1
Give your AI sub-agent a budget, not your keys
给AI子代理设预算和时限，别直接交出钥匙，安全又省心
Spawn a sub-agent in CrewAI, LangGraph, AutoGen, or a Claude sub-agent setup and check what it actually holds: your credentials. The parent's keys, at…
2
消息称日本经产省将为支持 Rapidus 再编列 1500 亿日元预算
IT之家 8 月 21 日消息，据《日本经济新闻》和彭博社报道，日本经济产业省计划在其 2027 财年预算案中编列一笔 1,500 亿日元 （IT之家注：现汇率约合 63.65 亿元人民币） 的开支， 以投资形式进一步支持该国先进半导体企业 Rapidus 。 考虑到 Rapidus 将其 2nm …
3
Fair ASR: Re-Evaluating Black-Box Jailbreaks under Shared Target-Call Budgets
挑战黑盒越狱评估的公平性，引入共享调用预算新框架，重新定义攻击成功率，值得AI安全研究者细读。
arXiv:2608.17360v1 Announce Type: cross Abstract: Reliable jailbreak evaluation is essential for assessing LLM safety, but most existing studies rely …
4
Why HTTP 429 Should Not Count Against a Job's Retry Budget
别再让限流响应耗尽重试次数，把429当延迟而不是失败，任务系统可靠性的关键设计。
Most retry implementations—including BullMQ, custom wrappers, and managed queues—treat every non-2xx HTTP response as an identical failure. Each attem…
5
Used EV buying guide: The $30,000 edition
三万美元预算淘二手电动车？哪些车型选择上百台、哪些只有孤品，一篇讲透选购门道
Depending on the model, you might have hundreds to choose from—or just one.
6
ComboShoppingBench: Evaluating LLM Agents for Budget-Constrained Basket Shopping with Coupons
LLM智能体在预算约束与优惠券组合购物中表现如何？全新基准ComboShoppingBench给出量化答案
arXiv:2608.09282v1 Announce Type: new Abstract: Real-world shopping often requires constructing a basket of complementary items rather than retrieving…
7
EcoAgent-Bench: Evaluating Economic Decision-Making in Budget-Constrained LLM Agents
评估LLM代理在预算约束下的经济决策，揭示资源效率与任务完成度的平衡新基准。
arXiv:2608.05519v1 Announce Type: new Abstract: Agent benchmarks usually measure task completion and treat resource use as an auxiliary statistic. In …
8
Adaptive Finite-Budget Training for CVaR Risk-Aware Q-Learning
用有限预算自适应训练风险敏感的Q学习，兼顾收益与尾部风险，金融决策新解法。
arXiv:2608.04305v1 Announce Type: new Abstract: Risk-aware Q-learning (RaQL) provides a model-free, two-timescale estimator for dynamic risk objective…
9
Cloudflare Wallet 推出：专为 AI 智能体设计的可编程钱包
为AI智能体打造的可编程钱包，精细控制预算与支付，让智能体独立完成付费任务。
IT之家 8 月 5 日消息，Cloudflare 昨日（8 月 4 日）发布公告，宣布推出 Cloudflare Wallet 功能， 针对 AI 智能体部署数字信用卡、身份证明（ID）以及消费上限机制。 博文指出当前 AI 智能体如果想要试用某个 API，必须先访问专为人类设计的登录页面，等待用…
10
AI coding agents are blowing through budgets — Replit, Kilo Code, and Symbotic explain how they're managing it
AI编码代理烧钱太快？看Replit、Kilo Code和Symbotic如何破解预算失控难题，开发者必读。
At Kilo Code, engineers are reading or writing code themselves only about 1% of the time now, according to co-founder Emilie Schario — the rest is age…
11
12
亚马逊 AI 超支五个月才被发现，单个项目花费 180 万美元超预算 860%
亚马逊AI项目超支860%五个月才被发现，企业部署AI的成本管控漏洞值得警惕。
IT之家 8 月 4 日消息，据英国《金融时报》，亚马逊内部已发现多起因 AI 部署失误和成本管控缺失而导致的“灾难性”预算超支案例。 在 7 月 29 日的一场员工会议上，亚马逊高级工程师向同事指出，在尝试将部分传统编程任务转交给 AI 处理时，出现了此前未预料到的“计划外支出”，且此类超支并非孤…
13
Flat Score, Amplified Failures: How the Error Budget Masks Damage in Quantized LLM Agents
4-bit量化看似无损，但工具调用型Agent的多轮任务中错误被悄悄放大，一份扎实的基准评测揭穿幻觉。
arXiv:2607.27275v1 Announce Type: new Abstract: Post-training quantization to 4-bit weights is widely reported to be nearly lossless. We test this cla…
14
OmniDelta: Skill-Driven Budget Allocation for Token Compression in OmniLLMs
针对全模态大模型，提出基于技能驱动的token压缩预算分配方法OmniDelta，提升效率保留关键信息。
arXiv:2607.25669v1 Announce Type: new Abstract: Emerging Omni-modal Large Language Models (OmniLLMs) enable unified understanding of text, audio, and …
15
WISERouter: LLM Routing with Workload Budget Constraint
新论文提出WISERouter，在预算约束下智能路由LLM请求，兼顾成本与性能。
arXiv:2607.23765v1 Announce Type: new Abstract: Large language models (LLMs) achieve impressive performance across multiple domains, but using the mos…
16
Automate 7 Projects Marketing with AI for $40
用AI自动化7个项目的市场营销，每月仅40美元，Claude+Codex双助手方案详解
Article URL: https://apsquared.co/posts/marketing-automation-claude-codex Comments URL: https://news.ycombinator.com/item?id=49061784 Points: 1 # Comm…
17
Show HN: LLM-budget-cap – atomic Redis spend cap for LLM APIs
基于Redis的原子性花费上限工具，防止AI API预算超支，支持毫秒级灵活窗口。
Article URL: https://github.com/Rentheria/llm-budget-cap Comments URL: https://news.ycombinator.com/item?id=49027830 Points: 1 # Comments: 0
18
From DevOps to SRE: Making the Transition
从DevOps转向SRE六年，真实分享思维本质差异：从“如何更快交付”到“如何稳定地更快交付”。
I moved from a DevOps title to an SRE title about 6 years ago. On paper, they look similar. In practice, the mindset is different. Here's what actuall…
19
An LLM API gateway with budget-limited tokens for your LLM API keys
为LLM API密钥设置预算限制令牌的网关，帮你控制调用成本与用量。
Article URL: https://github.com/er91/budget_aware_llm_api_gateway Comments URL: https://news.ycombinator.com/item?id=48974910 Points: 1 # Comments: 1
20
Budgeted Subset Refinement for Execution-Aware LLM Research Ideation
LLM研究想法生成新方法：预算子集细化让创意更易落地执行
arXiv:2607.14118v1 Announce Type: new Abstract: Large language models (LLMs) can generate research ideas that appear novel to expert reviewers, but re…