\u65e5\u671f\uff1a2026-04-29<\/p>\n
\u672c\u671f\u805a\u7126\uff1a\u91cd\u70b9\u5173\u6ce8\u6a21\u578b\u53d1\u5e03\u4e0e release notes\u3001\u5b98\u65b9 engineering blog\u3001AI coding \/ agent \/ SRE\u3001\u8bc4\u6d4b\u699c\u5355\u53d8\u5316\u3001\u5f00\u53d1\u8005\u5b9e\u8df5\u535a\u5ba2\u3001\u6846\u67b6\u751f\u6001\u3001\u5f00\u6e90\u6a21\u578b\u4e0e\u771f\u5b9e\u7528\u6237\u89c6\u89d2\uff1b\u5f53 HN\u3001Reddit\u3001Hugging Face \u7b49\u793e\u533a\u6e90\u53ef\u8bbf\u95ee\u65f6\u4f18\u5148\u7eb3\u5165\u3002<\/p>\n
Artificial Analysis \u6700\u65b0\u6a21\u578b\u6392\u540d\u89c2\u5bdf<\/strong>\uff08Artificial Analysis\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>Artificial Analysis \u6700\u65b0\u6a21\u578b\u6392\u540d\u663e\u793a\uff0cGPT-5.5 (xhigh) \u4ee5 60 \u5206\u7684\u667a\u80fd\u6307\u6570\u4f4d\u5c45\u699c\u9996\uff0cGPT-5.5 (high) \u4ee5 59 \u5206\u7d27\u968f\u5176\u540e\u3002Claude Opus 4.7 (Max Effort) \u4e0e Gemini 3.1 Pro Preview \u5e76\u5217\u7b2c\u4e09\uff0c\u5747\u4e3a 57 \u5206\u3002\u5728\u8f93\u51fa\u901f\u5ea6\u65b9\u9762\uff0cMercury 2 \u4ee5\u6bcf\u79d2 687 \u4e2a token \u9886\u5148\uff0cGranite 3.3 8B \u4ee5 333 t\/s \u4f4d\u5217\u7b2c\u4e8c\u3002\u5ef6\u8fdf\u6700\u4f4e\u7684\u662f Ministral 3 3B\uff080.45 \u79d2\uff09\u548c LFM2 24B A2B\uff080.50 \u79d2\uff09\u3002\u8be5\u5e73\u53f0\u76ee\u524d\u5171\u8bc4\u4f30\u4e86 361 \u4e2a\u6a21\u578b\uff0c\u63d0\u4f9b\u667a\u80fd\u3001\u901f\u5ea6\u3001\u4ef7\u683c\u7b49\u591a\u7ef4\u5ea6\u5bf9\u6bd4\u3002<\/p>\n English Summary:<\/strong> Artificial Analysis's latest model rankings show GPT-5.5 (xhigh) leading with an Intelligence Index score of 60, followed by GPT-5.5 (high) at 59. Claude Opus 4.7 (Max Effort) ties with Gemini 3.1 Pro Preview at 57 for third place. For output speed, Mercury 2 leads at 687 tokens\/s, with Granite 3.3 8B at 333 t\/s. Lowest latency models are Ministral 3 3B (0.45s) and LFM2 24B A2B (0.50s). The platform evaluates 361 models across intelligence, speed, price, and other metrics.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Featured An update on recent Claude Code quality reports<\/strong>\uff08Anthropic Engineering\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>Anthropic \u53d1\u5e03\u5173\u4e8e Claude Code \u8d28\u91cf\u95ee\u9898\u7684\u590d\u76d8\u62a5\u544a\u30024 \u6708 16 \u65e5 Opus 4.7 \u53d1\u5e03\u65f6\uff0c\u56e2\u961f\u4e3a\u51cf\u5c11\u6a21\u578b\u5197\u957f\u8f93\u51fa\u800c\u6dfb\u52a0\u7684\u7cfb\u7edf\u63d0\u793a\u8bcd\uff08\u9650\u5236\u5de5\u5177\u8c03\u7528\u95f4\u6587\u672c\u4e0d\u8d85\u8fc7 25 \u8bcd\u3001\u6700\u7ec8\u56de\u590d\u4e0d\u8d85\u8fc7 100 \u8bcd\uff09\u610f\u5916\u5bfc\u81f4\u667a\u80fd\u6c34\u5e73\u663e\u8457\u4e0b\u964d\u3002\u6b64\u5916\uff0c\u4e00\u9879\u7f13\u5b58\u4f18\u5316\u9519\u8bef\u5730\u4e22\u5f03\u4e86\u5148\u524d\u7684\u63a8\u7406\u5185\u5bb9\uff0c\u5bfc\u81f4\u4ee3\u7801\u5ba1\u67e5\u4ee3\u7406\u4e22\u5931\u4e0a\u4e0b\u6587\u3002\u53d1\u73b0\u95ee\u9898\u540e\uff0cAnthropic \u5df2\u4e8e 4 \u6708 7 \u65e5\u5c06\u6240\u6709\u7528\u6237\u9ed8\u8ba4\u8bbe\u7f6e\u6062\u590d\u4e3a Opus 4.7 \u4f7f\u7528 xhigh effort\uff0c\u5e76\u4e8e 4 \u6708 10 \u65e5\u5728 v2.1.101 \u7248\u672c\u4e2d\u4fee\u590d\u4e86\u7f13\u5b58 bug\u3002<\/p>\n English Summary:<\/strong> Anthropic published a postmortem on Claude Code quality issues. A system prompt change to reduce verbosity (limiting text between tool calls to \u226425 words and final responses to \u2264100 words), shipped with Opus 4.7 on April 16, unexpectedly degraded intelligence. Additionally, a caching optimization incorrectly dropped prior reasoning from conversation history, causing code review agents to lose context. Anthropic reverted the effort level defaults to xhigh for Opus 4.7 on April 7 and fixed the caching bug in v2.1.101 on April 10.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Scaling Managed Agents: Decoupling the brain from the hands<\/strong>\uff08Anthropic Engineering\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>Anthropic \u5de5\u7a0b\u56e2\u961f\u5206\u4eab Managed Agents \u67b6\u6784\u8bbe\u8ba1\u7406\u5ff5\uff0c\u6838\u5fc3\u601d\u60f3\u662f\u5c06"\u5927\u8111"\uff08Claude \u7684\u667a\u80fd\uff09\u4e0e"\u624b"\uff08\u5177\u4f53\u6267\u884c\u4efb\u52a1\u7684 harness\uff09\u89e3\u8026\u3002\u8be5\u5143\u67b6\u6784\u901a\u8fc7 Session \u6301\u4e45\u5316\u5b58\u50a8\u4e8b\u4ef6\u6d41\uff0c\u63d0\u4f9b getEvents() \u63a5\u53e3\u8ba9\u6a21\u578b\u7075\u6d3b\u68c0\u7d22\u4e0a\u4e0b\u6587\uff0c\u800c\u975e\u7b80\u5355\u7d2f\u79ef\u804a\u5929\u65e5\u5fd7\u3002\u8fd9\u79cd\u8bbe\u8ba1\u5141\u8bb8\u4e0d\u540c\u9886\u57df\u4f7f\u7528\u4e13\u95e8\u7684 harness\uff08\u5982 Claude Code \u6216\u7279\u5b9a\u4efb\u52a1\u4ee3\u7406\uff09\uff0c\u540c\u65f6\u4fdd\u6301\u4e0a\u4e0b\u6587\u7684\u53ef\u6062\u590d\u6027\u548c\u53ef\u67e5\u8be2\u6027\u3002\u56e2\u961f\u5f3a\u8c03\u5c06\u4e0a\u4e0b\u6587\u7ba1\u7406\u4e0b\u653e\u5230 harness \u5c42\uff0c\u4f7f\u7cfb\u7edf\u80fd\u9002\u914d\u672a\u6765\u6a21\u578b\u6f14\u8fdb\uff0c\u800c\u4e0d\u5fc5\u9884\u6d4b\u5177\u4f53\u7684\u4e0a\u4e0b\u6587\u5de5\u7a0b\u9700\u6c42\u3002<\/p>\n English Summary:<\/strong> Anthropic's engineering team shared the design philosophy behind Managed Agents, decoupling the "brain" (Claude's intelligence) from the "hands" (task-specific harnesses). The meta-architecture uses Sessions to durably store event streams, providing a getEvents() interface for flexible context retrieval rather than accumulating chat logs. This allows specialized harnesses for different domains while maintaining recoverable, queryable context.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Introducing Claude Opus 4.7<\/strong>\uff08Anthropic News\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>Anthropic \u6b63\u5f0f\u53d1\u5e03 Claude Opus 4.7\uff0c\u5728\u591a\u6b65\u9aa4\u4efb\u52a1\u6548\u7387\u4e0a\u521b\u4e0b\u5185\u90e8\u7814\u7a76\u4ee3\u7406\u57fa\u51c6\u6d4b\u8bd5\u7684\u6700\u4f73\u8868\u73b0\uff0c\u516d\u9879\u6a21\u5757\u603b\u5206 0.715 \u5e76\u5217\u7b2c\u4e00\uff0c\u957f\u4e0a\u4e0b\u6587\u6027\u80fd\u6700\u4e3a\u7a33\u5b9a\u3002\u5728\u91d1\u878d\u5206\u6790\u6a21\u5757\u5f97\u5206\u4ece 4.6 \u7248\u7684 0.767 \u63d0\u5347\u81f3 0.813\uff0c\u6f14\u7ece\u63a8\u7406\u80fd\u529b\u4e5f\u663e\u8457\u589e\u5f3a\u3002Databricks \u6d4b\u8bd5\u663e\u793a\u5176\u5728 OfficeQA Pro \u4e0a\u6587\u6863\u63a8\u7406\u9519\u8bef\u51cf\u5c11 21%\uff1bRakuten \u7684 SWE-Bench \u6d4b\u8bd5\u663e\u793a\u751f\u4ea7\u4efb\u52a1\u89e3\u51b3\u7387\u662f 4.6 \u7684\u4e09\u500d\uff0c\u4ee3\u7801\u8d28\u91cf\u548c\u6d4b\u8bd5\u8d28\u91cf\u5747\u6709\u4e24\u4f4d\u6570\u63d0\u5347\u3002\u5408\u4f5c\u4f19\u4f34\u53cd\u9988\u5176\u5728\u4ee3\u7406\u56e2\u961f\u534f\u4f5c\u3001\u5de5\u5177\u8c03\u7528\u51c6\u786e\u6027\u548c\u89c4\u5212\u80fd\u529b\u65b9\u9762\u8868\u73b0\u7a81\u51fa\u3002<\/p>\n English Summary:<\/strong> Anthropic officially launched Claude Opus 4.7, achieving the strongest efficiency baseline for multi-step work on internal research-agent benchmarks with a top score of 0.715 across six modules and the most consistent long-context performance. Financial analysis scores improved from 0.767 to 0.813, with stronger deductive reasoning. Databricks testing showed 21% fewer document reasoning errors on OfficeQA Pro; Rakuten's SWE-Bench showed 3x more production tasks resolved versus Opus 4.6, with double-digit gains in code and test quality.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n How Slack Manages Context in Long-running Multi-agent Systems<\/strong>\uff08InfoQ AI\/ML\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>Slack \u5de5\u7a0b\u5e08\u5206\u4eab\u5728\u957f\u671f\u8fd0\u884c\u591a\u4ee3\u7406\u7cfb\u7edf\u4e2d\u7ba1\u7406\u4e0a\u4e0b\u6587\u7684\u7ecf\u9a8c\u3002\u56e2\u961f\u91c7\u7528\u534f\u8c03\u5668\/\u5206\u53d1\u5668\u67b6\u6784\uff0c\u7531\u4e2d\u592e\u534f\u8c03\u5668\u4f5c\u4e3a\u51b3\u7b56\u8005\uff0c\u5c06\u8bf7\u6c42\u5206\u53d1\u7ed9\u4e13\u5bb6\u4ee3\u7406\u548c\u8bc4\u4f30\u4ee3\u7406\u3002\u4e3a\u907f\u514d\u4e0a\u4e0b\u6587\u7a97\u53e3\u586b\u6ee1\u5bfc\u81f4\u54cd\u5e94\u8d28\u91cf\u4e0b\u964d\uff0cSlack \u653e\u5f03\u7d2f\u79ef\u804a\u5929\u65e5\u5fd7\u7684\u505a\u6cd5\uff0c\u8f6c\u800c\u4f7f\u7528\u7ed3\u6784\u5316\u5185\u5b58\u3001\u9a8c\u8bc1\u673a\u5236\u548c\u63d0\u70bc\u7684"\u771f\u76f8"\u6765\u7ef4\u6301\u7cfb\u7edf\u4e00\u81f4\u6027\u3002\u534f\u8c03\u5668\u7ef4\u62a4\u7684\u65e5\u5fd7\u5305\u542b\u53d1\u73b0\u3001\u89c2\u5bdf\u3001\u51b3\u7b56\u3001\u95ee\u9898\u548c\u5047\u8bbe\uff0c\u4e3a\u6240\u6709\u4ee3\u7406\u63d0\u4f9b\u5171\u540c\u53d9\u4e8b\u3002\u8bc4\u4f30\u4ee3\u7406\u8d1f\u8d23\u9a8c\u8bc1\u4e13\u5bb6\u4ee3\u7406\u7684\u5de5\u4f5c\uff0c\u901a\u8fc7\u8bc4\u5206\u7cfb\u7edf\u8bc6\u522b\u591a\u65b9\u4f50\u8bc1\u7684\u53ef\u4fe1\u53d1\u73b0\uff0c\u9632\u6b62\u5e7b\u89c9\u6216\u6570\u636e\u8bef\u8bfb\u3002<\/p>\n English Summary:<\/strong> Slack engineers shared their approach to context management in long-running multi-agent systems. Using a coordinator\/dispatcher architecture, a central coordinator dispatches requests to expert and critic agents. To prevent context window overflow from degrading response quality, Slack moved away from accumulating chat logs toward structured memory, validation, and distilled truth. The coordinator's journal contains findings, observations, decisions, questions, and hypotheses, providing a common narrative.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Amazon is already offering new OpenAI products on AWS<\/strong>\uff08TechCrunch AI\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>OpenAI\u4e0e\u5fae\u8f6f\u4fee\u8ba2\u5408\u4f5c\u534f\u8bae\u3001\u7ed3\u675f\u4ea7\u54c1\u72ec\u5bb6\u6388\u6743\u540e\uff0cAWS\u8fc5\u901f\u5ba3\u5e03\u5728Amazon Bedrock\u5e73\u53f0\u4e0a\u7ebfOpenAI\u5168\u7cfb\u4ea7\u54c1\uff0c\u5305\u62ecGPT-5.5\u7b49\u524d\u6cbf\u6a21\u578b\u3001\u4ee3\u7801\u751f\u6210\u5de5\u5177Codex\u4ee5\u53ca\u5168\u65b0\u7684Bedrock Managed Agents\u4ee3\u7406\u670d\u52a1\u3002\u8be5\u4ee3\u7406\u670d\u52a1\u4e13\u4e3aOpenAI\u63a8\u7406\u6a21\u578b\u8bbe\u8ba1\uff0c\u63d0\u4f9b\u4ee3\u7406\u5f15\u5bfc\u4e0e\u5b89\u5168\u7ba1\u63a7\u529f\u80fd\u3002\u6b64\u4e3e\u6807\u5fd7\u7740OpenAI\u4e0e\u5fae\u8f6f\u5173\u7cfb\u6301\u7eed\u6076\u5316\uff0c\u53cc\u65b9\u5404\u81ea\u6295\u5411\u5bf9\u65b9\u6700\u5927\u7ade\u4e89\u5bf9\u624b\u2014\u2014OpenAI\u4e0eAWS\/Oracle\u5408\u4f5c\uff0c\u5fae\u8f6f\u5219\u4e0eAnthropic\u6df1\u5316\u8054\u76df\u3002\u4e9a\u9a6c\u900a\u8868\u793a\u8fd9\u53ea\u662f\u53cc\u65b9\u6df1\u5ea6\u5408\u4f5c\u7684\u5f00\u59cb\uff0c\u4f01\u4e1a\u5ba2\u6237\u73b0\u53ef\u5728\u719f\u6089\u7684AWS\u73af\u5883\u4e2d\u6784\u5efa\u5b89\u5168\u7684AI\u5e94\u7528\u3002<\/p>\n English Summary:<\/strong> Following OpenAI's revised agreement with Microsoft that ended exclusive licensing, AWS quickly announced the availability of OpenAI's full product suite on Amazon Bedrock, including GPT-5.5 frontier models, the Codex coding service, and the new Bedrock Managed Agents service designed for OpenAI's reasoning models with agent steering and security features. This move signals the deteriorating OpenAI-Microsoft relationship, with each turning to their partner's biggest rival\u2014OpenAI partnering with AWS and Oracle, while Microsoft deepens ties with Anthropic. Amazon calls this "the beginning of a deeper collaboration," allowing enterprise customers to build secure AI within their existing AWS environments.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Migrating a text agent to a voice assistant with Amazon Nova 2 Sonic<\/strong>\uff08AWS ML Blog\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>AWS\u673a\u5668\u5b66\u4e60\u535a\u5ba2\u53d1\u5e03\u6280\u672f\u6307\u5357\uff0c\u8be6\u89e3\u5982\u4f55\u4f7f\u7528Amazon Nova 2 Sonic\u5c06\u4f20\u7edf\u6587\u672c\u667a\u80fd\u4f53\u8fc1\u79fb\u4e3a\u5bf9\u8bdd\u5f0f\u8bed\u97f3\u52a9\u624b\u3002\u6587\u7ae0\u5bf9\u6bd4\u4e86\u6587\u672c\u4e0e\u8bed\u97f3\u667a\u80fd\u4f53\u5728\u8f93\u5165\u65b9\u5f0f\u3001\u54cd\u5e94\u98ce\u683c\u3001\u5ef6\u8fdf\u9884\u7b97\u3001\u8f6e\u6b21\u7ba1\u7406\u7b49\u65b9\u9762\u7684\u6838\u5fc3\u5dee\u5f02\uff0c\u5f3a\u8c03\u8bed\u97f3\u573a\u666f\u9700\u8981\u8d85\u4f4e\u5ef6\u8fdf\u3001\u53cc\u5411\u6d41\u5f0f\u4f20\u8f93\u3001\u6253\u65ad\u5904\u7406\uff08barge-in\uff09\u548c\u8bed\u97f3\u6d3b\u52a8\u68c0\u6d4b\uff08VAD\uff09\u3002Nova 2 Sonic\u4f5c\u4e3a\u539f\u751f\u8bed\u97f3\u5230\u8bed\u97f3\u6a21\u578b\uff0c\u5185\u7f6eASR\u3001\u63a8\u7406\u3001TTS\u80fd\u529b\uff0c\u652f\u6301\u5f02\u6b65\u5de5\u5177\u8c03\u7528\uff0c\u5141\u8bb8\u5728\u5de5\u5177\u6267\u884c\u671f\u95f4\u4fdd\u6301\u5bf9\u8bdd\u6d41\u7545\u3002\u6587\u7ae0\u63d0\u4f9b\u4e86\u57fa\u4e8eStrands Agents\u7684\u4ee3\u7801\u793a\u4f8b\uff0c\u5c55\u793a\u5982\u4f55\u590d\u7528\u73b0\u6709\u5de5\u5177\u548c\u7cfb\u7edf\u63d0\u793a\uff0c\u540c\u65f6\u9488\u5bf9\u8bed\u97f3\u573a\u666f\u4f18\u5316\u54cd\u5e94\u957f\u5ea6\u3001\u5ef6\u8fdf\u548c\u5bf9\u8bdd\u98ce\u683c\uff0c\u5b9e\u73b0\u4ece\u6587\u672c\u5230\u8bed\u97f3\u7684\u5e73\u6ed1\u67b6\u6784\u8fc1\u79fb\u3002<\/p>\n English Summary:<\/strong> AWS Machine Learning Blog published a technical guide on migrating text agents to conversational voice assistants using Amazon Nova 2 Sonic. The post compares key differences between text and voice agents in input methods, response styles, latency budgets, and turn-taking, emphasizing voice requirements for ultra-low latency, bidirectional streaming, barge-in handling, and voice activity detection (VAD). Nova 2 Sonic, a native speech-to-speech model with built-in ASR, reasoning, and TTS capabilities, supports asynchronous tool calling to maintain conversation flow during tool execution. The article provides code examples using Strands Agents demonstrating how to reuse existing tools and system prompts while optimizing response length, latency, and conversational style for voice interactions.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n [AINews] ImageGen is on the Path to AGI<\/strong>\uff08Latent Space\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>Latent Space\u7684AINews\u680f\u76ee\u63a2\u8ba8\u56fe\u50cf\u751f\u6210\u6a21\u578b\u5728\u901a\u5f80AGI\u9053\u8def\u4e0a\u7684\u4ef7\u503c\u3002\u6587\u7ae0\u6307\u51fa\uff0c\u5c3d\u7ba1\u5404\u5927\u5b9e\u9a8c\u5ba4\u7eb7\u7eb7\u8f6c\u5411\u4ee5\u7f16\u7801\u548c\u4f01\u4e1aAI\u4e3a\u91cd\u70b9\u7684\u65b9\u5411\uff08\u5982Anthropic\u6a21\u5f0f\uff09\uff0cGPT-Image-2\u4ecd\u5728\u521b\u610f\u5e94\u7528\u9886\u57df\u6301\u7eed\u7206\u53d1\uff0c\u4ece\u4e50\u9ad8\u98ce\u683c\u89d2\u8272\u8bbe\u8ba1\u5230\u6559\u80b2\u56fe\u8868\u3001\u6d41\u884c\u6587\u5316\u521b\u4f5c\u5747\u5c55\u73b0\u5f3a\u5927\u80fd\u529b\u3002\u4f5c\u8005\u8ba4\u4e3a\uff0c\u591a\u6a21\u6001\u8bed\u97f3\u548c\u89c6\u89c9\u751f\u6210\u80fd\u529b\uff08\u5305\u62ec\u900f\u660e\u80cc\u666f\u751f\u6210\uff09\u662f\u91ca\u653eAGI\u4e2d"\u901a\u7528\u6027"\u7684\u5173\u952e\u2014\u2014AI\u4e0d\u5e94\u4ec5\u9650\u4e8e\u7f16\u7a0b\u4efb\u52a1\u3002\u6587\u7ae0\u540c\u65f6\u62a5\u9053\u4e86OpenAI\u4e0e\u5fae\u8f6f\u5408\u4f5c\u8c03\u6574\u3001GPT-5.5\u57fa\u51c6\u6d4b\u8bd5\u8868\u73b0\u3001GitHub Copilot\u8f6c\u5411\u6309\u91cf\u8ba1\u8d39\u3001\u5c0f\u7c73\u5f00\u6e90MiMo-V2.5\u7cfb\u5217\u6a21\u578b\u3001Sakana\u7684Conductor\u591a\u667a\u80fd\u4f53\u7cfb\u7edf\u7b49\u6280\u672f\u52a8\u6001\uff0c\u5f3a\u8c03\u56fe\u50cf\u751f\u6210\u4e0e\u4ee3\u7801\u751f\u6210\u7684\u95ed\u73af\u6574\u5408\u6b63\u6210\u4e3a\u7ade\u4e89\u7126\u70b9\u3002<\/p>\n English Summary:<\/strong> Latent Space's AINews column explores how image generation models contribute to the path toward AGI. While major labs pivot toward coding and enterprise AI focus (the "Anthropic model"), GPT-Image-2 continues driving creative applications from Lego-style character designs to educational infographics and pop culture content. The authors argue that multimodal voice and visual generation capabilities\u2014including transparent background generation\u2014are key to unlocking the "General" in AGI, as AI shouldn't be limited to programming tasks. The piece also covers OpenAI's Microsoft partnership adjustments, GPT-5.5 benchmark performance, GitHub Copilot's shift to usage-based billing, Xiaomi's open-source MiMo-V2.5 models, Sakana's Conductor multi-agent system, and other technical developments, emphasizing that integrating image generation with coding workflows is becoming a competitive battleground.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n OpenAI models, Codex, and Managed Agents come to AWS<\/strong>\uff08OpenAI News\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>OpenAI\u5b98\u65b9\u5ba3\u5e03\u4e0eAWS\u6269\u5c55\u6218\u7565\u5408\u4f5c\uff0c\u5c06GPT\u6a21\u578b\u3001Codex\u7f16\u7a0b\u5de5\u5177\u548cManaged Agents\u4ee3\u7406\u670d\u52a1\u5f15\u5165Amazon Bedrock\u5e73\u53f0\uff0c\u73b0\u5df2\u5f00\u542f\u6709\u9650\u9884\u89c8\u3002\u4f01\u4e1a\u5ba2\u6237\u53ef\u5728\u73b0\u6709AWS\u73af\u5883\u4e2d\u76f4\u63a5\u8c03\u7528OpenAI\u80fd\u529b\uff0c\u590d\u7528\u5df2\u6709\u7684\u5b89\u5168\u63a7\u5236\u3001\u8eab\u4efd\u7cfb\u7edf\u548c\u91c7\u8d2d\u6d41\u7a0b\u3002Codex on Bedrock\u652f\u6301\u901a\u8fc7Bedrock API\u914d\u7f6e\uff0c\u517c\u5bb9Codex CLI\u3001\u684c\u9762\u5e94\u7528\u548cVS Code\u6269\u5c55\uff0c\u5ba2\u6237\u6570\u636e\u7531Bedrock\u5904\u7406\uff0c\u7b26\u5408\u6761\u4ef6\u7684\u4f7f\u7528\u53ef\u8ba1\u5165AWS\u4e91\u627f\u8bfa\u3002Bedrock Managed Agents\u7531OpenAI\u63d0\u4f9b\u6280\u672f\u652f\u6301\uff0c\u652f\u6301\u591a\u6b65\u9aa4\u5de5\u4f5c\u6d41\u3001\u5de5\u5177\u4f7f\u7528\u548c\u590d\u6742\u4e1a\u52a1\u6d41\u7a0b\uff0c\u5e2e\u52a9\u4f01\u4e1a\u4ece\u5b9e\u9a8c\u9636\u6bb5\u5feb\u901f\u8fc8\u5411\u751f\u4ea7\u90e8\u7f72\uff0c\u540c\u65f6\u4fdd\u6301\u4e0eAWS\u5b89\u5168\u5408\u89c4\u6807\u51c6\u7684\u4e00\u81f4\u6027\u3002<\/p>\n English Summary:<\/strong> OpenAI officially announced an expanded strategic partnership with AWS, bringing GPT models, Codex coding tools, and Managed Agents to Amazon Bedrock, now available in limited preview. Enterprise customers can access OpenAI capabilities within their existing AWS environments, leveraging current security controls, identity systems, and procurement workflows. Codex on Bedrock supports configuration via the Bedrock API and is compatible with Codex CLI, desktop app, and VS Code extension, with customer data processed by Bedrock and eligible usage counting toward AWS cloud commitments. Bedrock Managed Agents, powered by OpenAI, support multi-step workflows, tool use, and complex business processes, helping organizations move from experimentation to production while maintaining alignment with AWS security and compliance standards.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Physical AI that Moves the World \u2014 Qasar Younis & Peter Ludwig, Applied Intuition<\/strong>\uff08Latent Space\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>Latent Space\u64ad\u5ba2\u4e13\u8bbfApplied Intuition\u8054\u5408\u521b\u59cb\u4eba\u517cCEO Qasar Younis\u4e0eCTO Peter Ludwig\uff0c\u6df1\u5165\u63a2\u8ba8\u8fd9\u5bb6\u4f30\u503c150\u4ebf\u7f8e\u5143\u7684\u7269\u7406AI\u516c\u53f8\u5982\u4f55\u5c06AI\u90e8\u7f72\u5230\u91c7\u77ff\u8bbe\u5907\u3001\u65e0\u4eba\u673a\u3001\u5361\u8f66\u3001\u519b\u8230\u7b49\u6781\u7aef\u73af\u5883\u4e2d\u7684\u7269\u7406\u8f7d\u5177\u3002\u4e24\u4eba\u56de\u987e\u4e86\u516c\u53f8\u4eceYC\u65f6\u671f\u7684\u81ea\u52a8\u9a7e\u9a76\u5de5\u5177\u8d77\u6b65\uff0c\u9010\u6b65\u53d1\u5c55\u4e3a\u6db5\u76d6\u4eff\u771f\u3001\u64cd\u4f5c\u7cfb\u7edf\u548c\u57fa\u7840\u6a21\u578b\u7684\u7efc\u5408\u5e73\u53f0\u3002\u6838\u5fc3\u89c2\u70b9\u5305\u62ec\uff1a\u7269\u7406AI\u4e0e\u5c4f\u5e55AI\u7684\u672c\u8d28\u5dee\u5f02\u5728\u4e8e\u5b89\u5168\u5173\u952e\u6027\u8981\u6c42\uff1b\u5f53\u524d\u74f6\u9888\u5e76\u975e\u6a21\u578b\u667a\u80fd\uff0c\u800c\u662f\u5982\u4f55\u5728\u5ef6\u8fdf\u3001\u529f\u8017\u3001\u6210\u672c\u548c\u5b89\u5168\u7ea6\u675f\u4e0b\u5c06\u6a21\u578b\u90e8\u7f72\u5230\u5d4c\u5165\u5f0f\u786c\u4ef6\uff1b\u516c\u53f8\u6b63\u81f4\u529b\u4e8e\u6210\u4e3a"\u7269\u7406\u673a\u5668\u7684Android"\uff0c\u4e3a\u788e\u7247\u5316\u7684\u8f66\u8f7d\u8f6f\u4ef6\u6808\u63d0\u4f9b\u7edf\u4e00\u64cd\u4f5c\u7cfb\u7edf\u5c42\uff1b\u540c\u65f6\u5206\u4eab\u4e86\u5728\u7aef\u5230\u7aef\u81ea\u52a8\u9a7e\u9a76\u3001\u4e16\u754c\u6a21\u578b\u3001\u4eff\u771f\u9a8c\u8bc1\u3001\u7edf\u8ba1\u5b89\u5168\u8bc4\u4f30\u7b49\u524d\u6cbf\u9886\u57df\u7684\u6280\u672f\u5b9e\u8df5\u4e0e\u884c\u4e1a\u6d1e\u5bdf\u3002<\/p>\n English Summary:<\/strong> Latent Space podcast features an in-depth interview with Applied Intuition co-founder\/CEO Qasar Younis and CTO Peter Ludwig, exploring how the $15B physical AI company deploys AI to mining rigs, drones, trucks, warships, and vehicles in adversarial environments. The founders trace the company's evolution from YC-era autonomy tooling to a comprehensive platform spanning simulation, operating systems, and foundation models. Key insights include: the fundamental difference between physical and screen AI lies in safety-critical requirements; the current bottleneck isn't model intelligence but deploying models under latency, power, cost, and safety constraints onto embedded hardware; the company aims to become "Android for physical machines," providing a unified OS layer for fragmented vehicle software stacks; and technical practices in end-to-end autonomy, world models, simulation validation, and statistical safety assessment.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n OpenAI available at FedRAMP Moderate<\/strong>\uff08OpenAI News\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>OpenAI \u5ba3\u5e03 ChatGPT Enterprise \u548c OpenAI API \u5e73\u53f0\u5df2\u83b7\u5f97 FedRAMP Moderate \u6388\u6743\uff0c\u4f7f\u7f8e\u56fd\u8054\u90a6\u673a\u6784\u80fd\u591f\u5b89\u5168\u5730\u91c7\u7528\u524d\u6cbf AI \u6280\u672f\u3002\u8be5\u6388\u6743\u901a\u8fc7 FedRAMP 20x \u5feb\u901f\u901a\u9053\u5b8c\u6210\uff0c\u6807\u5fd7\u7740\u4e91\u539f\u751f\u5b89\u5168\u9a8c\u8bc1\u4e0e\u81ea\u52a8\u5316\u8bc4\u4f30\u7684\u65b0\u6a21\u5f0f\u3002\u8054\u90a6\u673a\u6784\u73b0\u53ef\u5728\u7b26\u5408\u5b89\u5168\u3001\u9690\u79c1\u548c\u6cbb\u7406\u8981\u6c42\u7684\u524d\u63d0\u4e0b\uff0c\u4f7f\u7528\u5305\u62ec GPT-5.5 \u5728\u5185\u7684\u6700\u5f3a\u6a21\u578b\u8fdb\u884c\u79d1\u7814\u8d77\u8349\u3001\u7ffb\u8bd1\u5206\u6790\u3001\u77e5\u8bc6\u7ba1\u7406\u7b49\u5de5\u4f5c\uff0c\u540c\u65f6\u5373\u5c06\u901a\u8fc7\u540c\u4e00\u73af\u5883\u8bbf\u95ee Codex Cloud\u3002\u8be5\u91cc\u7a0b\u7891\u6d88\u9664\u4e86\u653f\u5e9c\u673a\u6784\u5728\u5c16\u7aef AI \u4e0e\u53ef\u4fe1\u90e8\u7f72\u73af\u5883\u4e4b\u95f4\u7684\u9009\u62e9\u56f0\u5883\u3002<\/p>\n English Summary:<\/strong> OpenAI announced that ChatGPT Enterprise and the OpenAI API Platform have achieved FedRAMP Moderate authorization, enabling U.S. federal agencies to securely adopt frontier AI capabilities. The milestone was reached through the FedRAMP 20x accelerated pathway, representing a new model of cloud-native security validation and automated assessment. Federal agencies can now use OpenAI's most powerful models including GPT-5.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Reading today's open-closed performance gap<\/strong>\uff08Interconnects\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>\u672c\u6587\u6df1\u5165\u5206\u6790\u4e86\u5f00\u6e90\u4e0e\u95ed\u6e90\u5927\u6a21\u578b\u4e4b\u95f4\u7684\u6027\u80fd\u5dee\u8ddd\u8bc4\u4f30\u95ee\u9898\uff0c\u6307\u51fa\u5c06\u8fd9\u4e00\u590d\u6742\u52a8\u6001\u7b80\u5316\u4e3a\u5355\u4e00\u6570\u5b57\u4f1a\u63a9\u76d6\u5173\u952e\u7ec6\u8282\u3002\u4f5c\u8005\u6307\u51fa\uff0c\u5f53\u524d\u57fa\u51c6\u6d4b\u8bd5\u6bcf 12-18 \u4e2a\u6708\u5c31\u4f1a\u968f\u884c\u4e1a\u7126\u70b9\u8f6c\u79fb\u800c\u53d8\u5316\uff0c\u4ece\u65e9\u671f\u7684\u804a\u5929\u3001\u6570\u5b66\u80fd\u529b\u8f6c\u5411\u590d\u6742\u4ee3\u7801\u548c\u4ee3\u7406\u4efb\u52a1\u3002\u95ed\u6e90\u524d\u6cbf\u5b9e\u9a8c\u5ba4\u6b63\u6295\u5165\u5de8\u989d\u8d44\u91d1\u638c\u63e1\u4ee3\u7801\u548c\u7ec8\u7aef\u4efb\u52a1\uff0c\u540c\u65f6\u5411\u4f1a\u8ba1\u3001\u6cd5\u5f8b\u3001\u533b\u7597\u7b49\u4e13\u4e1a\u9886\u57df\u63a8\u8fdb\u3002\u5f00\u6e90\u6a21\u578b\u867d\u5728\u90e8\u5206\u57fa\u51c6\u4e0a\u63a5\u8fd1\u95ed\u6e90\u6a21\u578b\uff0c\u4f46\u5728\u9700\u8981\u4e13\u4e1a\u77e5\u8bc6\u548c\u7279\u5b9a\u5de5\u5177\u96c6\u6210\u7684\u65b0\u9886\u57df\u53ef\u80fd\u96be\u4ee5\u8ddf\u4e0a\u3002\u6587\u7ae0\u5f3a\u8c03\uff0c\u8bc4\u4f30\u590d\u6742\u8bed\u8a00\u6a21\u578b\u5de5\u4f5c\u6d41\u672c\u8eab\u4e5f\u662f\u5177\u6709\u6311\u6218\u6027\u7684\u7814\u7a76\u95ee\u9898\u3002<\/p>\n English Summary:<\/strong> This article provides an in-depth analysis of evaluating the performance gap between open and closed-source large language models, arguing that reducing this complex dynamic to a single number obscures crucial nuances. The author notes that benchmark focus shifts every 12-18 months as the industry evolves, moving from early chat and math capabilities toward complex coding and agentic tasks. Closed frontier labs are investing massive resources in mastering code and terminal tasks while pushing into specialized domains like accounting, law, and healthcare. While open models approach closed models on some benchmarks, they may struggle to keep pace in new areas requiring domain expertise and specific tool integrations.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Building an emoji list generator with the GitHub Copilot CLI<\/strong>\uff08GitHub AI\/ML\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>GitHub \u56e2\u961f\u5728 Rubber Duck Thursday \u76f4\u64ad\u6d3b\u52a8\u4e2d\u5c55\u793a\u4e86\u5982\u4f55\u4f7f\u7528 GitHub Copilot CLI \u6784\u5efa\u4e00\u4e2a\u8868\u60c5\u7b26\u53f7\u5217\u8868\u751f\u6210\u5668\u3002\u8be5\u9879\u76ee\u4f7f\u7528 OpenTUI \u6784\u5efa\u7ec8\u7aef\u754c\u9762\u3001GitHub Copilot SDK \u63d0\u4f9b AI \u80fd\u529b\u3001clipboardy \u5904\u7406\u526a\u8d34\u677f\u529f\u80fd\u3002\u5f00\u53d1\u8005\u901a\u8fc7 Plan \u6a21\u5f0f\u4e0e Claude Sonnet 4.6 \u534f\u4f5c\u5236\u5b9a\u65b9\u6848\uff0c\u518d\u7528 Claude Opus 4.7 \u5b9e\u73b0\u4ee3\u7801\uff0c\u6700\u7ec8\u5f97\u5230\u4e00\u4e2a\u53ef\u5c06\u666e\u901a\u5217\u8868\u81ea\u52a8\u8f6c\u6362\u4e3a\u5e26\u76f8\u5173\u8868\u60c5\u7b26\u53f7\u683c\u5f0f\u7684\u7ec8\u7aef\u5de5\u5177\u3002\u9879\u76ee\u5c55\u793a\u4e86 Copilot CLI \u7684\u591a\u6a21\u578b\u5de5\u4f5c\u6d41\u3001Autopilot \u6a21\u5f0f\u3001allow-all \u5de5\u5177\u6807\u5fd7\u4ee5\u53ca GitHub MCP \u670d\u52a1\u5668\u7b49\u7279\u6027\u7684\u5b9e\u9645\u5e94\u7528\u3002<\/p>\n English Summary:<\/strong> The GitHub team demonstrated building an emoji list generator using the GitHub Copilot CLI during their Rubber Duck Thursday livestream. The project uses OpenTUI for the terminal interface, GitHub Copilot SDK for AI capabilities, and clipboardy for clipboard functionality. Developers collaborated with Claude Sonnet 4.6 in Plan mode to create a strategy, then implemented the code with Claude Opus 4.7, resulting in a terminal tool that automatically converts plain lists into emoji-enhanced formats. The project showcases practical applications of Copilot CLI's multi-model workflows, Autopilot mode, allow-all tools flag, and GitHub MCP server integration.<\/p>\n \u539f\u6587\u94fe\u63a5<\/a><\/p>\n<\/li>\n Build a personal organization command center with GitHub Copilot CLI<\/strong>\uff08GitHub AI\/ML\uff09<\/p>\n \u4e2d\u6587\u6458\u8981\uff1a<\/strong>GitHub \u5de5\u7a0b\u5e08 Brittany Ellich \u5206\u4eab\u4e86\u5979\u5982\u4f55\u4f7f\u7528 GitHub Copilot CLI \u6784\u5efa\u4e2a\u4eba\u7ec4\u7ec7\u6307\u6325\u4e2d\u5fc3\uff0c\u4ee5\u89e3\u51b3\u6570\u5b57\u4fe1\u606f\u5206\u6563\u5728\u591a\u4e2a\u5e94\u7528\u4e2d\u7684\u95ee\u9898\u3002\u8be5\u9879\u76ee\u662f\u4e00\u4e2a Electron \u684c\u9762\u5e94\u7528\uff0c\u6574\u5408\u4e86\u65e5\u5386\u3001\u4efb\u52a1\u548c\u7b14\u8bb0\u7b49\u529f\u80fd\u5230\u7edf\u4e00\u7684\u89c6\u89c9\u754c\u9762\u4e2d\u3002\u5f00\u53d1\u91c7\u7528"\u5148\u89c4\u5212\u540e\u5b9e\u73b0"\u7684 AI \u8f85\u52a9\u5de5\u4f5c\u6d41\uff1a\u5148\u7528 Copilot \u8fdb\u884c\u9700\u6c42\u8bbf\u8c08\u548c\u65b9\u6848\u5236\u5b9a\uff0c\u518d\u4ea4\u7531 Agent Mode \u5b9e\u73b0\u4ee3\u7801\u3002Ellich \u540c\u65f6\u7ed3\u5408 VS Code \u7684\u540c\u6b65 Agent \u5f00\u53d1\u4e0e Copilot Cloud Agent \u7684\u5f02\u6b65\u4efb\u52a1\u5904\u7406\uff0c\u4ec5\u7528\u4e00\u5929\u5c31\u5b8c\u6210\u4e86 v1 \u7248\u672c\u3002\u9879\u76ee\u4f7f\u7528 React\u3001Vite\u3001Tailwind \u548c WorkIQ MCP \u7b49\u6280\u672f\u6808\u3002<\/p>\n English Summary:<\/strong> GitHub engineer Brittany Ellich shared how she built a personal organization command center using the GitHub Copilot CLI to solve digital fragmentation across multiple apps. The project is an Electron desktop application that unifies calendar, tasks, and notes into a single visual interface. The development followed an AI-assisted "plan-then-implement" workflow: first using Copilot for requirement interviews and planning, then delegating implementation to Agent Mode. Ellich combined synchronous Agent development in VS Code with asynchronous tasks via Copilot Cloud Agent, completing v1 in just one day. The tech stack includes React, Vite, Tailwind, and WorkIQ MCP.<\/p>\n