 

最惊险的引导加载程序代码

2025-12-20 分类：前沿哨所阅读(2) 评论(0) 赞(0)

智谱 GLM，支持多语言、多任务推理。从写作到代码生成，从搜索到知识问答，AI 生产力的中国解法。

本文讲述了作者在1999年获得一台HP 9000/720工作站后，尝试在其上运行OpenBSD操作系统的经历。文章详细描述了硬件背景、操作系统移植的挑战，以及开源社区中Michael Shalayeff和Matt Fredette等人的努力。尽管面临硬件故障和缺乏文档的困难，作者通过测试和参与，见证了PA-RISC架构上OpenBSD移植的进展。这段历史故事为科技读者提供了早期开源软件开发和硬件移植的宝贵见解，展示了开源社区的坚持和协作精神。

原文链接：Hacker News

赞(0)

未经允许不得转载：Toy Tech Blog » 最惊险的引导加载程序代码

分享到

评论抢沙发

快讯

Computer-Based Test Reform for IT Certification: Increased Difficulty and Emerging Technologies as New Challenges

The difficulty of the computer-based test for IT certification has significantly increased. Predicting essay topics is no longer effective, as the exam scope now covers cutting-edge technologies across all textbook sections. The proportion of past questions in case studies and comprehensive exams has decreased, while the focus on emerging technologies has grown, requiring candidates to have broader knowledge backgrounds. For example, in the Advanced Information System Management and Advanced System Architect exams, essay topics on cutting-edge technologies such as uncertain performance domains, cloud operations, and multimodal databases have stumped most candidates. Case studies now assess technical stacks like Redis and MongoDB, affecting embedded and front-end developers. Overall pass rates have dropped, reflecting the new challenges that the computer-based test reform poses to IT professionals.

原文链接：V2EX 分享发现

6分钟前
软考机考改革：难度升级，前沿技术成新挑战

软考机考后难度显著升级。论文押题失效，出题范围更前沿，覆盖教材各部分；案例和综合科目真题占比降低，前沿技术比例提高，要求考生知识背景更丰富。以高级信息系统管理和高级系统架构师为例，论文题目如不确定绩效域、云上运维、多模态数据库等前沿技术，拦住多数考生；案例考核Redis、MongoDB等技术栈，影响嵌入式、前端开发者。整体通过率降低，反映机考改革对IT专业人才的新挑战。

原文链接：V2EX 分享发现

6分钟前
Claude Chrome Extension: API Configuration and Usage Guide

This article provides a detailed guide on installing and configuring the Claude for Chrome extension, including modifying the baseurl in assets files, adding CSP rules to manifest.json, loading the extension in Chrome developer mode, and adding API keys in storage. The author shares their personal experience, highlighting that the extension's greatest value lies in its page font style resources and the reference value of tool descriptions. The article also mentions that OAuth has organizational limitations, but the ApiKey mode can solve this issue. For users wanting to use Claude AI features in Chrome browser, this guide provides practical technical reference.

Original Link:Linux.do

1小时前
MiniMax-M2.1 Model Review: Performance Comparable to Sonnet 4.5 with Image Recognition Capabilities

According to the Linux.do community, the MiniMax-M2.1 model has completed preliminary testing and evaluation. Based on test results, the model's capabilities are approximately equal to OpenAI's Sonnet 4.5, performing better than Haiku 4.5. The review notes that MiniMax-M2.1 has image recognition capabilities but shows some deficiencies in frontend knowledge. Community members are further testing its performance in real-world application scenarios, particularly whether it can be seamlessly integrated into CC systems. This evaluation provides AI professionals with a reference for new model performance, helping to understand the current development trends in AI models.

Original Link:Linux.do

1小时前
Claude Chrome扩展：API配置与使用指南

本文详细介绍了Claude for Chrome扩展的安装配置过程，包括修改assets文件中的baseurl、在manifest.json中添加CSP规则、在Chrome开发者模式下加载扩展，以及在存储处添加API密钥等步骤。作者分享了自己的使用体验，指出该扩展的最大价值在于页面字体样式等资源以及tool描述的参考价值。文章还提到OAuth存在组织限制，但ApiKey模式可以解决这一问题。对于想要在Chrome浏览器中使用Claude AI功能的用户，这篇指南提供了实用的技术参考。

原文链接：Linux.do

1小时前
MiniMax-M2.1模型评测：性能接近Sonnet 4.5，具备图像识别能力

据Linux.do社区消息，MiniMax-M2.1模型已完成初步测试评估。根据测试结果，该模型能力约等于OpenAI的Sonnet 4.5，表现优于Haiku 4.5。评测指出，MiniMax-M2.1具备图像识别能力，但在前端知识储备方面有所不足。社区成员正进一步测试其在实际应用场景中的表现，特别是能否在CC系统中无感切入。这一评测为AI从业者提供了新模型性能参考，有助于了解当前AI模型发展态势。

原文链接：Linux.do

1小时前