This project is a ready-to-run boilerplate for operating a "lights-on" software factory with Claude Code, built from the setup used on Votito. It argues that moving routine orchestration from LLM inference into deterministic tooling dramatically cuts token use and speeds workflows - the example orchestrator reportedly reduced token spend by about 80% versus a bare-bones Claude workflow. The package is presented as both a practical starter kit and an educational demo of software-factory concepts: secure sandboxing, deterministic command regulation, automated lint-driven feedback, a simple plan/implement workflow, and a generator that turns plans into multi-agent Claude workflow scripts.
Concrete components include a container image and makefile for running Claude in an isolated workspace, a bash-command regulator that deep-inspects commands and issues automated approve/deny decisions with corrective guidance, and a PostToolUse linter hook (an eslint plugin with example rules) that runs on every JavaScript edit to enforce domain rules and reduce manual review. The orchestrator moves gating, testing and script generation into command-line tools, while workspace/.claude contains agent, command and workflow code plus tests. Documentation files cover setup, settings, regulator rules, linter rules, orchestrator internals and usage; the project is MIT-licensed.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
MUSE is a new benchmark for evaluating text-to-CAD models on complex assembly tasks, focusing on functionality, manufacturability, and assemblability. Experiments reveal current models struggle with producing engineering-ready designs, highlighting the need for improved AI tools in industrial product design. (emergentmind.com)
Many people develop AI projects privately and consider sharing ideas to inspire others or incorporate feedback. Examples include a Linux desktop picture viewer by date, highlighting common needs in file management tools. (news.ycombinator.com)
![Tenstorrent raises $1.5B [$6.2B valuation] to expand its open compute ecosystem](https://hntoday.pixelshot.io/screenshots/20261020/c163ac92861a43ce_tn.jpeg)
Tenstorrent, an AI hardware startup led by semiconductor legend Jim Keller, raised $1.5 billion to expand its open compute ecosystem and challenge Nvidia. The company's scalable, energy-efficient AI platforms use RISC-V based accelerators and are already deployed in enterprise applications like AI trading agents. (siliconangle.com)
Token-maxxing has declined as OpenRouter's token usage plateaus, indicating a slowdown in growth. The platform's rankings show stable usage but no significant increase in token processing across models. (openrouter.ai)
Open weights models like Ornith-1.5 have largely disappeared, with recent releases becoming infrequent. Developers now appear more focused on monetization rather than consistent open-source releases. (news.ycombinator.com)
Company World Models are AI systems designed to observe business activities, generate strategic hypotheses, and coordinate actions through digital employees to improve company performance. They aim to enable better decision-making and growth for essential businesses by integrating AI as a strategic partner. (beaconholdings.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.