{"id":1002,"date":"2026-04-19T00:30:36","date_gmt":"2026-04-18T15:30:36","guid":{"rendered":"https:\/\/itexplore.org\/jp\/columns\/ai-evolution-frontier-life-sciences-robotics-desktop-ai\/"},"modified":"2026-04-19T00:30:36","modified_gmt":"2026-04-18T15:30:36","slug":"ai-evolution-frontier-life-sciences-robotics-desktop-ai","status":"publish","type":"post","link":"https:\/\/itexplore.org\/jp\/columns\/ai-evolution-frontier-life-sciences-robotics-desktop-ai\/","title":{"rendered":"AI\u9032\u5316\u306e\u6700\u524d\u7dda\uff1a\u30e9\u30a4\u30d5\u30b5\u30a4\u30a8\u30f3\u30b9\u3001\u30ed\u30dc\u30c6\u30a3\u30af\u30b9\u3001\u30c7\u30b9\u30af\u30c8\u30c3\u30d7AI\u306e\u9769\u65b0"},"content":{"rendered":"<p>\u672c\u65e5\u306e\u6ce8\u76eeAI\u30fb\u30c6\u30c3\u30af\u30cb\u30e5\u30fc\u30b9\u3092\u3001\u5c02\u9580\u7684\u306a\u5206\u6790\u3068\u5171\u306b\u304a\u5c4a\u3051\u3057\u307e\u3059\u3002<\/p>\n<div class=\"wp-block-vk-blocks-alert vk_alert alert alert-warning has-alert-icon\">\n<div class=\"vk_alert_icon\">\n<div class=\"vk_alert_icon_icon\"><i class=\"fa-solid fa-triangle-exclamation\" aria-hidden=\"true\"><\/i><\/div>\n<div class=\"vk_alert_icon_text\"><span>Warning<\/span><\/div>\n<\/div>\n<div class=\"vk_alert_content\">\n<p>\u3053\u306e\u8a18\u4e8b\u306fAI\u306b\u3088\u3063\u3066\u81ea\u52d5\u751f\u6210\u30fb\u5206\u6790\u3055\u308c\u305f\u3082\u306e\u3067\u3059\u3002AI\u306e\u6027\u8cea\u4e0a\u3001\u4e8b\u5b9f\u8aa4\u8a8d\u304c\u542b\u307e\u308c\u308b\u53ef\u80fd\u6027\u304c\u3042\u308b\u305f\u3081\u3001\u91cd\u8981\u306a\u5224\u65ad\u3092\u4e0b\u3059\u969b\u306f\u5fc5\u305a\u30ea\u30f3\u30af\u5148\u306e\u4e00\u6b21\u30bd\u30fc\u30b9\u3092\u3054\u78ba\u8a8d\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/div>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">GPT-Rosalind\uff1a\u30e9\u30a4\u30d5\u30b5\u30a4\u30a8\u30f3\u30b9\u7814\u7a76\u5411\u3051\u65b0\u30e2\u30c7\u30eb<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Introducing GPT-Rosalind for life sciences research | OpenAI<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p>OpenAI\u306f\u3001\u751f\u7269\u5b66\u3001\u5275\u85ac\u3001\u30c8\u30e9\u30f3\u30b9\u30ec\u30fc\u30b7\u30e7\u30ca\u30eb\u533b\u7642\u306e\u7814\u7a76\u3092\u52a0\u901f\u3055\u305b\u308b\u305f\u3081\u306e\u30d5\u30ed\u30f3\u30c6\u30a3\u30a2\u63a8\u8ad6\u30e2\u30c7\u30eb\u300c<strong>GPT-Rosalind<\/strong>\u300d\u3092\u767a\u8868\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u30e2\u30c7\u30eb\u306f\u3001\u5316\u5b66\u3001\u30bf\u30f3\u30d1\u30af\u8cea\u5de5\u5b66\u3001\u30b2\u30ce\u30df\u30af\u30b9\u306b\u304a\u3051\u308b\u6df1\u3044\u7406\u89e3\u3068\u30c4\u30fc\u30eb\u5229\u7528\u306e\u6539\u5584\u3092\u7d44\u307f\u5408\u308f\u305b\u3001\u79d1\u5b66\u7684\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u306b\u6700\u9069\u5316\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u65b0\u85ac\u958b\u767a\u306b\u306f\u901a\u5e3810\u301c15\u5e74\u304b\u304b\u308a\u3001\u7814\u7a76\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u306e\u8907\u96d1\u3055\u304c\u9032\u6b69\u3092\u59a8\u3052\u3066\u3044\u307e\u3059\u3002<strong>GPT-Rosalind<\/strong>\u306f\u3001\u8a3c\u62e0\u306e\u7d71\u5408\u3001\u4eee\u8aac\u751f\u6210\u3001\u5b9f\u9a13\u8a08\u753b\u306a\u3069\u306e\u591a\u6bb5\u968e\u7814\u7a76\u30bf\u30b9\u30af\u3092\u652f\u63f4\u3057\u3001\u7814\u7a76\u8005\u304c\u3088\u308a\u8fc5\u901f\u306b\u767a\u898b\u306e\u521d\u671f\u6bb5\u968e\u3092\u9032\u3081\u308b\u3053\u3068\u3092\u76ee\u7684\u3068\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u30e2\u30c7\u30eb\u306f\u3001<strong>ChatGPT<\/strong>\u3001<strong>Codex<\/strong>\u3001\u304a\u3088\u3073API\u3092\u901a\u3058\u3066\u3001\u9069\u683c\u306a\u9867\u5ba2\u5411\u3051\u306b\u7814\u7a76\u30d7\u30ec\u30d3\u30e5\u30fc\u3068\u3057\u3066\u63d0\u4f9b\u3055\u308c\u3066\u3044\u307e\u3059\u3002\u307e\u305f\u3001\u79d1\u5b66\u8005\u304c50\u4ee5\u4e0a\u306e\u79d1\u5b66\u30c4\u30fc\u30eb\u3084\u30c7\u30fc\u30bf\u30bd\u30fc\u30b9\u306b\u30e2\u30c7\u30eb\u3092\u63a5\u7d9a\u3067\u304d\u308b\u3001\u81ea\u7531\u306b\u30a2\u30af\u30bb\u30b9\u53ef\u80fd\u306a<strong>Life Sciences research plugin for Codex<\/strong>\u3082\u5c0e\u5165\u3055\u308c\u307e\u3057\u305f\u3002<\/p>\n<p><strong>Amgen<\/strong>\u3001<strong>Moderna<\/strong>\u3001<strong>Allen Institute<\/strong>\u3001<strong>Thermo Fisher Scientific<\/strong>\u306a\u3069\u306e\u9867\u5ba2\u3068\u5354\u529b\u3057\u3001\u7814\u7a76\u3068\u767a\u898b\u3092\u52a0\u901f\u3059\u308b\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u306b<strong>GPT-Rosalind<\/strong>\u3092\u9069\u7528\u3057\u3066\u3044\u307e\u3059\u3002\u30e2\u30c7\u30eb\u306f\u3001DNA\u306e\u69cb\u9020\u89e3\u660e\u306b\u8ca2\u732e\u3057\u305f<strong>Rosalind Franklin<\/strong>\u306b\u3061\u306a\u3093\u3067\u540d\u4ed8\u3051\u3089\u308c\u307e\u3057\u305f\u3002<\/p>\n<p>\u8a55\u4fa1\u3067\u306f\u3001\u5316\u5b66\u3001\u751f\u5316\u5b66\u3001\u30bf\u30f3\u30d1\u30af\u8cea\u7406\u89e3\u3001\u7cfb\u7d71\u767a\u751f\u5b66\u3001\u5b9f\u9a13\u8a08\u753b\u3001\u30c4\u30fc\u30eb\u5229\u7528\u306a\u3069\u306e\u79d1\u5b66\u7684\u30b5\u30d6\u30c9\u30e1\u30a4\u30f3\u306b\u304a\u3044\u3066\u3001<strong>GPT-Rosalind<\/strong>\u304c\u512a\u308c\u305f\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u793a\u3057\u307e\u3057\u305f\u3002\u7279\u306b\u3001<strong>BixBench<\/strong>\u3084<strong>LABBench2<\/strong>\u306a\u3069\u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3067\u3001\u4ed6\u306e\u30e2\u30c7\u30eb\u3092\u4e0a\u56de\u308b\u7d50\u679c\u3092\u51fa\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u30e2\u30c7\u30eb\u306f\u3001\u751f\u7269\u5b66\u7684\u8aa4\u7528\u306b\u5bfe\u3059\u308b\u5f37\u529b\u306a\u5b89\u5168\u7b56\u3092\u7dad\u6301\u3057\u3064\u3064\u3001\u4eba\u985e\u306e\u5065\u5eb7\u3092\u9032\u6b69\u3055\u305b\u308b\u305f\u3081\u306e\u79d1\u5b66\u8005\u3084\u7814\u7a76\u6a5f\u95a2\u306b\u63d0\u4f9b\u3055\u308c\u307e\u3059\u3002\u30a2\u30af\u30bb\u30b9\u306f\u3001\u6709\u76ca\u306a\u5229\u7528\u3001\u5f37\u529b\u306a\u30ac\u30d0\u30ca\u30f3\u30b9\u3068\u5b89\u5168\u76e3\u8996\u3001\u30a8\u30f3\u30bf\u30fc\u30d7\u30e9\u30a4\u30ba\u30b0\u30ec\u30fc\u30c9\u306e\u30bb\u30ad\u30e5\u30ea\u30c6\u30a3\u3092\u5099\u3048\u305f\u5236\u5fa1\u3055\u308c\u305f\u30a2\u30af\u30bb\u30b9\u3068\u3044\u30463\u3064\u306e\u539f\u5247\u306b\u57fa\u3065\u3044\u3066\u8a55\u4fa1\u3055\u308c\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/openai.com\/index\/introducing-gpt-rosalind\/\" target=\"_blank\" rel=\"noopener\">OpenAI \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> GPT-Rosalind is OpenAI's specialized AI model for life sciences, designed to accelerate drug discovery and biological research through advanced reasoning, tool integration, and enhanced understanding of complex scientific data, available via a controlled access program.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Editorial Staff<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p>OpenAI has introduced <strong>GPT-Rosalind<\/strong>, a frontier reasoning model designed to accelerate research across biology, drug discovery, and translational medicine. This model is optimized for scientific workflows, combining improved tool use with a deeper understanding of chemistry, protein engineering, and genomics.<\/p>\n<p>Drug development typically takes 10 to 15 years, with complex research workflows often hindering progress. <strong>GPT-Rosalind<\/strong> aims to help researchers navigate the early stages of discovery faster by supporting evidence synthesis, hypothesis generation, and experimental planning in multi-step research tasks.<\/p>\n<p>The model is available as a research preview in <strong>ChatGPT<\/strong>, <strong>Codex<\/strong>, and the API for qualified customers through a trusted access program. OpenAI has also introduced a freely accessible <strong>Life Sciences research plugin for Codex<\/strong>, enabling scientists to connect models to over 50 scientific tools and data sources.<\/p>\n<p>OpenAI is collaborating with customers such as <strong>Amgen<\/strong>, <strong>Moderna<\/strong>, the <strong>Allen Institute<\/strong>, and <strong>Thermo Fisher Scientific<\/strong> to apply <strong>GPT-Rosalind<\/strong> across workflows that accelerate research and discovery. The model is named after <strong>Rosalind Franklin<\/strong>, whose research was crucial in revealing the structure of DNA.<\/p>\n<p>Evaluations show that <strong>GPT-Rosalind<\/strong> delivers superior performance in scientific subdomains including chemistry, biochemistry, protein understanding, phylogenetics, experimental design, and tool usage. It has achieved leading results on benchmarks like <strong>BixBench<\/strong> and <strong>LABBench2<\/strong>, outperforming other models.<\/p>\n<p>Access to the model is provided to scientists and research organizations best positioned to advance human health, while maintaining strong safeguards against biological misuse. Access is evaluated based on three core principles: beneficial use, strong governance and safety oversight, and controlled access with enterprise-grade security.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">\u6025\u6210\u9577\u4e2d\u306e\u30ed\u30dc\u30c6\u30a3\u30af\u30b9\u30b9\u30bf\u30fc\u30c8\u30a2\u30c3\u30d7Physical Intelligence\u3001\u672a\u5b66\u7fd2\u30bf\u30b9\u30af\u3092\u5b9f\u884c\u53ef\u80fd\u306a\u65b0\u30ed\u30dc\u30c3\u30c8\u8133\u3092\u767a\u8868<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> Physical Intelligence, a hot robotics startup, says its new robot brain can figure out tasks it was never taught<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p>\u6025\u6210\u9577\u4e2d\u306e\u30ed\u30dc\u30c6\u30a3\u30af\u30b9\u30b9\u30bf\u30fc\u30c8\u30a2\u30c3\u30d7<strong>Physical Intelligence<\/strong>\u306f\u3001\u672a\u5b66\u7fd2\u306e\u30bf\u30b9\u30af\u3092\u81ea\u5f8b\u7684\u306b\u5b9f\u884c\u3067\u304d\u308b\u65b0\u3057\u3044\u30ed\u30dc\u30c3\u30c8\u8133\u3092\u767a\u8868\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u6280\u8853\u306f\u3001\u30ed\u30dc\u30c3\u30c8\u304c\u4e8b\u524d\u306b\u30d7\u30ed\u30b0\u30e9\u30e0\u3055\u308c\u3066\u3044\u306a\u3044\u72b6\u6cc1\u3067\u3082\u3001\u74b0\u5883\u3092\u7406\u89e3\u3057\u3001\u65b0\u3057\u3044\u554f\u984c\u3092\u89e3\u6c7a\u3059\u308b\u80fd\u529b\u3092\u5927\u5e45\u306b\u5411\u4e0a\u3055\u305b\u308b\u3068\u671f\u5f85\u3055\u308c\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u540c\u793e\u306e\u30ed\u30dc\u30c3\u30c8\u8133\u306f\u3001<strong>\u751f\u6210AI<\/strong>\u3068<strong>AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8<\/strong>\u306e\u539f\u5247\u3092\u6d3b\u7528\u3057\u3001\u8907\u96d1\u306a\u7269\u7406\u4e16\u754c\u3067\u306e\u63a8\u8ad6\u3068\u884c\u52d5\u8a08\u753b\u3092\u53ef\u80fd\u306b\u3057\u307e\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u3001\u88fd\u9020\u3001\u7269\u6d41\u3001\u30b5\u30fc\u30d3\u30b9\u696d\u306a\u3069\u3001\u591a\u69d8\u306a\u7523\u696d\u3067\u306e\u30ed\u30dc\u30c3\u30c8\u306e\u5c0e\u5165\u304c\u52a0\u901f\u3055\u308c\u308b\u53ef\u80fd\u6027\u304c\u3042\u308a\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u30d6\u30ec\u30fc\u30af\u30b9\u30eb\u30fc\u306f\u3001\u30ed\u30dc\u30c3\u30c8\u304c\u3088\u308a\u6c4e\u7528\u6027\u3092\u6301\u3061\u3001\u4eba\u9593\u304c\u4ecb\u5165\u3059\u308b\u3053\u3068\u306a\u304f\u3001\u4e88\u671f\u305b\u306c\u72b6\u6cc1\u306b\u9069\u5fdc\u3067\u304d\u308b\u3088\u3046\u306b\u306a\u308b\u3053\u3068\u3092\u610f\u5473\u3057\u307e\u3059\u3002\u3053\u308c\u306f\u3001<strong>\u6c4e\u7528AI\uff08AGI\uff09<\/strong>\u3078\u306e\u4e00\u6b69\u3068\u3082\u898b\u306a\u3055\u308c\u3066\u304a\u308a\u3001\u30ed\u30dc\u30c6\u30a3\u30af\u30b9\u5206\u91ce\u306b\u304a\u3051\u308b\u91cd\u8981\u306a\u9032\u5c55\u3067\u3059\u3002<\/p>\n<p><strong>Physical Intelligence<\/strong>\u306f\u3001\u3053\u306e\u65b0\u3057\u3044\u30ed\u30dc\u30c3\u30c8\u8133\u304c\u3001\u5f93\u6765\u306e\u30ed\u30dc\u30c3\u30c8\u304c\u76f4\u9762\u3057\u3066\u3044\u305f\u67d4\u8edf\u6027\u306e\u6b20\u5982\u3068\u3044\u3046\u8ab2\u984c\u3092\u514b\u670d\u3057\u3001\u3088\u308a\u8907\u96d1\u3067\u52d5\u7684\u306a\u74b0\u5883\u3067\u306e\u4f5c\u696d\u3092\u53ef\u80fd\u306b\u3059\u308b\u3068\u5f37\u8abf\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/techcrunch.com\/2026\/04\/16\/physical-intelligence-a-hot-robotics-startup-says-its-new-robot-brain-can-figure-out-tasks-it-was-never-taught\/\" target=\"_blank\" rel=\"noopener\">TechCrunch \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> Physical Intelligence's new robot brain, powered by Generative AI and AI Agents, enables robots to autonomously perform unlearned tasks, significantly increasing their versatility and adaptability in complex physical environments, marking a step towards AGI in robotics.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Connie Loizos<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>Physical Intelligence<\/strong>, a rapidly growing robotics startup, has unveiled a new robot brain capable of autonomously performing tasks it was never explicitly taught. This technology is expected to significantly enhance robots' ability to understand their environment and solve novel problems, even in unprogrammed situations.<\/p>\n<p>The company's robot brain leverages principles of <strong>Generative AI<\/strong> and <strong>AI Agents<\/strong> to enable reasoning and action planning in complex physical environments. This could accelerate the adoption of robots across various industries, including manufacturing, logistics, and service sectors.<\/p>\n<p>This breakthrough signifies that robots can become more versatile and adapt to unforeseen circumstances without human intervention. It is considered a significant step forward in the field of robotics, potentially moving closer to <strong>Artificial General Intelligence (AGI)<\/strong>.<\/p>\n<p><strong>Physical Intelligence<\/strong> emphasizes that this new robot brain overcomes the limitations of flexibility faced by traditional robots, allowing them to operate in more complex and dynamic settings.<\/p>\n<\/blockquote>\n<\/div>\n<div class=\"wp-block-group\" style=\"margin-top:40px;margin-bottom:40px\">\n<h2 class=\"wp-block-heading\">OpenAI\u3001Anthropic\u306b\u5bfe\u6297\u3057\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u5236\u5fa1\u3092\u5f37\u5316\u3057\u305fCodex\u3092\u767a\u8868<\/h2>\n<ul>\n<li><strong>\u539f\u984c:<\/strong> OpenAI takes aim at Anthropic with beefed-up Codex that gives it more power over your desktop | TechCrunch<\/li>\n<\/ul>\n<h3 class=\"wp-block-heading\">\u5c02\u9580\u30a2\u30ca\u30ea\u30b9\u30c8\u306e\u5206\u6790<\/h3>\n<div class=\"ai-summary-content\">\n<p><strong>OpenAI<\/strong>\u306f\u3001\u7af6\u5408\u3067\u3042\u308b<strong>Anthropic<\/strong>\u306b\u5bfe\u6297\u3059\u308b\u305f\u3081\u3001\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u74b0\u5883\u3067\u306e\u5236\u5fa1\u80fd\u529b\u3092\u5927\u5e45\u306b\u5f37\u5316\u3057\u305f<strong>Codex<\/strong>\u306e\u30a2\u30c3\u30d7\u30c7\u30fc\u30c8\u3092\u767a\u8868\u3057\u307e\u3057\u305f\u3002\u3053\u306e\u5f37\u5316\u3055\u308c\u305f<strong>Codex<\/strong>\u306f\u3001\u30e6\u30fc\u30b6\u30fc\u306e\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u4e0a\u3067\u3088\u308a\u5e83\u7bc4\u306a\u30bf\u30b9\u30af\u3092\u81ea\u52d5\u5316\u3057\u3001\u5b9f\u884c\u3067\u304d\u308b\u3088\u3046\u306b\u306a\u308a\u307e\u3059\u3002<\/p>\n<p>\u65b0\u3057\u3044<strong>Codex<\/strong>\u306f\u3001<strong>AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8<\/strong>\u3068\u3057\u3066\u306e\u6a5f\u80fd\u3092\u62e1\u5f35\u3057\u3001\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u9593\u306e\u9023\u643a\u3001\u30d5\u30a1\u30a4\u30eb\u64cd\u4f5c\u3001\u30b7\u30b9\u30c6\u30e0\u8a2d\u5b9a\u306e\u5909\u66f4\u306a\u3069\u3001\u8907\u96d1\u306a\u30de\u30eb\u30c1\u30b9\u30c6\u30c3\u30d7\u306e\u64cd\u4f5c\u3092\u81ea\u7136\u8a00\u8a9e\u306e\u6307\u793a\u306b\u57fa\u3065\u3044\u3066\u5b9f\u884c\u53ef\u80fd\u3067\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u3001\u958b\u767a\u8005\u3084\u4e00\u822c\u30e6\u30fc\u30b6\u30fc\u306e\u751f\u7523\u6027\u304c\u98db\u8e8d\u7684\u306b\u5411\u4e0a\u3059\u308b\u3053\u3068\u304c\u671f\u5f85\u3055\u308c\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u52d5\u304d\u306f\u3001<strong>OpenAI<\/strong>\u304c<strong>Anthropic<\/strong>\u306e<strong>Claude<\/strong>\u306a\u3069\u306e\u7af6\u5408\u30e2\u30c7\u30eb\u304c\u63d0\u4f9b\u3059\u308b\u9ad8\u5ea6\u306a\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u6a5f\u80fd\u306b\u76f4\u63a5\u5bfe\u6297\u3057\u3001\u30c7\u30b9\u30af\u30c8\u30c3\u30d7AI\u30a2\u30b7\u30b9\u30bf\u30f3\u30c8\u5e02\u5834\u3067\u306e\u4e3b\u5c0e\u6a29\u3092\u78ba\u7acb\u3057\u3088\u3046\u3068\u3059\u308b\u6226\u7565\u3092\u793a\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u5f37\u5316\u3055\u308c\u305f<strong>Codex<\/strong>\u306f\u3001\u30d7\u30ed\u30b0\u30e9\u30df\u30f3\u30b0\u30bf\u30b9\u30af\u3060\u3051\u3067\u306a\u304f\u3001\u65e5\u5e38\u7684\u306a\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u4f5c\u696d\u306b\u304a\u3044\u3066\u3082\u3001\u3088\u308a\u5f37\u529b\u306a<strong>LLM<\/strong>\u30d9\u30fc\u30b9\u306e\u81ea\u52d5\u5316\u3092\u63d0\u4f9b\u3057\u3001\u30e6\u30fc\u30b6\u30fc\u304c\u3088\u308a\u8907\u96d1\u306a\u30ef\u30fc\u30af\u30d5\u30ed\u30fc\u3092\u52b9\u7387\u7684\u306b\u7ba1\u7406\u3067\u304d\u308b\u3088\u3046\u652f\u63f4\u3057\u307e\u3059\u3002<\/p>\n<p>\ud83d\udc49 <strong><a href=\"https:\/\/techcrunch.com\/2026\/04\/16\/openai-takes-aim-at-anthropic-with-beefed-up-codex-that-gives-it-more-power-over-your-desktop\/\" target=\"_blank\" rel=\"noopener\">TechCrunch \u3067\u8a18\u4e8b\u5168\u6587\u3092\u8aad\u3080<\/a><\/strong><\/p>\n<\/div>\n<ul>\n<li><strong>\u8981\u70b9:<\/strong> OpenAI's updated Codex, with enhanced desktop control and AI Agent capabilities, directly challenges Anthropic by enabling more powerful, LLM-driven automation of complex multi-step tasks on user desktops, aiming to boost productivity and secure market leadership.<\/li>\n<li><strong>\u8457\u8005:<\/strong> Lucas Ropek<\/li>\n<\/ul>\n<blockquote class=\"wp-block-quote\"><p><span>English Summary:<\/span><\/p>\n<p><strong>OpenAI<\/strong> has announced a significant update to <strong>Codex<\/strong>, enhancing its control capabilities over desktop environments, a move aimed at competing with rival <strong>Anthropic<\/strong>. This beefed-up <strong>Codex<\/strong> will be able to automate and execute a wider range of tasks directly on a user's desktop.<\/p>\n<p>The new <strong>Codex<\/strong> expands its functionality as an <strong>AI Agent<\/strong>, capable of performing complex, multi-step operations based on natural language instructions, such as inter-application coordination, file manipulation, and system setting changes. This is expected to dramatically boost productivity for both developers and general users.<\/p>\n<p>This strategic move indicates <strong>OpenAI<\/strong>'s direct competition with advanced agent capabilities offered by models like <strong>Anthropic's Claude<\/strong>, aiming to establish leadership in the desktop AI assistant market.<\/p>\n<p>The enhanced <strong>Codex<\/strong> will provide more powerful <strong>LLM<\/strong>-based automation not only for programming tasks but also for everyday desktop operations, helping users efficiently manage more complex workflows.<\/p>\n<\/blockquote>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>OpenAI\u306eGPT-Rosalind\u306b\u3088\u308b\u30e9\u30a4\u30d5\u30b5\u30a4\u30a8\u30f3\u30b9\u7814\u7a76\u52a0\u901f\u3001Physical Intelligence\u306e\u672a\u5b66\u7fd2\u30bf\u30b9\u30af\u5b9f\u884c\u30ed\u30dc\u30c3\u30c8\u8133\u3001\u305d\u3057\u3066\u30c7\u30b9\u30af\u30c8\u30c3\u30d7\u5236\u5fa1\u3092\u5f37\u5316\u3057\u305fOpenAI Codex\u306e\u6700\u65b0\u52d5\u5411\u3092\u6df1\u6398\u308a\u3002<\/p>\n","protected":false},"author":1,"featured_media":6,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"vkexunit_cta_each_option":"","footnotes":""},"categories":[3],"tags":[8,17,348,349,16,57,350,347,51,186,26,15],"class_list":{"0":"post-1002","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","6":"hentry","7":"category-columns","8":"tag-ai","10":"tag-codex","11":"tag-gpt-rosalind","12":"tag-llm","13":"tag-openai","14":"tag-physical-intelligence","15":"tag-347","16":"tag-51","17":"tag-186","18":"tag-26"},"_links":{"self":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/1002","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/comments?post=1002"}],"version-history":[{"count":0,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/posts\/1002\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media\/6"}],"wp:attachment":[{"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/media?parent=1002"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/categories?post=1002"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/itexplore.org\/jp\/wp-json\/wp\/v2\/tags?post=1002"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}