{
    "success": true,
    "data": {
        "id": 1714884,
        "msgid": "singapores-foreign-ministers-homemade-ai-built-using-raspberry-pi-1777824255",
        "date": "2026-05-03 13:11:00",
        "title": "Singapore's Foreign Minister's Homemade AI Built Using Raspberry Pi",
        "author": "Soffya Ranti",
        "source": "KOMPAS",
        "tags": "",
        "topic": "Technology",
        "summary": "Singapore's Foreign Minister, Dr. Vivian Balakrishnan, has developed a personal AI assistant using open-source tools and a Raspberry Pi to support his diplomatic activities, including answering technical questions, conducting research, drafting speeches, and providing daily briefings. The system, which operates entirely locally to ensure data privacy, incorporates knowledge graphs from his personal documents and integrates with messaging apps, demonstrating how diplomats can gain a competitive edge by collaborating with AI. Balakrishnan, a former ophthalmologist with a systematic approach, views this as a practical step beyond mere policy debates on artificial intelligence.",
        "content": "<p>Amid the heated debates on artificial intelligence (AI) regulations\namong world leaders, Singapore\u2019s Foreign Minister, Dr.\u00a0Vivian\nBalakrishnan, has taken a far more concrete step. Rather than just\ndiscussing it, he has openly introduced a personal AI assistant that he\ndesigned and built himself to support his daily diplomatic activities.\nBalakrishnan describes his virtual assistant as a \u201csecond brain\u201d for a\ndiplomat. Its capabilities include answering technical questions,\nconducting topic research, drafting speeches, and providing daily\nbriefings. \u201cThis system has become invaluable; I wouldn\u2019t even dare to\nturn it off!\u201d he wrote in a Facebook post. Balakrishnan is not just a\npolitician jumping on the technology bandwagon. He is a former\nophthalmologist who graduated from the National University of Singapore\nand received a Presidential Scholarship in medicine in 1980, making him\nsomeone accustomed to systematic, evidence-based thinking. The virtual\nassistant system is built on two open-source foundations. The first is\nNanoClaw, an independent AI assistant based on the Claude model by\ndeveloper Gavriel Cohen, which runs locally on a Raspberry Pi and\nconnects directly to various messaging applications such as WhatsApp,\nTelegram, Slack, and Discord. Secondly, Balakrishnan applies the \u201cLLM\nWiki\u201d pattern proposed by former Tesla AI Director Andrej Karpathy, as a\nsolution to the AI\u2019s \u201camnesia\u201d problem, where it often loses\nconversation context with each new session. The system works by\nabsorbing all of Balakrishnan\u2019s speech drafts, articles, and web\nclippings, then processing them into a structured knowledge graph. Every\ntime it receives a question, the AI performs a semantic query and\ninjects relevant facts into its response, making it smarter the more it\nis used. For Balakrishnan, AI is not just a policy debate topic.\n\u201cDiplomats who learn to work with AI will have a significant advantage,\nand I believe that advantage starts now,\u201d he concluded. The tool named\nMnemon stores information in a SQLite database, which is then converted\ninto wiki pages and can be accessed via the Obsidian app on macOS or\niOS. The semantic search process runs entirely locally using the Ollama\nplatform on a Raspberry Pi 5, while voice message processing is done\ndirectly on the device via whisper.cpp. Thus, no confidential\nconversations have ever touched external servers. Furthermore, each chat\ngroup has its own Docker container and local memory, ensuring that data\nbetween groups remains securely isolated.<\/p>",
        "url": "https:\/\/jawawa.id\/newsitem\/singapores-foreign-ministers-homemade-ai-built-using-raspberry-pi-1777824255",
        "image": ""
    },
    "sponsor": "Okusi Associates",
    "sponsor_url": "https:\/\/okusiassociates.com"
}