Glance is a browser extension that fact-checks posts in your X/Twitter feed as you scroll. A small icon sits next to each post, reads the content, and surfaces missing context, disputed claims, and things that are being pushed back on in the replies. It works on any Chromium-based browser. We waited 6 months before building this because the economics looked impossible. A decent quality AI fact-checking analysis is $0.05-0.15 per post, a typical user scrolls hundreds of posts per seconds. The only way to make the math work was a pipeline that leverages the comment section to triage posts and analyze in depth only when it's necessary. It works more or less like this: 1. Local filter in-browser (free): short posts, already-seen content. 2. Small-model triage: does this post even make a factual worth checking? 3. Comment analysis (main path): pull the replies, analyze them alongside the post. 4. Full web-search analysis: only when steps 1-3 can't decide. Average cost landed at ~$0.0015 per post, which looks sustainable with a subscription model, and can definitely be optimized. Comments URL: https://news.ycombinator.com/item?id=47779599 Points: 1 # Comments: 0
Coding agents made everything feel fast again - except managing release binaries. Operations like these should be real commands you can run from your terminal: - "Deploy the login fix to staging" - "What's the functional difference between production and staging?" - "Notify me when any security fix hits production" - "Promote Jen's last changes to staging" - "Share the new beta API SDK with the partners at Acme" That's what Fly does. Every time a CI run pushes artifacts to the Fly Registry, it creates a release - including the binaries produced, an AI-generated summary, captured architectural decisions, merged PRs and commits. All of it is searchable. So when you ask your agent "when did we release a fix for the connection timeout" it runs a semantic search across all of that and finds the right evidence. No version numbers, no digging through run IDs, no Git hashes. Fly also understands your runtime environments and works with whatever deployment framework your coding agent already knows - Argo CD, Flux, etc. It tracks what gets deployed through them automatically, so "what's the latest change in production right now" gets a real answer. A few other things worth knowing: - Integrates out of the box with your GitHub repositories - Works with Claude Code, Cursor, Codex, etc. via MCP - Agentless integration with your runtime environments - Slack notifications in natural language - Fly watches releases and notifies with context when something matches - Runs Artifactory as a scalable package registry under the hood - Supports Docker, Helm, npm, PyPI, Maven, Gradle, Go, and .NET, with more package types coming - Once installed, your package managers just work: npm publish, pip install, docker push - all route through your private Fly registry automatically. No .npmrc editing, no token wrangling Take it for a spin and upgrade the way you manage your release binaries - and finally get real visibility into WHAT is actually running where. Happy to get into how it works, what else it handles, and where we're taking it next. https://jfrog.com/fly Comments URL: https://news.ycombinator.com/item?id=47764383 Points: 4 # Comments: 0
今天更新了ecc之后就一直出现 ai会一直卡在这里然后就不走了。各位佬友有没有什么解决方案 1 个帖子 - 1 位参与者 阅读完整话题
I hated the fact that my Hermes Agent can't check emails unprompted without burning tokens uselessly on heartbeats so I asked it to make this. It's a skill to set up customized cron jobs for apps you care about that runs every 5 minutes and only wakes your agent up when there's a real notification Even works for delegated subagents notifying the orchestrator directly Comments URL: https://news.ycombinator.com/item?id=47740042 Points: 1 # Comments: 0
丸红旗下公司斥资2.04亿欧元收购Factor Energy 85%股份。（新浪财经）