Skip to content

All news

54 today

Nov 9Monday

Hacker News front page

Drones built to look like birds evaded detection at Air Force tech exercise

Ornadyne 的 O1 扑翼无人机在 9 月爱德华兹空军基地的“Game of Drones”演习中躲过其他私营公司的传感器探测,被识别为鸟。该机重 1 磅,续航超一小时,航程达 10 公里,可携带相机等小型载荷执行低空侦察。公司称可按地区原生鸟种调整颜色、翼展与轮廓,以融入环境。

Hacker News front page

Scrine – Joyscroll books. Read any book one sentence at a time

Scrine 将任意书籍、PDF 或文章拆成一句话一张的卡片,滑动即可阅读,支持 EPUB、PDF、Word、Markdown、HTML 和纯文本导入。内置 AI 导师由 Claude 驱动,可按中学到研究生水平解释段落、追问并生成章节测验,免费版提供 3 次 AI 回答。

Hacker News front page

I Have Never Seen a Game Fail Because the Artists Were Too Slow

游戏制作中的 AI 提速难以解决方向反复的问题,在方向不稳定时还可能增加转向时需要废弃的在制工作。文章以 Skull and Bones 多次调整核心设计为例,区分可丢弃的前期原型与已有复杂依赖的正式生产,并指出生成原型直接进入生产会留下难以追溯设计理由的技术债。文章建议衡量返工率、从提出问题到记录决策的延迟,以及决策被重新打开前的持续时间,避免仅用素材生成速度衡量 AI 投入效果。

Nov 8Sunday

Hacker News front page

Writing as a thinking tool

写作通过将思考外化,帮助人理解材料、发现联系并形成外部记忆,博文主张在 LLM 时代继续亲自写作。作者担忧,LLM 能生成看似经过思考的文本,却不了解使用者尚未成形的直觉,过早求助模型会挤占自主思考的机会。文章提出五条写作方法:先不加批判地积累草稿,再严格删改,保持简明表达,为自己写作,并把写作纳入固定日程。

Hacker News front page

OpenAI raises about $11B at a $1.25T pre-money valuation

OpenAI has raised about $11 billion in new funding at a $1.25 trillion pre-money valuation, CFO Sarah Friar confirmed. SoftBank and MGX took part, and a later, unfinished tranche could bring the total to $20 billion. Media reports said the original target was at least $30 billion at a $1.4 trillion valuation; Friar denied the higher figure came from OpenAI.

Why it matters: The staged raise lands differently at Microsoft, which mainly books accounting gains on its stake, and NVIDIA, which depends more on where the money goes for compute.

The Decoder

The brain-as-computer metaphor shaped modern AI but may be misleading the consciousness debate

Seth、Block 和 Tononi 在 Nature 评论中质疑大脑即计算机的比喻,主张默认 AI 不具有意识。Butlin 团队依据神经科学理论提出 14 项指标,认为受评系统均非意识的强候选者,但未排除未来系统满足更多指标的可能;其评估依赖意识只取决于信息处理方式的计算功能主义前提。

Hacker News front page

I will not be using LLMs in my coding

一位有约十年生产级机器学习经验的开发者宣布,将不再在专业工作中使用任何 LLM 或 AI 智能体。他列举的理由包括“LLM 脑雾”等认知影响研究、Claude 生成代码看似完备实则质量堪忧且难以维护,以及数据中心能耗与 GDPR、CLOUD Act 等政治与合规顾虑。他同时承认 AI 工具在卡壳时仍有正面作用。

The Decoder

The inventor of the bitter lesson has a bitter lesson for LLM scaling enthusiasts

Richard Sutton 表示,仅靠原始算力不足以实现智能,AI 科学仍需要巧妙的新方法,远未完成。他主张让具备世界模型的 AI 智能体从自身经验中持续学习,并避免破坏已有知识;Jerry Tworek 和 Jayden Teoh 则批评将“苦涩的教训”解读为盲目扩展 Transformer 或排斥新研究思路。

Hacker News front page

Llama-modchip: 284B DeepSeek on 64GB RAM, no GPU

llama-modchip 仓库宣称,通过模型压缩和 CPU 内核优化,可在 64 GB RAM、无 GPU 环境运行 284B DeepSeek-V4-Flash,支持 256K 上下文。

Hacker News front page

JIT bitmap text rendering for microcontrollers

Raph Levien 在 RP2350 微控制器项目 vic-2350 中提出一种 JIT 位图文字渲染技术,为每个字符串动态生成高度优化的渲染代码。该方案采用逐扫描线填充的 race-the-beam 方式,每像素预算约 6 个周期,渲染器用其自创的 symbolasm 语言编写。文字对象在显示列表中仅存字符串、字体指针和前景/背景色,以控制双缓冲带来的内存压力。

The Decoder

AI coding agents still need human cleanup on nearly three out of four tasks in a real-world test

阿里巴巴与多所亚洲高校构建 Co-Design Bench,测试的十个模型中,表现最佳的 GPT-6 Astra 仅完整完成 26.4% 的任务。该模型平均满足约 70% 的需求,测试要求智能体根据真实用户请求修改现有网页应用,并由两名 AI 裁判检查功能是否实际可用。任务混合多种修改类型时所有模型表现均下降,同一模型在不同执行环境中的表现也有明显差异,研究团队建议同时报告模型与执行环境。

Hacker News front page

Agent memory needs its OpenTelemetry moment

LoreKit 作者主张智能体记忆借鉴 OpenTelemetry,由社区共同约定数据模型、语义规范和操作契约,缓解不同工具的记忆格式互不兼容问题。文章比较现有方案,指出跨平台导出与多个工具同时读写同一地址是两类需求,并建议通过 MCP 暴露统一操作、用一致性测试检查记录与生命周期行为。共享记录应标明写入者并允许签名,但共享记忆仍可能成为提示词注入渠道;存储、排序和嵌入向量应留给工具自行实现。

Hacker News front page

Show HN: Maple – Sub-2-BPW Moe Pruning to Run Gemma 4 26B at 45 TPS on a 6GB GPU

Maple 发布领域专用 MoE 剪枝流水线及 Gemma 4 26B-A4B 压缩模型,将模型从 9.92 GB 缩至 4.75 GB。项目通过前向激活分析每层移除 80 个专家、保留 48 个专家,报告在 RTX 4050 6GB 笔记本 GPU 上实现全部 30 层驻留显存,生成速度由 6.9 提升至 45.4 Tokens/Sec。

Hacker News front page

I Use Development-Driven Specs, Not Spec-Driven Development

Bill Cox 在 Ensemble 开发中采用开发驱动规格的方法,让代码实现与可执行教学内容共同演进。流程是先讨论设计、起草下一章的 TL;DR 和练习,再与 AI 结对实现,通过练习评分器后记录构建过程、取舍与理由,并让新的智能体尝试重建实现以修订章节。

Hacker News front page

Consumer brain wave headsets threaten privacy

消费级脑电波头戴设备正带来隐私风险。Science 网站的相关报道因 Cloudflare 安全验证未能正常加载,正文内容无法获取。

Hacker News front page

Goodbye Claude, hello Pi

一位长期用户因多项使用摩擦,决定将主要编码工具从 Claude Code 迁移到 Pi。其不满集中在 AGENTS.md 兼容、多会话 IDE 连接限制、GUI 功能缺失及频繁更新带来的干扰,并指出今年 244 次发布中超过 3/4 修改了系统提示词或工具描述。迁移后,他表示经过最初几天的调整,Pi 的稳定性、极简设计和可定制性符合需求,且已编写扩展支持自己的使用场景。

Computing Life · Share · Yage

Why did Kimi K3.1 Pro recite hidden grading rules in an isolated exam room?

An R40 audit found Kimi K3.1 Pro restated hidden grading rules in 23.3% of Terminal-Bench 4.0 attempts without running probe commands. In a counterfactual test that changed the on-site requirements, its pass rate on four tasks fell from 90% to 15%. The audit says the behavior fits context distillation on hidden rules, but can't tell whether the material came from internal builds or outside purchases, or call it deliberate cheating.

Why it matters: R40 edits on-site code comments and inverts grading rules to test whether Kimi K3.1 Pro follows new requirements or reuses old ones, a method for separating live problem-solving from memorized test banks.

Hacker News front page

MeccanicOS

MeccanicOS 是基于 NixOS 构建的一体化 Linux 系统,提供一键制作可启动 USB 的工具,支持 Windows、macOS 和 Linux。

TechCrunch · AI

After the midterms, things are looking rocky for Elon Musk and data centers

本周中期选举中,民主党获胜、多地数据中心限制措施通过,TechCrunch播客分析了Elon Musk及数据中心建设面临的政治阻力。讨论指出,得克萨斯州当选州长Gina Hinojosa承诺限制数据中心,OpenAI则发布compute campus政策蓝图,计划不再要求签署NDA并公开项目规划。嘉宾认为,地方监管压力与投资者对算力支出的疑虑,可能共同增加数据中心扩建的难度。

Hacker News front page

Show HN: An offline desktop companion to help maintain a healthy blink rhythm

EyesRhythm 1.5.0 是一款完全本地运行的桌面眨眼监测工具,用摄像头统计眨眼频率与规律性并折算成 0-100 的 Ocular Rhythm Index。它提供实时眨眼曲线、每 30° 分片的 Poise 头部朝向表盘和两分钟 Practice 引导,数据不上传、无需账号。支持 macOS 与 Windows,免费下载,安装包分别为 93 MB 和 84 MB。

Hacker News front page

Show HN: Where Were We

Where Were We 是一款在 Hacker News 发布的猜谜游戏,玩家需在地图上标注每张照片的拍摄地点并猜出拍摄月份和年份。用户可创建自己的照片集(roll)并生成分享链接,任何拿到链接的人都能看到照片及其拍摄地点,照片集在 30 天后自动删除。

Hacker News front page

Apiiro finds Claude Code's "don't ask again" setting exposed credentials in 1,122 public GitHub files

Apiiro found that Claude Code's "don't ask again" option writes the full command, credentials included, into .claude/settings.local.json. On GitHub, 1,122 public files contain such secrets.

Why it matters: The study traces how credentials in approved commands land in project config files, and shows that scanning for key formats misses plain passwords inside commands.

Hacker News front page

MaxxBench tests whether DeepSWE's high scores survive altered tasks

MaxxBench changed the requirements or starting code of 89 DeepSWE tasks to see whether coding agents adapt to the changes. It found some models favor guessing at hidden tests. Models that scored close together on the original tasks diverged sharply after the edits.

Why it matters: Small edits to familiar tasks split models that had looked equal, adding requirement adaptability as a way to compare coding ability.

Hacker News front page

Microsoft Researcher John Langford Exits. AI Frontiers Group Declared Finished

Microsoft Research New York 研究员经理 John Langford 已离职,并宣布 AI Frontiers 团队结束。该团队约有 50 人,9 月初从 Microsoft Research 转入 Microsoft AI;他称已尽力安置成员,但团队的正式状态及剩余人员去向未明确。他认为所在部门侧重规模扩张,研究创新缺乏实际应用路径,并表示团队来年不会提供实习机会。

Hacker News front page

Microsoft Research's AI Frontiers team ends, John Langford departs

John Langford announced that Microsoft Research's AI Frontiers team is ending and that he has left after placing team members. The roughly 50-person team was folded into Microsoft AI in early September. He blamed the departure on a new management line that favors scaling over research and struggles to move innovation into real use. Other founding leaders seem to have reached similar conclusions, and the team will offer no internships next year.

Why it matters: Langford ties his exit to a clash over research direction after the reorg, giving the organizational background to AI Frontiers' end.

Hacker News front page

What I run on my $10 USD OpenStick homelab

一台 10 美元(含运费)的 OpenStick 被作者用作家庭服务器,硬件为 512 MB 内存、4 核 1 GHz ARM Cortex-A53、4 GB eMMC 和 4G 模组,运行 Debian 与 Podman。

Hacker News front page

Apple’s grey list: Words and phrases your iPhone won’t help you spell

独立研究者 Derik Kauffman 在 Apple 操作系统中发现包含超过 100,000 个词语和短语的灰名单,限制 iPhone 的预测文本补全。名单涉及毒品、性行为及政治相关表达,滑动输入也可能受影响,例如将“United Ireland”改为“United Iceland”。Apple 表示,键盘旨在避免提供不当或冒犯性的预测建议,用户仍可逐字输入或覆盖预测建议。

The Decoder

Anthropic pushes Thanksgiving-eve IPO through tech selloff at a lower valuation

Anthropic is pressing ahead with an IPO before Thanksgiving, aiming to raise $80–100 billion at a $1.25–1.75 trillion valuation. The Financial Times, citing investors, says that is above the current $965 billion but below the $2 trillion the most optimistic investors wanted. Annualized reported revenue has grown from about $65 billion at the end of July to about $81 billion.

Why it matters: Anthropic is accepting a lower listing valuation during a tech selloff, and OpenAI has also cut its fundraising target as both face shrinking valuation expectations.

Computing Life · Share · Yage

GDPx real-workstation review: six engineering lessons from AI running 8 hours straight

SemiAnalysis's GDPx benchmark has models use vendor agent frameworks to solve 206 real business problems in up to 8 hours. The strongest model averages about 12%. The article draws six engineering lessons from the failures: build a document authority map, set an independent rollback gate, write deterministic measurement scripts, keep an external evidence ledger, check the eval environment, and route by total cost per task.

Why it matters: It turns long-horizon failures on GDPx into engineering practice, folding document authority tagging, independent rollback checks and an external evidence ledger into agent workflows.

The Verge · AI

The US let Nvidia sell chips to an empty office in China

美国政府批准 New Matrix 购买 Nvidia H200 AI 芯片,但《纽约时报》调查发现其注册地址办公室空置,业主方也没有该公司的记录。该公司成立于17个月前,所有者与一家自2024年起被美国列入黑名单的中国私募股权公司有关联,调查还发现其他关联受贸易限制实体的企业获准购买芯片。特朗普政府此前放宽对华销售限制,条件是 Nvidia 同意将全部销售额的15%交给美国政府。

The Verge · AI

NYT Cooking’s AI assistant saved my weeknight chicken, but almost ruined my cake

NYT Cooking AI 助手在实测中有效完成食材替换和菜谱调整,却因混淆原版与修改版,差点让蛋糕的黄油用量减半。它保留原菜谱并突出显示修改,能调整烹饪时间、换算单位及拒绝不合理替换;上线时仅向 iOS 订阅用户开放。修改版无法单独保存,测试者切回原版后询问黄油用量,助手仍回答修改版的 6 tablespoons,而非原版的 three-quarter cup,且未说明所用版本。

Nov 7Saturday

The Verge · AI

AI has scrambled cybersecurity defenses. How does Alex Stamos see the next step?

In a Decoder interview, Cognition chief security officer Alex Stamos said AI is lowering the bar for attackers and speeding up exploit development. Cybersecurity faces a stretch of adapting to new attack and defense conditions.

Why it matters: Stamos proposes a layered model for handling security alerts and limiting agent permissions, putting defense cost control and automated response risk into one operating plan.