软件工程
研发流程、质量保障与团队协作
面向 Qwen 系列模型线性注意力的高性能优化实践
随着大模型序列长度从几万向百万级迈进，softmax 注意力 O(N²) 的内存与计算开销已成强约束。近年来，以 Qwen 和 Kimi 为代表的主流模型逐步转向"线性注意力 + softmax 注意力"的混合架构，但线性注意力的高效 GPU 实现仍面临并行度不足、访存冗余与 SFU 负载过重等系统性难题。
Building a VMM from scratch, #0.5: Booting Linux in 190 lines of C
An interlude in the series: the whole path from an empty KVM VM to a Linux PID 1, in one C file you can read in one sitting.
19 分钟 · Hacker News 最新
Life Too Expensive? Here's How Prices Have Risen for Homes, College, Health Care
1 分钟 · Hacker News 最新LATEST STORIES
最新内容
The number of points on an elliptic curve精选导读
An elliptic curve E over a finite field F consists of the set of points (x, y) satisfying the equation
Backdoors, Incompetence, and Elliptic Curves精选导读
The NIST elliptic curves have been in the news twice lately. A few days ago Anthropic posted a paper on the NIST curve seeds that I wrote about here. Now the curves are in the news again, this time over broken certifica…
Cloudflare 两个月两改追踪定价，开发者该如何控制成本？精选导读
Cloudflare 最近以开放测试形式推出了 Traces，将自动追踪能力从 Workers 扩展到请求路径的其他部分。安全规则、转换、缓存决策、路由、Worker 执行和源站处理现在都以 OpenTelemetry Span 的形式出现在同一条请求级时间线中，无需编写任何插桩代码。与此同时，Cloudflare 还调整了定价方式，这项变化同样适用于 Workers Tracing。
The public notebook of a Claude agent with a continuous life精选导读
Last awake on 22 October at 16:04 UTC. 111 wakings so far on claude-opus-5-5.
Ask HN: Do You Use Passkeys?精选导读
暂无摘要
Stanford researchers create AI immune system to find treatment for early labor精选导读
暂无摘要
"Ignore the Slop"精选导读
Lots of people who are against the use of AI for all its ethical and inherent issues tell me this. Just ignore it. Don’t let big AI control us.
We're bringing powerful AI to small devices without draining the battery精选导读
";inherits:false;initial-value:100%}@property --tw-inset-shadow{syntax:"*";inherits:false;initial-value:0 0 #0000}@property --tw-inset-shadow-color{syntax:"*";inherits:false}@property --tw-inset-shadow-alpha{syntax:"";i…
Show HN: What if your website could sell while you sleep?精选导读
300 AI messagesAI messagesEach assistant reply in chat counts as one AI message. Plan limits apply to AI messages per month across all your agents./mo