Llama-modchip: 284B DeepSeek on 64GB RAM, no GPU
Llama-modchip aims to run a 284-billion-parameter DeepSeek model using 64GB of RAM without a GPU. The provided information does not explain its technical approach. (github.com)
A DIY woodworker recreates an expensive professional up-and-over tool chest: closed, it is a carry box with a top handle; open, its front swings over the handle and lies flat on the body, forming a taller cabinet. The builder used Claude Opus 5.5 to develop the mechanism, starting with a Python hinge simulation and proceeding to 3D models, drawings, parts lists, sourcing, and laser-cut steel files. The finished chest uses 12 mm birch plywood, 5 mm steel links, and a beech-dowel handle, measures 486 × 350 × 268 mm when closed, and stands 701 mm tall when open. Materials and parts cost about €220.
The guide explains the four-bar hinge design, cutting and assembling the two trapezoidal box sections, mounting the handle plates, and fitting pivots only after the body is complete so they align accurately. The simulation reports 109.6 mm and 121.5 mm links, a 0.03 mm error when the lid rests flat, 16.4 mm minimum handle clearance, and no collision with the body. Plans, templates, CAD files, and a parts list accompany the instructions. The builder says the project worked well, apart from replacing a panel cut incorrectly, and found the finished chest useful for carrying household tools together. Key takeaways include the use of three-position Burmester synthesis for hinge design and shoulder screws and threaded inserts in woodworking; the AI-assisted process consumed about 28 million tokens.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Llama-modchip aims to run a 284-billion-parameter DeepSeek model using 64GB of RAM without a GPU. The provided information does not explain its technical approach. (github.com)
A GNU maintainer’s controversial take prompted debate about AI-generated code in open-source projects. Commenters discussed project leaders who use AI tools while distancing themselves from code ownership. (social.treehouse.systems)
ThinkStream says AI can generate work, but people still need to understand it. The provided page content contains only an anti-bot prompt. (thinkstream.app)
Ensemble is a book and open-source project for building an AI coding agent that helps create the next edition. It advocates development-driven specifications, where specs evolve alongside implementation instead of dictating development upfront. (github.com)
Maple uses manifold-activated pruning to compress mixture-of-experts language models below 2 bits per weight. Its developers say it runs Gemma 4 26B at 45 tokens per second on a 6GB GPU. (github.com)
A longtime Claude Code user is switching to Pi after growing frustrated with incompatible instruction-file formats, limited VSCode session support, and bugs in the GUI extension. They say Claude Code has prioritized adding features over fixing reliability and usability problems. (caleb.software)
Today's best Hacker News stories, summarized and screenshotted, one email a day.