, research, and comparison documentation and is released under the MIT License.