🌎 Monitor the world

Same AI agent. Same task. Same model. Token cost varied by 30x. A new study from researchers at Michigan, Stanford, MIT, Google DeepMind & All Hands AI just exposed the biggest blocker to enterprise AI agent deployment: cost predictability. They analyzed 8 frontier models (GPT-5, Claude Sonnet 4.5, Gemini 3 Pro, Kimi-K2, etc.) on real coding tasks. The findings are wild: 👉 Agentic tasks consume ~1,000x more tokens than a regular chat or reasoning task. 👉 The same task, same model, same setup → token usage varied by up to 30x across runs. 👉 On identical workloads, Kimi-K2 and Claude Sonnet 4.5 burned 1.5M more tokens than GPT-5. The most counterintuitive finding? More tokens did NOT mean better outcomes. Accuracy peaked at intermediate cost — then flattened or declined. Expensive runs were dominated by repeated file views, redundant edits, and aimless exploration. The agents weren’t reasoning harder. They were looping. Even more striking: when asked to predict their own token cost before execution, frontier models hit only 0.39 correlation at best — and systematically underestimated real usage. This is why the AI infrastructure layer is about to explode: → Orchestration → Caching → Observability → Budget enforcement → Routing & model selection → Memory management Each becomes mission-critical. It’s also why enterprises can’t deploy agents at scale yet. The bottleneck has shifted. The challenge is no longer model intelligence. It’s controlling the variance of autonomous execution. Paper: “How Do AI Agents Spend Your Money?” (Bai et al., 2026) arxiv.org/abs/2604.22750

Hot take: AI won’t cheapen human work. It’ll nuke the middle and split the market for thought in two. 🔻One side: infinite output. Drafts, code, decks, analysis, strategy docs. A teenager with Claude and Codex now ships what used to need a staffed team and three weeks of meetings. 🔺Other side: judgment. Knowing what’s true. What’s noise. What breaks at scale. What compounds. What’s a hallucination wrapped in confidence. The first market is collapsing into abundance. The second is becoming a scarcity market. When everyone can generate, the only thing that matters is taste, discernment, and the ability to say: “this is the one worth betting on.” AI makes production cheap. Conviction gets expensive.

Major life hack: Don't complain, ever. Nobody likes a complainer. They drain the energy of everyone around them. It's exhausting spending time around someone who constantly complains about things outside their control. If it’s within your control, go do something about it. If it’s not, you’re just wasting energy thinking about it. Complaining gives too much power to the thing. Take back that power.

1/ Wingbits is now powering the flight tracking and GPS interference layer on @worldmonitorai, one of the fastest-growing open-source intelligence platforms in the world. Excited to be working with @eliehabib on this. 🧵

Askeri, finans, enerji, havacılık, siber tehditler, altyapı verileri.. tek arayüzde birleştiren YZ destekli açık kaynaklı küresel durum izleme platformu. Ollama desteği ile lokalde çalıştırabiliyor. @eliehabib ▪️Github: github.com/koala73/worldmoni… ▪️Deneme: worldmonitor.app

😭 😭 😭

别只盯新闻了，这个工具能帮你看到世界真实运行的样子 大多数人看世界的方式是：刷新闻。 今天哪里打仗了，明天哪里降息了，后天哪个大佬发推特了。 问题是，新闻是被编辑筛选过的。你看的是别人想让你看的。 如果你想自己判断局势——不是看结论，而是看原始信号——你需要的是情报工具，不是新闻客户端。 WorldMonitor 是什么 一个开源的全球态势感知系统。 核心能力： 435 个信息源，15 个类别 军事、金融、能源、气候、航空、网络攻击、基础设施……全部聚合在一个界面里。 不是简单的 RSS 聚合。它用本地 AI（Ollama）对信息做合成简报，交叉验证，风险评分。 双地图引擎 45 个数据图层 3D 地球 WebGL 平面地图。 图层包括： • 军事部署和演习 • 灾害位置和影响范围 • 经济数据变化 • 基础设施状态（电网、港口、管道） • 航班实时追踪 跨流关联分析 系统不只是展示数据。它会自动关联： • 军事信号 能源价格 • 灾害信号 供应链中断 • 经济数据 市场异动 每个国家有一个"情报指数"，覆盖 12 个维度的风险评分。 金融雷达 92 个股票交易所、大宗商品、加密货币，7 指标综合分析。 不只是价格，还有市场结构和资金流动信号。 谁适合用 不是所有人。 如果你只是想知道"今天发生了什么"，刷新闻就够了。 但如果你需要： • 做地缘风险研判（咨询、智库） • 管理跨境业务（能源、物流、供应链） • 做宏观投资（需要对冲尾部风险） • 做开源情报研究（OSINT） 这个工具值得看。 特别是它的本地 AI 能力。所有分析都在你自己的机器上跑，数据不上传。 在地缘情报领域，这点很重要。 怎么装 两种方式。 方式一：直接下载桌面客户端 支持 macOS（ARM / Intel）、Windows、Linux AppImage。 下载地址：worldmonitor.app/api/downloa… 方式二：自己部署 git clone github.com/koala73/worldmoni… cd worldmonitor npm install npm run dev 启动后打开 localhost:5173，不需要任何环境变量。 如果想用本地 AI 分析，需要安装 Ollama。 一个细节 项目用 AGPL-3.0 协议。 个人、研究、教育用途免费。商业用途需要单独授权。 如果你想做 SaaS 或者商业化产品，需要联系作者买许可证。 这不是"免费开源随便用"那种项目。作者很明确：非商用免费，商用付费。 为什么值得试 大多数情报工具是闭源的、收费的、面向企业客户的。 WorldMonitor 是开源的，可以自己部署，支持本地 AI。 它不是给你"结论"的工具。它是给你"信号"的工具。 新闻告诉你"发生了什么"。 这个工具帮你看到"正在发生什么"——以及不同事件之间的关联。 如果你需要独立判断局势，而不是被喂结论，试试看。 项目地址：github.com/koala73/worldmoni… (github.com/koala73/worldmoni…) 在线体验：worldmonitor.app 文档：docs.worldmonitor.app

News been heavy? This helped me zoom out.. 👀😶

Gość stworzył duży, interaktywny agregator monitorujący w czasie rzeczywistym wiadomości z całego świata, podzielony na 15 kategorii (konflikty, technologia, finanse itd). To jest projekt który ma w przeciągu miesiąca najwięcej gwiazdek na githubie - może kogoś zainteresuje 😉 worldmonitor (.) app

World Monitor: Real-time global intelligence dashboard - AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface. GitHub: github.com/koala73/worldmoni…

🚨معلومة | منصة ذكاء اصطناعي تراقب الحروب وسلاسل الإمداد لحظة بلحظة منصة WorldMonitor تراقب بيانات الطائرات والملاحة والشحن العالمي لرسم خريطة مباشرة للصراعات والمخاطر الجيوسياسية، تقارير تشير إلى أن هذا النوع من التحليل قد يقلل أخطاء التوقع بـ25٪.

I want world monitor on every screen in that bar x.com/ohryansbelt/status/202…

In a time of escalating conflict, parsing news and data to make sense of ongoing events is an overwhelming task. LAU alumnus and trustee @eliehabib (BS '94), CEO of Anghami and OSN , took it upon himself to build a solution: World Monitor, an open-source, AI-powered platform that maps global events in real time, drawing from 450 verified sources without editorial interference or delay. It now serves 4 million visitors across 174 countries. Read the full interview: news.lau.edu.lb/2026/lau-alu…

The Entrepreneur Middle East team remains on the ground in the #UAE, navigating this uncertainty alongside you—something our second March 2026 issue, 𝐓𝐡𝐞 𝐑𝐞𝐬𝐢𝐥𝐢𝐞𝐧𝐜𝐞 𝐈𝐬𝐬𝐮𝐞 𝐏𝐚𝐫𝐭 𝟐, clearly reflects. This issue’s cover features @eliehabib, co-founder of @anghami, and founder of #WorldMonitor, who anticipated the world’s growing need for real-time insight amid rising uncertainty. Also in this issue, we honor the crisis leadership of the UAE’s President H.H. Sheikh Mohamed bin Zayed Al Nahyan (@MohamedBinZayed), Minister of State for International Cooperation H.E. Reem Al Hashimy, and Dubai Crown Prince, UAE Deputy Prime Minister and Minister of Defence, and Chairman of the Executive Council of Dubai, H.H. Sheikh Hamdan (@HamdanMohammed). We also bring you a collection of founder stories on how #GCC businesses are responding to regional tensions — hear from: @Virtuzone_UAE, @Buzzybeedubai, #Minoa, @VentureSouq, #CruiseMaster, the British Chamber of Commerce Dubai, INAS EXIM, The Lab Holding, @CarabiaCars, and @Pangaea_X_. Read all this and more on @issuu: issuu.com/entmagazineme/docs…

WORLDMONITOR IS EXPANDING FROM WAR TRACKING INTO AI FORECASTS AND GLOBAL SUPPLY CHAIN INTELLIGENCE. It now pulls real time signals on the world, from critical corridors and shipping rates to weighted forecasts built from live data.