用创业者的成本思维使用Claude：缓存不变的系统提示词与工具定义，输入成本能降到约十分之一。
I am running an AI-native startup. Early on though, I would spend so much time staying ahead of every new AI hot take that I lost focus on what I need…
超级启动器SuperCmd V2大幅优化中文识别，双击修饰键呼出，效率控必试
>下载少数派2.0客户端、关注少数派公众号，解锁全新阅读体验📰>实用、好用的正版软件，少数派为你呈现🚀 查看全文
多模态大模型提速新思路：按角色分区裁剪视觉token，兼顾效率与精度
arXiv:2608.07088v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) encode images as long visual token sequences, making prefil…
EASy为LLM智能体系统提速增效，直击推理与行动协同的优化痛点
arXiv:2608.04588v1 Announce Type: cross Abstract: Agentic systems have emerged as a promising paradigm for solving complex tasks by coordinating speci…
多模态大模型如何在推理中平衡效率与效果？这篇ACM MM论文提出无训练注意力引导切换，值得关注。
arXiv:2608.03450v1 Announce Type: cross Abstract: Reasoning in Multimodal Large Language Models (MLLMs) requires both fine-grained visual perception a…
稀疏注意力中因果信息流被重新定义，Stem提出新视角，为长上下文模型效率与准确性兼得提供关键思路。
arXiv:2603.06274v2 Announce Type: replace-cross Abstract: The quadratic computational complexity of self-attention remains a fundamental bottleneck fo…
结构感知数据组织新方法，直击LLM后训练效率痛点，值得技术党深读
arXiv:2607.27273v1 Announce Type: new Abstract: Post-training of large language models is expensive, and existing efficiency improvements mainly focus…
用XPath代替循环查询，一招让Power Automate流程更高效。
When I started building Power Automate solutions, my answer to almost every array was the same: Use "Apply to each." It worked. My flows ran successfu…
结构化9-pass预计算，让RAG运行时直接查索引而非重新推导，效率拉满的开源新方案。
Hi everyone, I've been working on a knowledge retrieval system that can create distributable applications for several months and I'd love to share it …
谷歌打造新款AI芯片Frozen v2，专为提升Gemini模型效率，2028年面世，巨额投资押注未来。
Alphabet, Google's parent company, is reportedly working on a new chip designed to make its Gemini models run much more efficiently.
开源的控制框架，让AI代理在迭代中知道何时停止，避免无限自嗨式的优化浪费资源。
Article URL: https://github.com/Danny-de-bree/bound Comments URL: https://news.ycombinator.com/item?id=48957765 Points: 2 # Comments: 1
针对大模型CoT推理中“过度思考”的冗余步骤，用最优停止策略动态剪枝，降本增效不丢精度。
arXiv:2607.11089v1 Announce Type: new Abstract: Large Language Models (LLMs) have achieved remarkable success in complex reasoning tasks through Chain…
OpenAI Codex与ChatGPT Work对Plus/Pro/Business订阅者暂时取消5小时使用限制，并预告GPT-5.6 Sol模型将优化效率，提升任务完成量。
IT之家 7 月 13 日消息，OpenAI Codex 工程负责人蒂博 · 索蒂奥（IT之家注：Thibault Sottiaux）今天在 X 平台宣布，Codex 和 ChatGPT Work 两大工具已暂时取消 Plus / Business 和 Pro 订阅的 5 小时使用限制。 这名员工还…
提出DeLS-Spec方法，解耦长短上下文实现并行投机解码，大幅提升大模型推理效率。
arXiv:2607.07409v1 Announce Type: new Abstract: Speculative decoding accelerates LLM inference by drafting multiple tokens and verifying them in paral…
一次建索引，查询不再递归：Skill Retriever用10K分类体系让AI Agent技能召回从243次调用降到几发，带评分和加载级别。
Article URL: https://github.com/ChonSong/skill-retriever Comments URL: https://news.ycombinator.com/item?id=48827939 Points: 1 # Comments: 0
防提示注入+节省token，基准测试第一的AI安全工具，效率与安全兼得。
Prompt injection and token savings - #1 in benchmarks Discussion | Link
提出条件性记忆访问机制，引导长上下文LLM更智能地选择注意力范围，显著提升效率与性能，ICML 2026最新成果。
arXiv:2603.17484v2 Announce Type: replace-cross Abstract: Language models struggle to generalize beyond pretraining context lengths, limiting long-hor…
提出PLoRA方法，实现大语言模型的高效并发LoRA训练，显著提升微调效率与资源利用率。
arXiv:2508.02932v2 Announce Type: replace Abstract: Low-Rank Adaptation (LoRA) has gained popularity as a fine-tuning approach for Large Language Mode…
让LLM agent告别昂贵低效：Auto编译器将实时行为中确定性部分提取为验证程序，实现智能体行为的精准编译与加速。
arXiv:2607.04542v1 Announce Type: cross Abstract: Every LLM agent run re-derives its behavior token by token on a frontier model: brilliant, expensive…