BD Tech Talks
Accessible analysis of AI, robotics, and emerging tech.
讀取它的 RSS所屬板塊:具身智慧
BD Tech Talks 最新
- How Next-Concept Modeling rethinks AI text generation2 天前
- Rethinking software security from first principles in the age of AI8 天前
- Unpacking the success of Jev and AI decision models26 天前
- BDH-CQ combines in-context learning with reasoning outside the token stream2026-10-06
- How LLM watermarking can change AI agent behavior2026-09-28
- AI developer & agentic AI events you shouldn’t miss this fall/winter (2026)2026-09-13
- OpenAI’s Cursor cutoff makes the ultimate business case for open-source AI2026-09-04
- Life, death, and the inevitable rise of AI: An argumentative reassessment2026-08-29
- The silent data leak hidden inside encrypted reasoning traces of frontier AI models2026-08-17
- Why the agent harness matters as much as the model in AI security2026-08-11
- Beyond ReAct: Building the modern AI agent stack for massive tool ecosystems2026-08-05
- 為什麼Goodfire的區塊稀疏特徵提取器是AI可解釋性的一項突破2026-07-29
- 超越被動RAG：如何為AI代理實施主動記憶重構2026-07-21
- 自我提升的Harness如何重寫代理工程手冊2026-07-13
- OneOdio Studio Max 2評測：科技愛好者的超低延遲耳機2026-07-12
- Nvidia的ASPIRE框架如何透過自我提升AI加速機器人程式設計2026-07-06
- AI軍備競賽從智慧模型轉向全棧基礎設施2026-06-30
- 解密迴圈工程：從AI代理獲得更多成效，避免迴圈過度使用2026-06-22
- 為何LLM應停止朗聲思考（以及思考鏈之後的發展）2026-06-15
- 超越感覺編碼：Codev 3.0如何打造AI驅動的開發團隊2026-06-08
- 代理人AI的未來關鍵在於組件2026-06-01
- Cursor的Composer 2.5如何使用自我蒸餾，在程式碼領域超越前沿LLM2026-05-25
- 垂直整合作為AI基礎設施：21D全拱植入系統教會我們如何建構自主臨床AI2026-05-24
- 為什麼沙盒化OpenClaw無法阻止資料外洩2026-05-18
- Google帶來多標記預測Gemma 4 LLMs2026-05-11
- 記憶稀疏注意力如何將LLM記憶擴展至1億個標記2026-05-04
- Claude Code正在將API金鑰洩漏至公開套件註冊表2026-04-27
- Anthropic的MCP漏洞：當「預期行為」變成供應鏈噩夢2026-04-20
- LLM自我蒸餾的悖論：更快的推理，更弱的泛化能力2026-04-13
- 為什麼架構工程正成為新的 AI 护城河2026-04-06