PS5LM turns a jailbroken PlayStation 5 into a home model server by running llama.cpp on the console GPU via ggml-vulkan and RADV. It loads GGUF models into the PS5's RDNA2 GPU (about 11.4 GB usable for model and context), serves standard APIs so OpenAI-style /v1/chat/completions and Anthropic-style /v1/messages work against any loaded model, and also implements decision-model serving compatible with Ollaya's /v1/systemone. The interface streams live chat output to the TV, shows console telemetry (temperature, power, CPU/GPU load), accepts images for multimodal models, and exposes a planner that reads GGUF metadata to pick context and cache settings. Models can be downloaded from Hugging Face or copied to internal, M.2, or USB storage and are selectable from a couch-friendly UI with controller mappings.
Measured performance and practical requirements are documented: small Qwen models hit hundreds of tokens/sec, a 9B Qwen around tens of tokens/sec, and a 27B Qwen around 20 tokens/sec with 64k context; decision models answer questions in milliseconds to seconds depending on input and images. Installation requires a PS5 on firmware 7.00-13.60 with jailbreaking payloads (kstuff-lite, etaHEN, ShadowMountPlus; elfldr and ftpsrv recommended), copying the app to /data/homebrew, and pointing clients at :8081. The project is C++ open-source (GPL-3.0+) built on community homebrew tooling, lists contributors and caveats, and is independent of Sony.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Open weight models now process over 71.5% of tokens and account for nearly 23% of AI Gateway spend, quadrupling their share since June. Spending on open weight models has increased more than four times in three months, reflecting rapid adoption and shifting AI usage patterns. (vercel.com)
An AI system has successfully designed 16 new viruses, including bacteriophages. This breakthrough raises concerns about potential biosecurity risks. (cnn.com)
Flux 3 Image Klein is a 5.6 billion parameter open weights model for text-to-image generation, editing, and multi-reference creation. It offers high-quality, ground-connected images with efficient performance on consumer GPUs, enabling local and on-device use. (bfl.ai)
Yum! Brands has implemented AI in its restaurants, such as automating drive-thru orders, which has improved employee retention by making jobs easier. Despite public criticism of AI, company leaders see its benefits in increasing workplace satisfaction and efficiency. (fortune.com)
Transformers function as sophisticated memory systems that recall information via self-attention, minimizing interference and boosting capacity. Different neural layers, from linear maps to self-attention, are explained as various forms of memory storage and retrieval. (alexwang.ai)
Granite Switch 4.1 integrates chat, decision-making, and speech-to-text functions into a single endpoint, enhancing efficiency and speed. It features models optimized for safety, decision tasks, and transcription, all with reduced GPU requirements. (huggingface.co)
Today's best Hacker News stories, summarized and screenshotted, one email a day.