Pacer is a lightweight macOS menu-bar utility that tells whether AI coding subscriptions (Claude Code, Codex, Antigravity, Cursor, Gemini/Copilot and others) will last until their next reset by displaying "speed" - the projected percent of a billing/session window at reset if current usage continues. Two concentric rings show session and week projections for the active account; green ≤80%, amber ≤100%, and a solid red disc if projection exceeds 100%. Clicking the icon opens a card with per-account/session/week speeds, a single actionable sentence of advice (e.g., reduce or shift model usage), per-model cost-weighting inferred from your local transcripts, and notifications when pace crosses thresholds. It also reports API-equivalent cost over rolling day/week/30-day windows.
The tool is implemented as one Node script plus a Swift menu-bar app, samples every 10 minutes (launchd/Homebrew service), stores a JSON status file (~/.claude/pacer/status.json) other tools can read, and requires macOS 13+, Node ≥18 and Xcode CLT. Projections are computed as used% divided by elapsed fraction for week windows and by a recent-rate extrapolation (last 45 minutes) for a 5-hour session; the headline is the worst window. It reads providers' own usage endpoints via logins already stored on the machine (never refreshing or writing tokens), fits per-model weights from local transcripts in ~/.claude/projects, and supports multiple remembered accounts with per-account windows and alerts.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
An AI agent was able to design a complete RISC-V CPU based on a 219-word specification sheet within 12 hours. This demonstrates significant advancements in automated hardware design using artificial intelligence. (tomshardware.com)
AI researchers estimate a 10% median chance that AI could cause human extinction or severe disempowerment, up from 5% a year earlier. Experts and forecasters remain cautious, acknowledging the uncertainty and potential risks of advanced AI development. (virev.ai)
OpenAI released an AI model designed to solve mathematical problems, prompting a response from mathematician Terence Tao. The event raises questions about AI's role in advancing mathematics research and education. (mathstodon.xyz)
Self-modeling interventions, which involve a model recognizing and reporting on itself, can influence and reduce emergent misalignment in AI systems. These findings suggest that metacognitive processes may play a role in shaping how models generalize and behave, offering potential pathways for AI safety improvements. (lesswrong.com)
An AI developer has launched open-source, Adobe-like creative software suites using AI reverse engineering, aiming for full feature parity. The projects, built with Anthropic’s Claude Opus, are in early alpha but plan to reach near-complete functionality within months. (arstechnica.com)
A new open vision-language model has been developed for various medical applications, enabling improved interpretation of medical data. The model aims to enhance diagnostics and treatment through advanced AI understanding of medical images and language. (nature.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.