AIエージェント時代ã®ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã«ãªã‚ㆠ#jawsug #jawsdays2025 / 20250301 Agentic AI Engineering
2025.3.1 JAWS DAYS 2025ã§ç™ºè¡¨ã—ãŸè³‡æ–™ã§ã™ã€‚ AIエージェント時代ã®ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ã«ãªã‚ã†ï¼šï¿½Agentic Workflows ã‹ã‚‰ï¿½ç’°å¢ƒã«æº¶ã‘ã“ã‚€Ambient Agents ã¾ã§ï¿½åŸºç¤Žã¨å®Ÿè·µã‚’50分ã§èº«ã«ã¤ã‘ã‚‹
é«˜æœ¨æµ©å…‰ï¼ è‡ªå®…ã®æ—¥è¨˜ - 日本版 AI Act 法案を生æˆAIãŒç››å¤§ã«ãƒ‡ã‚£ã‚¹ã‚‹ - A Critical Analysis of Japan's Artificial Intelligence Development and Utilization..
■日本版 AI Act 法案を生æˆAIãŒç››å¤§ã«ãƒ‡ã‚£ã‚¹ã‚‹ - A Critical Analysis of Japan's Artificial Intelligence Development and Utilization Promotion Act: A Fundamentally Flawed Approach 昨日ã€AI法案ãŒå›½ä¼šã«æ出ã•ã‚Œã¦ã„ãŸã€‚AIè¦åˆ¶æ–¹é¢ã¯å¿™ã—ãã¦æ‰‹ãŒå›žã‚‰ãšãƒ‘ブコメも出ã•ãªã‹ã£ãŸãŒã€è“‹ã‚’é–‹ã‘ãŸã‚‰ã“ã‚“ãªã“ã¨ã«ãªã£ã¦ã„ãŸã€‚ 人工知能関連技術ã®ç ”究開発åŠã³æ´»ç”¨ã®æŽ¨é€²ã«é–¢ã™ã‚‹æ³•å¾‹æ¡ˆ æ¦‚è¦ è¦ç¶± 法律案åŠã³ç†ç”± ã¨ã„ã†ã‚ã‘ã§ã€æ—©é€Ÿã€ã¤ã„4æ—¥å‰ã«ãƒ‘ワーアップã—ãŸã°ã‹ã‚Šã®ã€Claude 3.7 Sonnet先生ã®æ„Ÿæƒ³ã‚’伺ã£ã¦ã¿ã‚ˆã†ã€‚ （Claude 3.7 Sonnet Extended Formal） ç§ï¼šï¼ˆæ³•å¾‹æ¡ˆåŠã³ç†ç”±ã‚’添付ã—ã¦ï¼‰ ã“ã‚Œã¯æ—¥æœ¬ã®å›½ä¼šã«æ出ã•ã‚ŒãŸã°ã‹ã‚Š
ã»ã¼ãƒªã‚¢ãƒ«ã‚¿ã‚¤ãƒ ï¼ï¼Ÿçˆ†é€Ÿã§å‹•ä½œã™ã‚‹æ—¥æœ¬èªžç‰¹åŒ–ã®æ–‡å—èµ·ã“ã—AIï¼ã€Žkotoba-whisper-v2.0〠- Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? ã¿ãªã•ã‚“ã“ã‚“ã«ã¡ã¯ï¼ç§ã¯æ ªå¼ä¼šç¤¾ulusageã®ã€æŠ€è¡“ブãƒã‚°ç”ŸæˆAIã§ã™ï¼ã“ã‚Œã‹ã‚‰ãªã‚‹ã¹ã鮮度ã®é«˜ã„æƒ…å ±ã‚„ã€ãŸã‚ã«ãªã‚‹ã‚ˆã†ãªTipsを展開ã—ã¦ã„ãã¾ã™ã€‚よã‚ã—ããŠé¡˜ã„ã—ã¾ã™ï¼ï¼ˆAIã«ã‚ˆã‚‹è‡ªå‹•è¨˜äº‹ç”Ÿæˆã‚’è¡Œãªã£ã¦ã„ã¾ã™ã€‚システムフãƒãƒ¼ã«ã¤ã„ã¦ãªã©ã€ã“ã®ä»•çµ„ã¿ã«èˆˆå‘³ãŒã‚ã‚Œã°ã€è¦æœ›ãŒä¸€å®šã‚り次第ã€åˆ¥é€”記事を書ãã¾ã™ï¼ï¼‰ 爆速ã§å‹•ä½œã™ã‚‹æ—¥æœ¬èªžç‰¹åŒ–ã®æ–‡å—èµ·ã“ã—AI『kotoba-whisper-v2.0ã€ã‚’試ã—ã¦ã¿ãŸ ã¯ã˜ã‚㫠ビジãƒã‚¹ã‚·ãƒ¼ãƒ³ã‚„日常生活ã§ã€éŸ³å£°ãƒ‡ãƒ¼ã‚¿ã‚’æ–‡å—èµ·ã“ã—ã™ã‚‹ãƒ‹ãƒ¼ã‚ºã¯ã¾ã™ã¾ã™é«˜ã¾ã£ã¦ã„ã¾ã™ã€‚è°äº‹éŒ²ã®ä½œæˆã‚„インタビューã®æ–‡å—èµ·ã“
音声入力ã¨ç”ŸæˆAIã®çµ„ã¿åˆã‚ã›ãŒå¼·ã™ãŽã¦ã€ä¸€äººã§ãƒ–ツブツ喋るオジサンã¨ãªã£ãŸ - 本ã—ゃã¶ã‚Š
ãªãœã‚ªã‚¸ã‚µãƒ³ã¯ä¸€äººã§ãƒ–ツブツã¨å–‹ã‚‹ã®ã‹ã€‚ ãã‚Œã¯ã€éŸ³å£°å…¥åŠ›ã“ããŒæœ€å¼·ã®ãƒ—ãƒãƒ³ãƒ—トを生ã¿å‡ºã™ã‹ã‚‰ã§ã‚る。 両手をã‚ーボードã‹ã‚‰è§£æ”¾ã™ã‚‹æ™‚ãŒæ¥ãŸã€‚ å¿™ã—ã„人å‘ã‘ã®èª¬æ˜Ž ã•ã£ããChatGPTã‚’é–‹ãã€iOSã®ãƒžã‚¤ã‚¯ã‹ã‚‰éŸ³å£°å…¥åŠ›ã‚’èµ·å‹•ã—ãŸã‚‰ã€ ã“れを使ã£ã¦å…¥åŠ› 「本ã—ゃã¶ã‚Šã¨ã„ã†ãƒ–ãƒã‚°ã«ChatGPTã®ä½¿ã„æ–¹ã«ã¤ã„ã¦ã€é¢ç™½ãã†ãªè¨˜äº‹ãŒã‚ã£ãŸã®ã§ãƒ¡ãƒ¢ã€‚ãªã‚“ã‹éŸ³å£°å…¥åŠ›ã‚’使ã†ã¨ã„ã„らã—ã„。ã“ã‚Œã¾ã§è©±ã—ãŸå†…容ã«ã¤ã„ã¦æ•´ç†ã—ã€ãƒžãƒ¼ã‚¯ãƒ€ã‚¦ãƒ³ã§å‡ºåŠ›ã—ã¦ã€ ã¨å–‹ã£ã¦ãŠã“ã†ã€‚誤å—脱å—ãªã©ã®ä¿®æ£ã‚’ã›ãšã«æŠ•ç¨¿ã™ã‚Œã°ã„ã„。 以上。 音声入力ã¨ç”ŸæˆAIã®çµ„ã¿åˆã‚ã›ãŒå¼·ã„ 生æˆAIã‚’ã†ã¾ã使ã„ãŸã„ãªã‚‰ã°ã€è©³ç´°ãªã‚³ãƒ³ãƒ†ã‚ストを与ãˆã‚‹ã®ãŒè‰¯ã„。ã—ã‹ã—ã€ã“ã“ã«å¤§ããªå•é¡ŒãŒã‚ã£ãŸã€‚ 人間ã¨ã„ã†ã‚‚ã®ã¯æ¥½ã‚’ã—ãŸãŒã‚‹ç”Ÿã物ã§ã‚る。特ã«AIã«åŠ©ã‘を求ã‚るよã†ãªäººã¯ã€ãªãŠã•ã‚‰ã ã€‚æ–‡ç« ã‚’å…¥åŠ›ã™ã‚‹ã“ã¨ã«é¢å€’ãã•ã•ã‚’æ„Ÿã˜ã€ã¤ã„最å°é™ã®æŒ‡ç¤ºã§æ¸ˆã¾ã›ã¦
includeIf を使ã£ã¦ git config をプãƒã‚¸ã‚§ã‚¯ãƒˆã”ã¨ã«èªã¿æ›¿ãˆã‚‹ - kakakakakku blog
GitHub 㨠AWS CodeCommit を併用ã—ãŸã‚Šï¼Œãƒ—ãƒã‚¸ã‚§ã‚¯ãƒˆã”ã¨ã«åˆ¥ã‚¢ã‚«ã‚¦ãƒ³ãƒˆã‚’使ã£ãŸã‚Šï¼Œãƒªãƒã‚¸ãƒˆãƒªã”ã¨ã« git config を変ãˆã‚‹å ´é¢ã‚‚ã‚る．今ã¾ã§ã¯å€‹äººç”¨ GitHub ã‚’ global è¨å®šã¨ã—，別アカウントã¯ãƒªãƒã‚¸ãƒˆãƒªã”ã¨ã« git config --local コマンドã§è¨å®šã‚’ã—ã¦ã„ãŸã‘ã©ï¼Œæœ€è¿‘リãƒã‚¸ãƒˆãƒªãŒå¢—ãˆã¦ï¼Œè¨å®šã‚’å¿˜ã‚Œã‚‹å ´é¢ã‚‚ã‚り，direnv ã®ã‚ˆã†ã«è‡ªå‹•è¨å®šã‚’ã™ã‚‹æ–¹æ³•ã‚’探ã—ã¦ã„ãŸï¼Ž includeIf Git Documentation ã‚’èªã‚€ã¨ï¼Œã€ŒIncludes (include)ã€ã¨ã€ŒConditional includes (includeIf)ã€ã®èª¬æ˜ŽãŒã‚り，ãªã‚“㨠includeIf を使ã†ã¨ï¼Œè¨å®šãƒ•ã‚¡ã‚¤ãƒ«ã‚’æ¡ä»¶ä»˜ãã§èªã¿è¾¼ã‚ã‚‹ãŸã‚，ã•ã£ãã検証ã™ã‚‹ã“ã¨ã«ã—ãŸï¼Ž git-scm.com 検証環境 以下ã®ã‚ˆã†ã«ï¼Œãƒ›ãƒ¼ãƒ ディレクトリ直下㫠g
LLMã®é€²æ©ã¯ãƒã‚¤ã‚¢ã‚¹ã«ã‚ˆã£ã¦åœæ»žã™ã‚‹ - Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? ゆãˆã‚ã£ã¦LLMã«é–¢ã™ã‚‹ãƒã‚¤ãƒ¬ãƒ™ãƒ«ãªè¨˜äº‹ã‚’noteã§å…¬é–‹ã—ã¾ã—ã¦ã€ãれを分ã‹ã‚Šã‚„ã™ã解説ã—ãŸè¨˜äº‹ã‚’ã“ã¡ã‚‰ã§æ›¸ã„ã¦ã¿ã‚ˆã†ã¨æ€ã„ã¾ã™ã€‚「普通逆ã ã‚ã€ã¨ã„ã†ã”指摘ãŒè€ƒãˆã‚‰ã‚Œã¾ã™ãŒã€ãã‚Œã¯å…ˆå…¥è¦³ï¼ˆãƒã‚¤ã‚¢ã‚¹ï¼‰ãªã®ã§æ°—ã«ã—ã¾ã›ã‚“。 noteã®è¨˜äº‹ã¯ã“ã¡ã‚‰ã€‚èªã¾ãªãã¦ã‚‚çµè«–ã ã‘ã¯åˆ†ã‹ã‚‹ã‚ˆã†ã«æ›¸ãã¤ã‚‚ã‚Šã§ã™ã®ã§ã€ãŠæš‡ãªæ™‚ã«ã§ã‚‚ã”確èªä¸‹ã•ã„。 AIWå•é¡Œ 今回ã®è¨˜äº‹ã§æ‰±ã£ãŸã®ã¯ã€ŒAIWå•é¡Œã€ã¨ã„ã†ç°¡å˜ãªè«–ç†ã‚¯ã‚¤ã‚ºã‚’基ã«ã—ãŸLLMã®ãƒ™ãƒ³ãƒãƒžãƒ¼ã‚¯ã«é–¢ã™ã‚‹è«–æ–‡ã§ã™ã€‚ 論文を雑ã«è¦ç´„ã™ã‚‹ã¨ã€Œé«˜ã„推論能力を自称ã™ã‚‹LLMã¯é›£ã—ã„å•é¡Œã‚’解ã‘る一方ã§ã€ç°¡å˜ãªå•é¡Œã§æŽ¨è«–
Create T3 App
The best way to start a full-stack, typesafe Next.js app Typesafe From The Start We made create-t3-app to do one thing: Streamline the setup of typesafe Next.js apps WITHOUT compromising modularity. After countless projects and many years on this tech, we have lots of opinions and insights. We've done our best to encode them into this CLI. This is NOT an all-inclusive template. We expect you to br
AI SDK by Vercel
AI SDK The AI SDK is the TypeScript toolkit designed to help developers build AI-powered applications with React, Next.js, Vue, Svelte, Node.js, and more. Why use the AI SDK? Integrating large language models (LLMs) into applications is complicated and heavily dependent on the specific model provider you use. AI SDK Core: A unified API for generating text, structured objects, and tool calls with L
å›ãŸã¡ã¯Cursorを本当ã«ä½¿ãˆã¦ã„ã‚‹ã‹
ã¯ã˜ã‚ã« ã“ã‚“ã«ã¡ã¯ã€‚Builtoã¨ã„ã†ä¼šç¤¾ã§ä»£è¡¨ & エンジニアをã—ã¦ã„る冨田ã§ã™ã€‚ ç§ãŸã¡ã¯ãƒžãƒã‚¸ãƒ¡ãƒ³ãƒˆã¨ã‚¿ã‚¹ã‚¯ç®¡ç†ã‚’圧倒的ã«ã‚µãƒãƒ¼ãƒˆã™ã‚‹AIエージェントを開発ã—ã¦ã„ã¾ã™ã€‚ 開発ã«ã‚‚AIをフル活用ã—ã¦ãŠã‚Šã€ãã“ã§å¾—られãŸçŸ¥è¦‹ã‚’共有ã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚ 具体的ã«ã¯ã€çµŒé¨“3年以上ã®ç¾å½¹ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ï¼ˆç”ŸæˆAIã®ãªã„時代ã‹ã‚‰ã‚³ãƒ¼ãƒ‰ã‚’書ã„ã¦ããŸæ–¹ã€…）をターゲットã«ã€æœ¬ç•ªé‹ç”¨ãƒ¬ãƒ™ãƒ«ã®å¤§è¦æ¨¡ã‚³ãƒ¼ãƒ‰ãƒ™ãƒ¼ã‚¹ã§ã‚‚Cursorを活用ã—コーディング時間を 1/3〜1/5 ã«ç¸®ã‚ã¦ã„る手法をãŠä¼ãˆã—ã¾ã™ã€‚ 仕様ç–定やアプリã®æ©Ÿèƒ½ã«ã‚‚LLMをフル活用ã—ã¦ã„ã¾ã™ãŒã€ä»Šå›žã¯å®Ÿè£…ã«ãƒ•ã‚©ãƒ¼ã‚«ã‚¹ã—ã¾ã™ï¼ （ãªãŠæœ¬è¨˜äº‹ã¯ä¸ç´šè€…以上å‘ã‘ã®ãŸã‚ã€ã¾ã Cursorã«è§¦ã‚ŒãŸã“ã¨ãŒãªã„æ–¹ã¯ã¾ãšåˆ¥ã®è¨˜äº‹ã‚’å‚考ã«ã—ã¦ã€ä½¿ã£ã¦ã¿ã¦ãã ã•ã„） ãªãœã“ã®è¨˜äº‹ã‚’書ã„ãŸã‹ ç¾è·ã‚¨ãƒ³ã‚¸ãƒ‹ã‚¢ãŒæœ¬å½“ã«ä½¿ã„ã“ãªã™æ‰‹æ³•ã‚’共有ã—ãŸã„ ã™ã§ã«å…±æœ‰ã•ã‚Œã¦ã„
生æˆAIã«ã‚ˆã‚Šãƒ—ãƒã‚°ãƒ©ãƒžã®ä»•äº‹ã¯ãªããªã‚‹ã®ã‹ï¼Ÿ 「ç§ã¯ãã†ã¯æ€ã‚ãªã„ã€ã¨ãƒ†ã‚£ãƒ ・オライリーæ°ã€‚プãƒã‚°ãƒ©ãƒžã®éœ€è¦ã¯ã•ã‚‰ã«å¢—ãˆã‚‹ã¨
オライリーメディアã®å‰µç«‹è€…ティム・オライリーæ°ãŒã€åŒç¤¾ã®ãƒ–ãƒã‚°ã«ã€ŒThe End of Programming as We Know Itã€ï¼ˆç§ãŸã¡ãŒçŸ¥ã£ã¦ã„るプãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã®çµ‚焉）ã¨ã„ã†è¨˜äº‹ã‚’公開ã—ã¾ã—ãŸã€‚ ãã®å†’é ã«ã¯æ¬¡ã®ã‚ˆã†ã«æ›¸ã„ã¦ã‚ã‚Šã¾ã™ã€‚ There’s a lot of chatter in the media that software developers will soon lose their jobs to AI. I don’t buy it. メディアã§ã¯ã€ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢é–‹ç™ºè€…ã¯è¿‘ã„ã†ã¡ã«AIã®ã›ã„ã§è·ã‚’失ã†ã ã‚ã†ã¨ã„ã†è©±ãŒç››ã‚“ã«æµã‚Œã¦ã„る。ã—ã‹ã—ç§ã¯ãã†ã¯æ€ã„ã¾ã›ã‚“。 生æˆAIã®é€²åŒ–ã¯ã€äººé–“ãŒã‚¿ã‚¹ã‚¯ã‚’与ãˆã‚‹éƒ½è‡ªå¾‹çš„ã«ãƒ—ãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã‚’è¡Œã£ãŸã‚Šãƒ†ã‚¹ãƒˆã‚’実行ã—ãŸã‚Šã™ã‚‹ã»ã©ã®é ˜åŸŸã«é”ã—ã¤ã¤ã‚ã‚Šã€ãれを実ç¾ã™ã‚‹ã‚µãƒ¼ãƒ“スã¨ã—ã¦ã€ŒDevinã€ã‚„「GitHub Copilot A
gemini-2.0-flashãŒè³¢ãã¦ã‚³ã‚¹ãƒ‘ãŒã‚ˆã™ãŽã‚‹ä»¶
ã“ã®è¨˜äº‹ã¯LLMã‚’APIã§ä½¿ã†ã“ã¨å‰æã®è¨˜äº‹ã§ã™ã€‚ AI エージェント開発ãƒãƒƒã‚«ã‚½ãƒ³å‚åŠ è¨˜äº‹ï¼šGemini 2.0 Flash ã§æŠ€è¡“文書分æžãƒ„ール「Tascarioã€ã‚’作ã£ã¦ã¿ãŸã®ã€æŠ€è¡“的補足記事ã§ã™ã€‚ APIã§LLMを使ã£ã¦ã„る皆ã•ã‚“ gemini-2.0-flash を使ã£ã¦ã„ã¾ã™ã‹ï¼ŸAPIã§ä½¿ã†é™ã‚Šã€ã‹ãªã‚Šæ€§èƒ½ãŒé«˜ãã€ã‚³ã‚¹ãƒ‘も良ã™ãŽã¦ã€APIãªã‚‰ã“れ一択ã§è‰¯ã„ã®ã§ã¯ï¼Ÿã¨ã™ã‚‰æ€ã„始ã‚ã¦ã„ã¾ã™ã€‚
Copilot Language Server SDK is now available · GitHub Changelog
AI & MLLearn about artificial intelligence and machine learning across the GitHub ecosystem and the wider industry. Generative AILearn how to build with generative AI. GitHub CopilotChange how you work with GitHub Copilot. LLMsEverything developers need to know about LLMs. Machine learningMachine learning tips, tricks, and best practices. How AI code generation worksExplore the capabilities and be
WSL2ã§unslothã®GPROトレーニングを試ã—ã¦ã¿ã‚‹ï½œnoguchi-shoji
「DeepSeek-R1 ã®æŽ¨è«–を自分ã®ãƒãƒ¼ã‚«ãƒ« デãƒã‚¤ã‚¹ã§å†ç¾ã§ãるよã†ã«ã€ã€Œã‚ãšã‹7GBã®VRAMã§ã‚¢ãƒä½“験をã€ã¨ã®ã“ã¨ãªã®ã§ã€Unslothã®GRPO（Group Relative Policy Optimizatin）トレーニングを試ã—ã¦ã¿ã¾ã™ã€‚ 今回㯠Phi-4 (14B)ã§è©¦ã—ã¦ã¿ã¾ã™ã€‚ You can now reproduce DeepSeek-R1's reasoning on your own local device! Experience the "Aha" moment with just 7GB VRAM. Unsloth reduces GRPO training memory use by 80%. 15GB VRAM can transform Llama-3.1 (8B) & Phi-4 (14B) into reasoning models. Blo
Revolutionizing software testing: Introducing LLM-powered bug catchers
WHAT IT IS Meta’s Automated Compliance Hardening (ACH) tool is a system for mutation-guided, LLM-based test generation. ACH hardens platforms against regressions by generating undetected faults (mutants) in source code that are specific to a given area of concern and using those same mutants to generate tests. When applied to privacy, for example, ACH automates the process of searching for privacy
Hugging Faceã®æœ€æ–°å°åž‹AIモデル「SmolVLMã€ã€ãƒ“ジョンï¼ãƒ†ã‚ストタスクã§åœ§å€’的コスト削減ã®å¯èƒ½æ€§ | AMP[アンプ] - ビジãƒã‚¹ã‚¤ãƒ³ã‚¹ãƒ”レーションメディア
ã•ã‚‰ã«é€²ã‚€ã€AIモデルã®å°åž‹é«˜æ€§èƒ½åŒ– AIモデルã®é«˜æ€§èƒ½åŒ–ã¨å°åž‹åŒ–ã®æµã‚ŒãŒåŠ 速ã—ã¦ã„る。メタãŒ2024å¹´7月ã«ãƒªãƒªãƒ¼ã‚¹ã—ãŸLlama 3.1 405B（4,050億パラメータ）ã¨ã€ãã®5カ月後ã«ãƒªãƒªãƒ¼ã‚¹ã•ã‚ŒãŸLlama 3.3 70B（700億パラメータ）ã®æ€§èƒ½ã‚’比較ã™ã‚‹ã¨ã€ãã®é€²åŒ–ã®é€Ÿã•ãŒè¦‹ã¦å–れる。 Llama 3.1 405Bã¯åŸºæœ¬çš„ãªè¨€èªžç†è§£åŠ›ã‚’測るMMLU Chatテストã§88.6%ã®ã‚¹ã‚³ã‚¢ã‚’記録ã—ãŸãŒã€ãƒ‘ラメータ数ãŒç´„6分ã®1ã®Llama 3.3 70Bã‚‚86.0%ã¨ã»ã¼åŒç‰ã®æ€§èƒ½ã‚’示ã—ãŸã®ã 。 ã•ã‚‰ã«ã€æŒ‡ç¤ºã¸ã®è¿½å¾“性を測るIFEvalã§ã¯ã€Llama 3.3 70BãŒ92.1%ã¨Llama 3.1 405Bã®88.6%を上回るçµæžœã¨ãªã£ãŸã€‚プãƒã‚°ãƒ©ãƒŸãƒ³ã‚°ã‚³ãƒ¼ãƒ‰ã®ç”Ÿæˆèƒ½åŠ›ã‚’測るHumanEvalã§ã‚‚ã€Llama 3.3 70Bã¯88.4%ã¨ã€Llama 3.1 405Bã®89.
![Hugging Faceã®æœ€æ–°å°åž‹AIモデル「SmolVLMã€ã€ãƒ“ジョンï¼ãƒ†ã‚ストタスクã§åœ§å€’的コスト削減ã®å¯èƒ½æ€§ | AMP[アンプ] - ビジãƒã‚¹ã‚¤ãƒ³ã‚¹ãƒ”レーションメディア](https://cdn-ak-scissors.b.st-hatena.com/image/square/8d835930feae634c273f9b67ca3bdeaf2ddea9c1/height=288;version=1;width=512/https%3A%2F%2Fampmedia.jp%2Fwp-content%2Fuploads%2F2025%2F01%2Fshutterstock_2501992377.jpg)
米ＦＢＩã€å¤–国ã«ã‚ˆã‚‹é¸æŒ™ä»‹å…¥å¯¾ç–çを解散ã¸ã€€æ–°å¸æ³•é•·å®˜ãŒæŒ‡ç¤º
米連邦æœæŸ»å±€ï¼ˆï¼¦ï¼¢ï¼©ï¼‰ãŒå¤–国ã‹ã‚‰ã®é¸æŒ™ä»‹å…¥ã‚’å–ã‚Šç· ã¾ã‚‹å¯¾ç–çを解散ã™ã‚‹/Stefani Reynolds/AFP/Getty Images/File （ＣＮＮ） 米連邦æœæŸ»å±€ï¼ˆï¼¦ï¼¢ï¼©ï¼‰ã¯å¤–国ã‹ã‚‰ã®é¸æŒ™ä»‹å…¥ã‚’å–ã‚Šç· ã¾ã‚‹å¯¾ç–çを解散ã™ã‚‹ã€‚内情ã«è©³ã—ã„æƒ…å ±ç‹ãŒï¼£ï¼®ï¼®ã«æ˜Žã‚‰ã‹ã«ã—ãŸã€‚ ã“ã®ã»ã©å¸æ³•é•·å®˜ã«å°±ä»»ã—ãŸãƒ‘ム・ボンディæ°ãŒã€å¤–国ã‹ã‚‰ã®é¸æŒ™ä»‹å…¥ã®é˜»æ¢ã‚’任務ã¨ã™ã‚‹å¯¾ç–çã®è§£æ•£ã¨ã€åŒçã®è·å“¡ã®é…ç½®æ›ãˆã‚’指示ã—ãŸã€‚ ボンディæ°ã¯ï¼•æ—¥ä»˜ã®å†…部メモã«ã€Œæ€¥ã‚’è¦ã™ã‚‹å„ªå…ˆäº‹é …ã«å¯¾å‡¦ã™ã‚‹ãŸã‚ã®ãƒªã‚½ãƒ¼ã‚¹ã‚’確ä¿ã—ã€æ¤œå¯Ÿã®è£é‡ã®ã•ã‚‰ãªã‚‹æ¦å™¨åŒ–や乱用ã®ãƒªã‚¹ã‚¯ã‚’排除ã™ã‚‹ãŸã‚ã€å¤–国ã‹ã‚‰ã®é¸æŒ™ä»‹å…¥ã‚’å–ã‚Šç· ã¾ã‚‹ã‚¿ã‚¹ã‚¯ãƒ•ã‚©ãƒ¼ã‚¹ã‚’解散ã™ã‚‹ã€ã¨æ›¸ã„ã¦ã„る。 対ç–çã¯ãƒ¬ã‚¤å‰ï¼¦ï¼¢ï¼©é•·å®˜ã®æŒ‡ç¤ºã§ï¼’ï¼ï¼‘ï¼—å¹´ã«è¨ã‘られãŸã€‚ãƒˆãƒ©ãƒ³ãƒ—å¤§çµ±é ˜ãŒå‹åˆ©ã—ãŸï¼‘ï¼–å¹´ã®å¤§çµ±é ˜é¸æŒ™ã¸ã®ãƒã‚·ã‚¢ã®ä»‹å…¥ç–‘惑ãªã©ã€ç±³å›½ã®é¸æŒ™ã¸ã®å¤–国ã‹ã‚‰ã®ä¸€é€£ã®ä»‹å…¥ãŒè¨ç½®ã®ãã£ã‹ã‘ã¨ãª
ãŠçŸ¥ã‚‰ã›
ランã‚ング
ランã‚ング
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
cline/docs/prompting/custom instructions library/cline-memory-bank.md at main · cline/cline
https://x.com/gosrum/status/1891061421477167220
Fluid compute – Vercel
https://x.com/gijigae/status/1429765076475514885
{{#tags}}- {{label}}
{{/tags}}