探討開源模型生態系的複利效應：中國「開源優先」與高參與度 AI 生態的啟示★ 75
Interconnects (Nathan L.)33 days agoOpinionThis article delves into how the open-source AI model ecosystem achieves exponential growth through "compounding effects," using China's highly engaged…
在 AWS 上進行基礎模型訓練與推理的建構基石 (Building Blocks)★ 75
Hugging Face Blog34 days agoTutorialIn the era of generative AI, training and deploying foundation models with billions of parameters faces enormous computational and architectural challenges…
Superset 如何在 Vercel 上構建 AI Agent 專屬的 IDE
Vercel Changelog35 days agoCommentaryAs AI agents rise to prominence, traditional code editors can no longer meet developers' needs for debugging, observing, and orchestrating agents. Superset is…
來自中國頂尖 AI 實驗室的內部觀察筆記★ 75
Interconnects (Nathan L.)38 days agoCommentaryThis in-depth piece from Interconnects founder Nathan Lambert documents his key observations after personally visiting several of China's top AI laboratories —…
vLLM V0 到 V1 的演進：在強化學習（RL）中「正確性重於修正」的實踐★ 75
Hugging Face Blog39 days agoOpinionThis blog post published by the ServiceNow AI team delves into the major transition of the open-source large language model inference engine vLLM from V0 to…
Hugging Face 為 Open ASR 排行榜引入「防刷榜機制」，使用私有測試數據打擊 Benchmaxxer★ 75
Hugging Face Blog40 days agoReleaseHugging Face has recently made a major update to its popular Open ASR (Automatic Speech Recognition) leaderboard, aimed at combating the increasingly serious…
蒸餾恐慌：為什麼將「知識蒸餾」稱為安全攻擊是極其糟糕的趨勢★ 75
Interconnects (Nathan L.)41 days agoOpinionIn the field of machine learning, "knowledge distillation" is a well-established technique that generally refers to using the output data generated by a…
IBM 推出 Granite 4.1 開源大語言模型：架構設計與構建技術深度解析★ 75
Hugging Face Blog46 days agoReleaseIBM has published a detailed blog post on Hugging Face outlining the construction technology and architectural design behind its latest generation of…
DeepInfra 正式加入 Hugging Face 推理服務商（Inference Providers）陣容 🔥★ 72
Hugging Face Blog47 days agoReleaseHugging Face's official blog has announced that DeepInfra — a well-known high-performance, low-cost serverless inference platform — has officially joined…
NVIDIA 推出 Nemotron 3 Nano Omni：支援長文本的多模態智慧模型，專為文件、語音與影片 Agent 設計★ 75
Hugging Face Blog47 days agoReleaseNVIDIA has officially launched a new lightweight multimodal model, "Nemotron 3 Nano Omni." This model is designed to deliver powerful multimodal intelligence…
2026 Vercel AI Accelerator 成果回顧：引領下一代 AI 應用與 Agent 開發浪潮
Vercel Changelog47 days agoBusinessVercel recently published a comprehensive retrospective on its 2026 AI Accelerator program. The program is designed to provide technical support, funding, and…
DeepSeek-V4：Agent 真正能派上用場的百萬 Token 超長上下文★ 85
Hugging Face Blog52 days agoReleaseDeepSeek has introduced its next-generation open-source model, DeepSeek-V4, whose most attention-grabbing feature is an ultra-long context window of up to 1…
如何在 Chrome 擴充功能中使用 Transformers.js 運行本地 AI 模型★ 75
Hugging Face Blog53 days agoTutorialAs browser-side computing power continues to improve, deploying AI models directly on the user's local device has become a popular trend. Hugging Face has…
QIMMA ⛰：首個品質優先的阿拉伯語大型語言模型（LLM）排行榜
Hugging Face Blog54 days agoReleaseThe Technology Innovation Institute (TII) of the United Arab Emirates — the organization behind the well-known open-source model Falcon — has officially…
AI 與網路安全的未來：為什麼「開放」至關重要★ 75
Hugging Face Blog55 days agoOpinionAs artificial intelligence (AI) technology undergoes explosive growth, cybersecurity has become a focal point of concern for governments and enterprises…
解讀當前開源與閉源 AI 模型的性能差距：超越單一評估指標的迷思★ 75
Interconnects (Nathan L.)55 days agoOpinionIn today's AI landscape, the performance gap between open-weights models (such as Meta's Llama family) and closed-source models (such as OpenAI's GPT and…
Import AI 454：自動化對齊研究、中國 AI 模型安全評估與全新 4 位元浮點格式 HiFloat4★ 75
Import AI (Jack Clark)55 days agoCommentaryIn this issue of Import AI 454, written by Jack Clark, the author begins by posing a thought-provoking question about finance and sociology: "At what point…
Hugging Face 推出 transformers-to-mlx：讓 Apple Silicon 運行 AI 模型更簡單的重大整合★ 80
Hugging Face Blog60 days agoReleaseThis article from the official Hugging Face blog, titled "The PR you would have opened yourself," focuses on the introduction of a brand-new technical…
使用 Sentence Transformers 訓練與微調多模態嵌入與 Reranker 模型★ 80
Hugging Face Blog60 days agoTutorialAs multimodal AI has become widespread, integrating data from different modalities — text, images, and more — into a single vector space and performing…
預測 2026 年年中：我對開源 AI 模型的幾點賭注與開閉源差距分析★ 75
Interconnects (Nathan L.)60 days agoOpinionIn this forward-looking article on the state of AI in mid-2026, Interconnects founder Nathan Lambert takes a deep dive into the dynamic gap between open-weight…
如何使用 Seedance 2.0 製作出色的 AI 影片★ 70
Replicate Blog61 days agoTutorial### A New Era of AI Video Generation: Why Now Is the Best Time With the rapid evolution of generative AI technology, video generation has transformed from an…
Nathan Lambert 的最新進展：ATOM Report、Post-Training 課程、新書與持續進行的 AI 研究★ 70
Interconnects (Nathan L.)61 days agoReleaseNathan Lambert, a prominent AI expert, former Alignment Scientist at Hugging Face, and founder of the popular newsletter Interconnects, recently wrote about…
【Nathan Lambert 觀點】建立「開源模型聯盟」的必然需求與挑戰★ 75
Interconnects (Nathan L.)64 days agoOpinionIn this article published in Interconnects, veteran AI expert Nathan Lambert explores the structural crisis facing the open-source AI ecosystem and proposes a…
Claude 神話與對開源權重模型無謂的恐慌★ 75
Interconnects (Nathan L.)66 days agoOpinionIn this opinion piece published in Interconnects, prominent AI policy and technology critic Nathan Lambert delivers a sharp critique of the excessive panic…
Sentence Transformers 推出多模態嵌入與重排（Reranker）模型支援★ 78
Hugging Face Blog67 days agoReleaseThe popular open-source library `sentence-transformers` from Hugging Face has received a major update, officially introducing native support for Multimodal…
Waypoint-1.5：讓家用 GPU 也能運行高保真度互動式虛擬世界★ 75
Hugging Face Blog67 days agoReleaseAs artificial intelligence advances toward Embodied AI and real-world physical interaction, high-fidelity 3D simulation environments have long been an…
Gemma 4 與開源模型成功的關鍵：為什麼基準測試分數不再是唯一指標★ 75
Interconnects (Nathan L.)72 days agoCommentaryThis article takes a deep dive into the release of Google's latest open-source model Gemma 4, using it as an opportunity to re-examine the core factors that…
Gemma 4：同等參數規模下最強大的開源模型，專為進階推理與 Agent 工作流打造★ 85
Google DeepMind Blog73 days agoReleaseGoogle DeepMind has today officially released its latest generation of open-source model series — Gemma 4. The company positions it as "the smartest and most…
Vercel AI Gateway 正式支援 Google Gemma 4 模型★ 70
Vercel Changelog73 days agoReleaseVercel has announced in its product Changelog that the Vercel AI Gateway service now officially supports Google's latest open-source model, Gemma 4. The Vercel…
Google 發表 Gemma 4：專為裝置端設計的前沿多模態開放模型★ 85
Hugging Face Blog74 days agoReleaseGoogle and Hugging Face have jointly announced a new generation of open-weight models — "Gemma 4." This model represents a major breakthrough in on-device AI…