GrapheneOS – break free from Google and Apple [ENG 🇬🇧] .. 🇬🇧->🇵🇱 Przejdź do polskiej wersji tego wpisu / Go to polish version of this post Table of contents: Introduction (feel free to skip) What is GrapheneOS Dedicated devices List of supported devices (February 2026) My smartphone choice Installing GrapheneOS We need Phone preparation Unlocking the bootloader Downloading and flashing the system image Re-locking the bootloader Restoring the OEM lock My vision of using GrapheneOS Additional user profile Obtainium List of my open-source applications Aurora Store List of apps whose functionality without GMS I have verified Full control over app permissions Private space List of apps in my private space Summary Supporting the GrapheneOS project Introduction (feel free to skip) Just a year ago, I was really deep into the Apple ecosystem. It seemed like there was no turning back from the orchard for me. Phone, laptop, watch, tablet, video and music streaming, cloud storage, and even a key tracker. All from one manufacturer. Plus shared family photo albums, calendars, and even shopping lists. However, at some point, I discovered Plenti , a company that rents a really wide range of different devices at quite reasonable prices. Casually, I threw the phrase “samsung fold” into the search engine on their website and it turned out that the Samsung Galaxy Z Fold 6 could be rented for just 250-300 PLN per month. That was quite an interesting option, as I was insanely curious about how it is to live with a foldable phone, which after unfolding becomes the equivalent of a tablet. Plus, I would never dare to buy this type of device, because firstly, their price is astronomical, and secondly, I have serious doubts about the longevity of the folding screen. I checked the rental conditions from Plenti and nothing raised my suspicions. Renting seemed like a really cool option, so I decided to get the Fold 6 for half a year. That’s how I broke out of the orchard and slightly reopened the doors to my heart for solutions without the apple logo. I even wrote a post about the whole process – I betrayed #TeamApple for broken phone . What I’m getting at is that this is how Android returned to my living room and I think I started liking it anew. My adventure with Samsung ended after the planned 6 months. The Galaxy Z Fold 6 is a good phone, and the ability to unfold it to the size of a tablet is an amazing feature. However, what bothered me about it was: after folding it was terribly thick, it couldn’t be used in a case, because all the covers either didn’t fit or slipped off the part that has screens on both sides, naked, it was very blocky, even sharp, which caused me discomfort; holding it was simply unpleasant, paying 300 PLN (~80 USD) for rent is a good short-term solution to get something to test, but not in the long run. All the points above made me give up on extending the rental and start wondering what to do next. Interestingly, I liked Android enough that I didn’

Source: Hacker News | Original Link

Is Show HN Dead? No, But It’s Drowning ← Back Is Show HN Dead? No, But It’s Drowning Feb 17, 2026 · Arthur Cnops A few days ago I posted to Show HN . I had good fun building that useless little internet experience. The post quickly disappeared from Show HN’s first page, amongst the rest of the vibecoded pulp. And to be clear, I’m fine with that. The behavior on Show HN was interesting to see though. So I pulled the data. The Short Version Show HN of course isn’t dead. You could even say it’s more alive than ever. What has changed is the volume of posts and engagement per post. It’s only natural when more projects are being built in a single weekend. There’s less “Proof of Work”. From the business side of this, Johan Halse recently called this the Sideprocalypse : the end of the small indie developer’s dream. Every idea has been built, marketed better, and SEO’d into oblivion by someone with more money. Some cool projects aren’t getting through this noise, which is a pity. Here are a few I thought were interesting: Neohabit OpenRun uForwarder I just upvoted them! Now, let’s look at some data. Volume Is Exploding The Graveyard Is Growing Show HN started out better than regular submissions. Now it’s significantly worse . The Shrinking Window How long does a Show HN post stay on page 1 before being pushed off? During peak hours (US daytime): Discussion Is Dying Too So Is Show HN Dead? No. There’s just more noise, and less opportunity to get attention and have a discussion with other folks on HN about your project. Some gems go completely unnoticed. Maybe something for HN to think about: how do these subjective “gems” get more spotlight? How does HN remain the coolest place to talk about the coolest tech?

Source: Hacker News | Original Link

GrapheneOS – break free from Android and iOS [ENG 🇬🇧] .. 🇬🇧->🇵🇱 Przejdź do polskiej wersji tego wpisu / Go to polish version of this post Table of contents: Introduction (feel free to skip) What is GrapheneOS Dedicated devices List of supported devices (February 2026) My smartphone choice Installing GrapheneOS We need Phone preparation Unlocking the bootloader Downloading and flashing the system image Re-locking the bootloader Restoring the OEM lock My vision of using GrapheneOS Additional user profile Obtainium List of my open-source applications Aurora Store List of apps whose functionality without GMS I have verified Full control over app permissions Private space List of apps in my private space Summary Supporting the GrapheneOS project Introduction (feel free to skip) Just a year ago, I was really deep into the Apple ecosystem. It seemed like there was no turning back from the orchard for me. Phone, laptop, watch, tablet, video and music streaming, cloud storage, and even a key tracker. All from one manufacturer. Plus shared family photo albums, calendars, and even shopping lists. However, at some point, I discovered Plenti , a company that rents a really wide range of different devices at quite reasonable prices. Casually, I threw the phrase “samsung fold” into the search engine on their website and it turned out that the Samsung Galaxy Z Fold 6 could be rented for just 250-300 PLN per month. That was quite an interesting option, as I was insanely curious about how it is to live with a foldable phone, which after unfolding becomes the equivalent of a tablet. Plus, I would never dare to buy this type of device, because firstly, their price is astronomical, and secondly, I have serious doubts about the longevity of the folding screen. I checked the rental conditions from Plenti and nothing raised my suspicions. Renting seemed like a really cool option, so I decided to get the Fold 6 for half a year. That’s how I broke out of the orchard and slightly reopened the doors to my heart for solutions without the apple logo. I even wrote a post about the whole process – I betrayed #TeamApple for broken phone . What I’m getting at is that this is how Android returned to my living room and I think I started liking it anew. My adventure with Samsung ended after the planned 6 months. The Galaxy Z Fold 6 is a good phone, and the ability to unfold it to the size of a tablet is an amazing feature. However, what bothered me about it was: after folding it was terribly thick, it couldn’t be used in a case, because all the covers either didn’t fit or slipped off the part that has screens on both sides, naked, it was very blocky, even sharp, which caused me discomfort; holding it was simply unpleasant, paying 300 PLN (~80 USD) for rent is a good short-term solution to get something to test, but not in the long run. All the points above made me give up on extending the rental and start wondering what to do next. Interestingly, I liked Android enough that I didn’t

Source: Hacker News | Original Link

WD and Seagate confirm: Hard drives for 2026 sold out | heise online heise+ entdecken Suchen Abo Suchen Alle Magazine im Browser lesen Anzeige Special: Zusammen das Datacenter weiterentwickeln Secure IT für Unternehmen Newsletter heise-Bot Push -Nachrichten Anzeige Special: Zusammen das Datacenter weiterentwickeln Secure IT für Unternehmen Advertisement Advertisement With Western Digital and Seagate, two of the three remaining hard drive manufacturers worldwide have confirmed that they have already sold their production quotas for 2026 completely or almost completely. The situation is likely to be similar for Toshiba. Continue after ad The statements come from the company CEOs in the analyst conferences on the latest business reports. They thus confirm speculation from September 2025 . And there are already initial agreements for the year 2026. They largely come from hyperscalers who want hard drives for their AI data centers, for example to store training data on them. These include Amazon (AWS), Google, Microsoft (Azure), Meta, and OpenAI. What the CEOs say Western Digital CEO Tiang Yew Tan said : “We’re pretty much sold out for calendar year 2026. We have firm purchase orders with our top seven customers through calendar year 2026. We also have in place robust commercial agreements with three of our top five customers, two through calendar year 2027 and one through calendar year 2028.” Seagate CEO William Mosley explained : “Our nearline capacity is fully allocated through calendar year 2026, and we expect to begin accepting orders for the first half of calendar year 2027 in the coming months. Further out, demand visibility is strengthening based on the long-term agreements in place with major cloud customers through calendar 2027. Additionally, multiple cloud customers are discussing their demand growth projections for calendar 2028, underscoring that supply assurance remains their highest priority.” By nearline, Seagate understands its HDD offering with a server focus. Some of these models also spill over into retail. According to the business report , nearline recently accounted for 87 percent of Seagate’s hard drive sales. The division is increasingly displacing the remaining offerings. A year earlier, nearline was still at 83 percent. Videos by heise mehr Videos c’t 3003 heise & ct Peertube No Production Expansion for Now Continue after ad According to Mosley, Seagate is not expanding its production capacities for now. Growth is to come only from higher-capacity hard drives, not from additional unit numbers. CFO Gianluca Romano added: “If in a quarter, we can produce a little bit more, of course, we will sell those exabytes in the open market at a good profitability. But I would say the vast, vast majority of the volume is already allocated.” Toshiba has not been listed on the stock exchange since 2023. As a result, only rudimentary financial information is publicly available. The company no longer provides outlooks like its competitors.

Source: Hacker News | Original Link

今年春晚机器人“浓度”超高，总导演于蕾回应_腾讯新闻 今年春晚机器人“浓度”超高，总导演于蕾回应 政事儿 2026-02-17 10:30 发布于 北京 新京报政事儿官方账号 2月16日晚，2026年中央广播电视总台春节联欢晚会开播。宇树科技、松延动力、魔法原子、银河通用等公司的 具身智能机器人 在春晚舞台高频曝光，冲上微博热搜。 据看看新闻Knews微博消息， 总台马年春晚总导演于蕾说，春晚直播结束，主创团队就像考试结束，接下来就是等待观众“阅卷”。大家很享受筹备和直播的过程，会为零点准时倒计时欢呼，也会为了虚拟团队呈现欢呼。 今年的春晚机器人“浓度”超高，希望通过节目让观众看到中国机器人产业立体化的状态。 2026年春晚总导演于蕾此前在采访中表示，春晚节目将让观众感受到科技走进生活，展示当下的科创时代，多家具身智能机器人分别以语言类演出、歌曲演出、武术演出、微电影演出等形式亮相春晚舞台。此外，春晚视觉团队 结合AI技术尝试突破创新，所有视觉变化都在央视一号演播厅内完成，给观众带来新体验。 机器人在春晚舞台上再次“出圈” 春晚舞台成为具身智能集中展示的窗口。 2025年春晚，宇树科技机器人因参与《秧BOT》表演产生“现象级破圈 ” 。今年春晚，宇树科技机器人表演武术节目《武BOT》，与河南塔沟武术学校的孩子们共同表演武术，呈现了连续翻筋斗、跳马、后空翻等动作，并和小朋友们“比武”，展示醉拳、双节棍等武术技能，宇树科技身高1.8米的机器人Unitree H2登台亮相，表演舞剑。 宇树科技创始人王兴兴在采访中表示，今年机器人在快速奔跑中完成了穿插变阵和武术动作。这种高动态、高协同的 集群控制技术 是首次亮相，“这个动作非常实用，为后续机器人在其他场景集群或单台机器人调度做好了铺垫”。 在小品《奶奶的最爱》节目中，具身智能企业松延动力的所有机器人产品全部亮相，包括N系列、E系列、小布米与 仿生人形机器人 。春晚舞台上，小布米等人形机器人展示了后空翻、侧手翻、舞蹈等技能，仿生机器人则还原了“机器人 蔡明 ”，与扮演孙子的 王天放 “插科打诨”。 三十年前，央视春晚舞台上的小品《机器人趣话》中，那个淳朴、热情、奔放的机器人“菜花”第一次在观众的脑海中构建了对人形机器人的想象。三十年后，在2026年总台春晚的现场，当年“菜花”的扮演者蔡明，真正与机器人同台。 松延动力CMO张淼表示，这是春晚历史上人形机器人第一次走进语言类小品，“多家机器人企业选择上春晚，是行业面向商业化关键节点的默契选择。机器人也丰富了春晚节目形式，展示我国科技实力。”张淼谈到，希望借助春晚演出，推动企业商业化进程。 此外，在歌曲表演《智造未来》节目中，魔法原子机器人给演员伴舞。同时，歌曲《智造未来》将 VR眼镜 、AI、机器人、交互手表等前沿科技元素融入歌词中。银河通用参与贺岁微电影《我最难忘的今宵》的演出。 机器人密集登陆各大卫视 除了总台春晚，各大卫视春晚也引入了机器人表演。例如，智元机器人亮相河南卫视，众擎机器人亮相深圳卫视，光谷东智机器人亮相湖北卫视。 近日，智元机器人陆续登陆河南卫视等多家平台，并针对不同平台的调性定制了差异化的内容。比如，2月14日，智元机器人登陆河南卫视春晚，带来节目《功夫机缘》，灵犀X2在嵩山少林寺演绎了传统武术。 智元机器人相关负责人表示，希望通过春晚演出，展示科技可以有温度，机器人是可以成为伙伴的，“它们不是一堆代码的集合，而是能在舞台上撒娇、能逗人笑、能让人感觉亲切的生动个体。对产业来说，这是具身智能从B端走向C端认知的一个标志。” 综合来源：中国证券报

Source: Tencent News | Original Link

俄美乌第三轮会谈今天举行，将讨论领土问题，俄乌代表或进行双边接触_腾讯新闻 俄美乌第三轮会谈今天举行，将讨论领土问题，俄乌代表或进行双边接触 九派新闻 2026-02-17 11:14 发布于 湖北 武汉晨报九派新闻官方账号 据@CCTV国际时讯，今天（2月17日），俄美乌三方第三轮会谈将在瑞士日内瓦举行。按照瑞士方面的通报，本次会谈闭门举行，媒体不得入内。俄罗斯总统新闻秘书佩斯科夫2月16日表示，本轮会谈将讨论更广泛的问题，包括领土问题。还有消息称，会谈期间，俄乌代表有可能进行双边接触。 会谈前夕，俄罗斯外交部发言人扎哈罗娃在接受总台记者专访时表示，俄方始终主张通过和平方式解决当前问题，而谈判必须考虑到各方的利益。扎哈罗娃强调，谈判必须要达成结果，而不是成为某种挑衅行为的掩护。 据新华国际，美国总统特朗普16日施压乌克兰“快速”同俄罗斯达成协议。 特朗普当天在总统专机“空军一号”上回答记者提问时说，17日在日内瓦举行的新一轮美俄乌会谈很重要，也将“非常容易”。 特朗普特别点名乌克兰：“乌克兰最好快速回到谈判桌前……我们已就位，希望他们来。” 新一轮美俄乌会谈将于17日在瑞士日内瓦举行。俄罗斯总统新闻秘书佩斯科夫16日说，会谈将讨论更广泛的问题，其中包括领土问题。 俄罗斯外交界消息人士16日深夜称，俄总统助理梅金斯基率领的俄方代表团已经出发，预计17日清晨抵达瑞士日内瓦，参加俄美乌三方会谈。会谈期间，俄乌可能进行双边接触。 消息人士说，俄罗斯代表团将于日内瓦当地时间17日6时左右抵达。美方在技术层面参与协调俄代表团经由欧盟地区抵达日内瓦的相关许可事宜。 消息人士说，本次三方会谈计划讨论解决乌克兰问题的关键因素，包括军事、政治和人道主义等领域。会谈将持续多久目前不得而知。在俄美乌会谈期间，俄罗斯与乌克兰有可能进行双边接触。 △俄罗斯总统新闻秘书佩斯科夫（资料图） 俄罗斯总统新闻秘书佩斯科夫16日在莫斯科说，本次三方日内瓦会谈将讨论更广泛的问题，其中包括领土问题。 今年以来，俄美乌三方在阿联酋首都阿布扎比举行两轮会谈，讨论俄乌军队脱离接触的可能性、停火监督机制和设立协调中心等，但关键的领土问题未得到解决。 来源：综合新华国际、央视新闻 编辑：凌山 审核：宣萱 声明：此文版权归原作者所有，若有来源错误或者侵犯您的合法权益，您可通过邮箱与我们取得联系，我们将及时进行处理。邮箱地址：[email protected]

Source: Tencent News | Original Link

南部战区新闻发言人发表谈话_腾讯新闻 南部战区新闻发言人发表谈话 南部战区 2026-02-17 10:46 发布于 广东 中国人民解放军南部战区官方账号 南部战区新闻发言人翟士臣海军大校表示，2月15日至16日，中国人民解放军南部战区组织海空兵力位南海海域进行战备警巡。菲律宾拉拢域外国家组织所谓“联合巡航”，搅局南海，破坏地区和平稳定。战区部队将坚决捍卫国家领土主权和海洋权益，坚定维护地区和平稳定。 Senior Captain Zhai Shichen, spokesperson for the PLA Southern Theater Command stated，from February 15 to 16, the naval and air forces dispatched by the PLA Southern Theater Command conducted combat readiness patrols in the South China Sea. The Philippines co-opted countries outside the region to organize the so-called joint patrols, disrupting peace and stability in the region. Forces of PLA Southern Theater Command will resolutely safeguard China’s territorial sovereignty and maritime rights and interests, and firmly uphold regional peace and stability. 南部战区发布 主编：汪尚建  编审 ：蔡梦华  责编：宋旭东 刊期： 第 5150 期

Source: Tencent News | Original Link

GitHub – matt1398/claude-devtools Skip to content You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert matt1398 / claude-devtools Public Notifications You must be signed in to change notification settings Fork 47 Star 370 License MIT license 370 stars 47 forks Branches Tags Activity Star Notifications You must be signed in to change notification settings matt1398/claude-devtools main Branches Tags Go to file Code Open more actions menu Folders and files Name Name Last commit message Last commit date Latest commit History 110 Commits 110 Commits .claude .claude .github/ workflows .github/ workflows build build public public resources resources scripts scripts src src test test .dockerignore .dockerignore .editorconfig .editorconfig .gitattributes .gitattributes .gitignore .gitignore .nvmrc .nvmrc .prettierignore .prettierignore .prettierrc.json .prettierrc.json CHANGELOG.md CHANGELOG.md CLAUDE.md CLAUDE.md CODE_OF_CONDUCT.md CODE_OF_CONDUCT.md CONTRIBUTING.md CONTRIBUTING.md Dockerfile Dockerfile LICENSE LICENSE README.md README.md SECURITY.md SECURITY.md docker-compose.yml docker-compose.yml electron.vite.config.ts electron.vite.config.ts eslint.config.js eslint.config.js knip.json knip.json package.json package.json pnpm-lock.yaml pnpm-lock.yaml pnpm-workspace.yaml pnpm-workspace.yaml postcss.config.cjs postcss.config.cjs tailwind.config.js tailwind.config.js tsconfig.json tsconfig.json tsconfig.node.json tsconfig.node.json tsconfig.test.json tsconfig.test.json vite.standalone.config.ts vite.standalone.config.ts vitest.config.ts vitest.config.ts vitest.critical.config.ts vitest.critical.config.ts View all files Repository files navigation claude-devtools Terminal tells you nothing. This shows you everything. A desktop app that reconstructs exactly what Claude Code did — every file path, every tool call, every token — from the raw session logs already on your machine. 100% free, open source. No API keys. No configuration. Just download, open, and see everything Claude Code did. demo.mp4 Installation Platform Download Notes macOS (Apple Silicon) .dmg Download the arm64 asset. Drag to Applications. On first launch: right-click → Open macOS (Intel) .dmg Download the x64 asset. Drag to Applications. On first launch: right-click → Open Linux .AppImage / .deb / .rpm / .pacman Choose the package format for your distro (portable AppImage or native package manager format). Windows .exe Standard installer. May trigger SmartScreen — click “More info” → “Run anyway” Docker docker compose up Open http://localhost:3456 . See Docker / Standalone Deployment for details. The app reads session logs from ~/.claude/ — the data is already on your machine. No setup, no API keys, no login. Why This Exists Claude Code stopped telling you what it’s doing. Recent Claude Code updates repl

Source: Hacker News | Original Link

0byte Intro to PyTorch Easy to follow, visual introduction. 10th FEBRUARY, 2026 Table of Contents: what is PyTorch tensors basics autograd for automatic differentiation autograd in practice building a simple neural network YouTube what is PyTorch? PyTorch is currently one of the most popular deep learning frameworks. It is an open-source library built upon the Torch Library (it’s no longer in active development), and it was developed by Meta AI (previously Facebook AI). It is now part of the Linux Foundation . tensor basics Machine Learning (ML) is all about numbers. Tensor is a specialised container for those numbers. You might know tensors from maths or physics, but in machine learning, a tensor is simply PyTorch’s data type for storing numbers. Think of it like a more powerful version of a list or array. Tensors hold your training data and the weights Weights are numbers that determine how important each input is to the final decision. your model learns. What makes tensors special is that they come packed with useful functions. When you create a new tensor, you need to fill it with starting values. PyTorch offers many initialisation functions : torch.rand() , torch.randn() , torch.ones() , the list goes on. But what’s the difference between them? If they give you random numbers, which random numbers? And why there are so many ways to initialise it? The best way to understand is to see it. If we create thousands of random values using each function and plot them as a histogram, we can see exactly what to expect: import torch rand_sample = torch.rand(10000) randn_sample = torch.randn(10000) zeros_sample = torch.zeros(10) ones_sample = torch.ones(10) arange_sample = torch.arange(0, 10) linspace_sample = torch.linspace(0, 10, steps=5) eye_sample = torch.eye(5) empty_sample = torch.empty(10) Seeing it as a histogram paints a very clear picture. torch.rand() initialises tensor with random values between 0 and 1. torch.randn() with values that are mostly clustering around 0 and torch.eye() gives an identity matrix , and torch.empty() is… wait. Not empty? In theory is not empty. It allocates memory but does not initialise it, so the tensor contains whatever values happened to already be in that memory. If you see zeros, that’s just coincidence. torch.zeros() explicitly fills the tensor with zeros, whereas torch.empty() makes no guarantees at all – you should always write to it before reading from it. “Hey! What about own my data?” I’ve got you. Initialising tensors with random noise can be helpful, but ultimately, you want your own data for training. Let’s start with a simple example, you have data structure as follows: Bedrooms Size (m²) Age (years) Price (£k) 2 65 15 285 3 95 8 425 4 120 25 380 3 88 42 295 5 180 3 675 2 58 50 245 # Each row is one house: [bedrooms, bathrooms, size, age, price] houses = torch.tensor([ [2, 65, 15, 285], [3, 95, 8, 425], [4, 120, 25, 380], [3, 88, 42, 295], [5, 180, 3, 675], [2, 58, 50, 245] ], dtype=torch.float32) Y

Source: Hacker News | Original Link

Sina Visitor System

Source: Weibo Hot Search | Original Link

A Deep Dive into Apple’s .car File Format · DBG.RE Table of Contents Every modern iOS, macOS, watchOS, and tvOS application uses Asset Catalogs to manage images, colors, icons, and other resources. When you build an app with Xcode, your .xcassets folders are compiled into binary .car files that ship with your application. Despite being a fundamental part of every Apple app, there is little to none official documentation about this file format. In this post, I’ll walk through the process of reverse engineering the .car file format, explain its internal structures, and show how to parse these files programmatically. This knowledge could be useful for security research and building developer tools that does not rely on Xcode or Apple’s proprietary tools. As part of this research, I’ve built a custom parser and compiler for .car files that doesn’t depend on any of Apple’s private frameworks or proprietary tools. To make this research practical, I’ve compiled my parser to WebAssembly so it runs entirely in your browser, so no server uploads required. You can drop any .car file into the interactive demo below to explore its content. I’m considering open-sourcing these tools, but no promises yet! What is a .car File? When you call UIImage(named: “MyImage”) in Swift (or [UIImage imageNamed:@”MyImage”] in Objective-C), the private CoreUI.framework opens your app’s Assets.car file and retrieves the appropriate image for the current device context (screen scale, appearance mode, size class, etc) The “CAR” extension likely stands for “Compiled Asset Record” based on method names found in Xcode’s IBFoundation framework. You can inspect car files using Apple’s closed-source tools: actool : Compiles asset catalogs (usually invoked by Xcode) assetutil : Processes and inspects car files Running assetutil -I Assets.car produces a JSON dump of all assets: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 [ { “Appearances” : { “UIAppearanceAny” : 0 }, “AssetStorageVersion” : “Xcode 12.5 (12E262) via IBCocoaTouchImageCatalogTool” , “Authoring Tool” : “@(#)PROGRAM:CoreThemeDefinition PROJECT:CoreThemeDefinition-491\n” , “CoreUIVersion” : 691 , “DumpToolVersion” : 972.1 , “Key Format” : [ “kCRThemeScaleName” , “kCRThemeIdiomName” , “kCRThemeSubtypeName” , “kCRThemeDimension2Name” , “kCRThemeIdentifierName” , “kCRThemeElementName” , “kCRThemePartName” ], “MainVersion” : “@(#)PROGRAM:CoreUI PROJECT:CoreUI-691.2\n” , “Platform” : “ios” , “PlatformVersion” : “13.1” , “SchemaVersion” : 2 , “StorageVersion” : 17 , “Timestamp” : 1769308120 }, { “AssetType” : “Icon Image” , “BitsPerComponent” : 8 , “ColorModel” : “RGB” , “Colorspace” : “srgb” , “Compression” : “lzfse” , “Encoding” : “ARGB” , “Icon Index” : 1 , “Idiom” : “phone” , “Name” : “AppIcon” //… } The Foundation: BOM File Format Opening a .car file in a hex editor immediately reveals its foundation: the file begins with the ASCII magic BOMStore . BOM (Bill of Materia

Source: Hacker News | Original Link

No content available.

Source: Hacker News | Original Link

No content available.

Source: 36Kr | Original Link

No content available.

Source: 36Kr | Original Link

印孚瑟斯与Anthropic合作推出高端企业级人工智能解决方案-36氪 账号设置 我的关注 我的收藏 申请的报道 退出登录 登录 搜索 36氪Auto 数字时氪 未来消费 智能涌现 未来城市 启动Power on 36氪出海 36氪研究院 潮生TIDE 36氪企服点评 36氪财经 职场bonus 36碳 后浪研究所 暗涌Waves 硬氪 氪睿研究院 媒体品牌 企业号 企服点评 36Kr研究院 36Kr创新咨询 企业服务 核心服务 城市之窗 政府服务 创投发布 LP源计划 VClub VClub投资机构库 投资机构职位推介 投资人认证 投资人服务 寻求报道 36氪Pro 创投氪堂 企业入驻 创业者服务 创投平台 AI测评网 首页 快讯 资讯 推荐 财经 AI 自助报道 四川 最新 创投 汽车 科技 专精特新 直播 视频 专题 活动 搜索 寻求报道 我要入驻 城市合作 02 月 17 印孚瑟斯与Anthropic合作推出高端企业级人工智能解决方案 2026-02-17 14:29 分享至 打开微信“扫一扫”，打开网页后点击屏幕右上角分享按钮 据印度科技公司印孚瑟斯（Infosys）向证券交易所发布的公告显示，该公司将与Anthropic合作，为电信、金融服务、制造业及软件开发领域的企业开发并提供高端企业级人工智能解决方案。双方合作将首先从电信行业启动，由Anthropic设立专门中心，用于构建和部署适配行业特定业务的人工智能智能体。随后，合作范围将扩展至其他行业。印孚瑟斯的Topaz 平台与Anthropic的Claude大模型将实现融合，用于自动化复杂工作流程、加快软件交付，并为各行业打造具备自主执行能力的人工智能解决方案。（新浪财经） 下一篇 丹纳赫接近以100亿美元收购迈心诺 据英国金融时报援引未具名知情人士消息，美国医疗设备制造商丹纳赫（Danaher）即将达成一项近100亿美元的交易，收购医疗科技公司迈心诺（Masimo）。若不出现最后时刻意外，收购协议最早可能于周二宣布。丹纳赫与迈心诺均未立即回应媒体置评请求。（新浪财经） 50分钟前 24小时热榜 查看更多榜单 帮长辈清手机成过年“基操”，这份2026年最新攻略请收好 揭秘春晚“两个蔡明”背后，这家机器人公司的百日奋战 如何通过三个问题识破谎言 6小时前 年轻人春节回家账单：花销超2万，机票占一半 23小时前 在AI和机器人里找节目：马年春晚背后的科技大战揭秘 5小时前 这个春节999元租个人形机器人拜年，然后呢？ 4小时前 “今年春节，我终于跟它们和解了” 4小时前 最长春节假期，你需要的异地过年攻略 4小时前 关于36氪 城市合作 寻求报道 我要入驻 投资者关系 商务合作 关于我们 联系我们 加入我们 36氪欧洲站 36氪欧洲站 36氪欧洲站 Ai产品日报 网络谣言信息举报入口 热门推荐 热门资讯 热门产品 文章标签 快讯标签 合作伙伴 36氪APP下载 iOS & Android 本站由 阿里云 提供计算与安全服务 违法和不良信息、未成年人保护举报电话：010-89650707 举报邮箱：[email protected] 网上有害信息举报 © 2011~ 2026 北京多氪信息科技有限公司 | 京ICP备12031756号-6 | 京ICP证150143号 | 京公网安备11010502057322号 意见反馈 36氪APP 让一部分人先看到未来 36氪 鲸准 氪空间 推送和解读前沿、有料的科技创投资讯 一级市场金融信息和系统服务提供商 聚焦全球优秀创业者，项目融资率接近97%，领跑行业

Source: 36Kr | Original Link

Kimi连续融资超12亿美元，估值翻倍突破100亿美元-36氪 账号设置 我的关注 我的收藏 申请的报道 退出登录 登录 搜索 36氪Auto 数字时氪 未来消费 智能涌现 未来城市 启动Power on 36氪出海 36氪研究院 潮生TIDE 36氪企服点评 36氪财经 职场bonus 36碳 后浪研究所 暗涌Waves 硬氪 氪睿研究院 媒体品牌 企业号 企服点评 36Kr研究院 36Kr创新咨询 企业服务 核心服务 城市之窗 政府服务 创投发布 LP源计划 VClub VClub投资机构库 投资机构职位推介 投资人认证 投资人服务 寻求报道 36氪Pro 创投氪堂 企业入驻 创业者服务 创投平台 AI测评网 首页 快讯 资讯 推荐 财经 AI 自助报道 四川 最新 创投 汽车 科技 专精特新 直播 视频 专题 活动 搜索 寻求报道 我要入驻 城市合作 02 月 17 Kimi连续融资超12亿美元，估值翻倍突破100亿美元 2026-02-17 14:37 分享至 打开微信“扫一扫”，打开网页后点击屏幕右上角分享按钮 在完成上一轮5亿美元融资仅一个多月后，月之暗面Kimi新一轮超7亿美元的融资即将完成交割，本轮由阿里、腾讯、五源、九安等老股东联合领投。此外，其新一轮100-120亿美元估值的融资亦已经开启，照此计算，Kimi最新估值直接翻倍并突破100亿美元大关，总额超12亿美元的连续两轮融资也是近一年来大模型行业的最高融资金额。（财联社） 原文链接 下一篇 2026春节档票房破10亿 36氪获悉，据大麦娱乐旗下灯塔专业版数据，截至2月17日13时04分，2026春节档（2月15日-2月23日），档期总票房（含预售）突破10亿。千问发布的数据显示，过去两天，用户通过AI在大麦买电影票的订单环比增长372倍，其中来自三四线城市的AI订单量暴涨782倍。 1小时前 24小时热榜 查看更多榜单 帮长辈清手机成过年“基操”，这份2026年最新攻略请收好 揭秘春晚“两个蔡明”背后，这家机器人公司的百日奋战 如何通过三个问题识破谎言 6小时前 年轻人春节回家账单：花销超2万，机票占一半 23小时前 在AI和机器人里找节目：马年春晚背后的科技大战揭秘 5小时前 这个春节999元租个人形机器人拜年，然后呢？ 4小时前 “今年春节，我终于跟它们和解了” 4小时前 最长春节假期，你需要的异地过年攻略 4小时前 关于36氪 城市合作 寻求报道 我要入驻 投资者关系 商务合作 关于我们 联系我们 加入我们 36氪欧洲站 36氪欧洲站 36氪欧洲站 Ai产品日报 网络谣言信息举报入口 热门推荐 热门资讯 热门产品 文章标签 快讯标签 合作伙伴 36氪APP下载 iOS & Android 本站由 阿里云 提供计算与安全服务 违法和不良信息、未成年人保护举报电话：010-89650707 举报邮箱：[email protected] 网上有害信息举报 © 2011~ 2026 北京多氪信息科技有限公司 | 京ICP备12031756号-6 | 京ICP证150143号 | 京公网安备11010502057322号 意见反馈 36氪APP 让一部分人先看到未来 36氪 鲸准 氪空间 推送和解读前沿、有料的科技创投资讯 一级市场金融信息和系统服务提供商 聚焦全球优秀创业者，项目融资率接近97%，领跑行业

Source: 36Kr | Original Link

飞猪：假期前三天出行的出境游机票、当地玩乐等服务订单量同比增长均超30%-36氪 账号设置 我的关注 我的收藏 申请的报道 退出登录 登录 搜索 36氪Auto 数字时氪 未来消费 智能涌现 未来城市 启动Power on 36氪出海 36氪研究院 潮生TIDE 36氪企服点评 36氪财经 职场bonus 36碳 后浪研究所 暗涌Waves 硬氪 氪睿研究院 媒体品牌 企业号 企服点评 36Kr研究院 36Kr创新咨询 企业服务 核心服务 城市之窗 政府服务 创投发布 LP源计划 VClub VClub投资机构库 投资机构职位推介 投资人认证 投资人服务 寻求报道 36氪Pro 创投氪堂 企业入驻 创业者服务 创投平台 AI测评网 首页 快讯 资讯 推荐 财经 AI 自助报道 四川 最新 创投 汽车 科技 专精特新 直播 视频 专题 活动 搜索 寻求报道 我要入驻 城市合作 02 月 17 飞猪：假期前三天出行的出境游机票、当地玩乐等服务订单量同比增长均超30% 2026-02-17 14:44 分享至 打开微信“扫一扫”，打开网页后点击屏幕右上角分享按钮 36氪获悉，“史上春节假期”进入节中出游高峰期。飞猪数据显示，假期前三天出行或入住的旅游服务预订量同比去年大幅增长。其中，酒店间夜量同比去年增长超80%，门票张数同比去年增长54%。跨境旅行也迎来高峰时刻。飞猪数据显示，假期前三天出行的出境游机票、当地玩乐等服务订单量同比去年增长均超30%，海外游客的入境游机票张数同比去年劲增345%。 下一篇 Kimi连续融资超12亿美元，估值翻倍突破100亿美元 在完成上一轮5亿美元融资仅一个多月后，月之暗面Kimi新一轮超7亿美元的融资即将完成交割，本轮由阿里、腾讯、五源、九安等老股东联合领投。此外，其新一轮100-120亿美元估值的融资亦已经开启，照此计算，Kimi最新估值直接翻倍并突破100亿美元大关，总额超12亿美元的连续两轮融资也是近一年来大模型行业的最高融资金额。（财联社） 13分钟前 24小时热榜 查看更多榜单 帮长辈清手机成过年“基操”，这份2026年最新攻略请收好 揭秘春晚“两个蔡明”背后，这家机器人公司的百日奋战 如何通过三个问题识破谎言 6小时前 年轻人春节回家账单：花销超2万，机票占一半 23小时前 在AI和机器人里找节目：马年春晚背后的科技大战揭秘 5小时前 这个春节999元租个人形机器人拜年，然后呢？ 4小时前 “今年春节，我终于跟它们和解了” 4小时前 最长春节假期，你需要的异地过年攻略 4小时前 关于36氪 城市合作 寻求报道 我要入驻 投资者关系 商务合作 关于我们 联系我们 加入我们 36氪欧洲站 36氪欧洲站 36氪欧洲站 Ai产品日报 网络谣言信息举报入口 热门推荐 热门资讯 热门产品 文章标签 快讯标签 合作伙伴 36氪APP下载 iOS & Android 本站由 阿里云 提供计算与安全服务 违法和不良信息、未成年人保护举报电话：010-89650707 举报邮箱：[email protected] 网上有害信息举报 © 2011~ 2026 北京多氪信息科技有限公司 | 京ICP备12031756号-6 | 京ICP证150143号 | 京公网安备11010502057322号 意见反馈 36氪APP 让一部分人先看到未来 36氪 鲸准 氪空间 推送和解读前沿、有料的科技创投资讯 一级市场金融信息和系统服务提供商 聚焦全球优秀创业者，项目融资率接近97%，领跑行业

Source: 36Kr | Original Link

Four Column ASCII Robbie’s Garbage, Collected. Four Column ASCII January 31, 2017 I found this gem on Hacker News the other day. User soneil posted to a four column version of the ASCII table that blew my mind. I just wanted to repost this here so it is easier to discover. Here’s an excerpt from the comment: I always thought it was a shame the ascii table is rarely shown in columns (or rows) of 32, as it makes a lot of this quite obvious. eg, http://pastebin.com/cdaga5i1 It becomes immediately obvious why, eg, ^[ becomes escape. Or that the alphabet is just 40h + the ordinal position of the letter (or 60h for lower-case). Or that we shift between upper & lower-case with a single bit. You know in ASCII there are 32 characters at the beginning of the table that don’t represent a written symbol. Backspace, newline, escape – that sort of thing. These are called control characters . In the terminal you can type these control characters by holding the CTRL (control characters, get it?) key in combination with another key. For example, as many experienced vim users know pressing CTRL+[ in the terminal (which is ^[ in caret notation ) is the same as pressing the ESC key. But why is the escape key triggered by the [ character? Why not another character? This is the insight soneil shares with us. Remember that ASCII is a 7 bit encoding. Let’s say the following: The first two bits denote the group of the character (2^2 so 4 possible values) The remaining five bits describe a character (2^5 so 32 possible values) In the linked table, which I reproduce below, the four groups are represented by the columns and the rows represent the values. 00 01 10 11 NUL Spc @ ` 00000 SOH ! A a 00001 STX ” B b 00010 ETX # C c 00011 EOT $ D d 00100 ENQ % E e 00101 ACK & F f 00110 BEL ‘ G g 00111 BS ( H h 01000 TAB ) I i 01001 LF * J j 01010 VT + K k 01011 FF , L l 01100 CR – M m 01101 SO . N n 01110 SI / O o 01111 DLE 0 P p 10000 DC1 1 Q q 10001 DC2 2 R r 10010 DC3 3 S s 10011 DC4 4 T t 10100 NAK 5 U u 10101 SYN 6 V v 10110 ETB 7 W w 10111 CAN 8 X x 11000 EM 9 Y y 11001 SUB : Z z 11010 ESC ; [ { 11011 FS < \ | 11100 GS = ] } 11101 RS > ^ ~ 11110 US ? _ DEL 11111 Now in this table, look for ESC. It’s in the first group, fifth from the bottom. It’s in the first column so its group has bits ’00’, the row has bits ‘11011’. Now look on the same line, what else is there? Yep, the ‘[‘ character is there, be it in a different column: 10 11011 means [ 00 11011 means ESC So when we you type CTRL+[ for ESC , you’re asking for the equivalent of the character 11011 ( [ ) out of the control set. Pressing CTRL simply sets all bits but the last 5 to zero in the character that you typed. You can imagine it as a bitwise AND. 10 11011 ([) & 00 11111 (CTRL) = 00 11011 (ESC) This is why ^J types a newline, ^H types a backspace and ^I types a tab. This is why if you cat -A a Windows text file, it has ^M printed all over (meaning CR , because newlines are CR+LF on Windows). Liking this? Follow me:

Source: Hacker News | Original Link

SvarDOS Main page I Packages I Help I Forum SvarDOS – an open-source DOS distribution for PCs of the 1980-2000 era SvarDOS is an open-source project that is meant to integrate the best out of the currently available DOS tools, drivers and games. DOS development has been abandoned by commercial players a long time ago, mostly during early nineties. Nowadays it survives solely through the efforts of hobbyists and retro-enthusiasts, but this is a highly sparse and unorganized ecosystem. SvarDOS aims to collect available DOS software and make it easy to find and install applications using a network-enabled package manager (like apt-get, but for DOS and able to run on a 8086 PC). Minimalist and 8086-compatible Once installed, SvarDOS is a minimalistic DOS system that offers only a DOS kernel, a command interpreter and the most basic tools for system administration. It is then up to the user to install additional packages. Care is taken so SvarDOS remains 8086-compatible, at least in its most basic (core) configuration. Open-source SvarDOS is published under the terms of the MIT license. This applies only to SvarDOS-specific files, though – the auxilliary packages supplied with SvarDOS may be subject to different licenses (GPL, BSD, Public Domain, Freeware…). Multilingual The system can be set up in a wide selection of languages: English, German, French, Polish, Russian, Italian and more. No “versions” SvarDOS is a rolling release, meaning that it does not adhere to the concept of “versions”. Once the system is installed, its packages can be kept up-to-date using the SvarDOS online update tools (pkg & pkgnet). Community and help Need to get in touch? Wish to submit some packages, translate SvarDOS to your language, or otherwise contribute? Or maybe you’d like some information about SvarDOS? Come visit the SvarDOS community forum . You may also wish to take a look at the project’s bug tracker . Downloads SvarDOS is available in a variety of installation images for different sizes of floppy disks. The CD image presented below contains only a virtual floppy to install the bare-bone SvarDOS system. If you’d like a CD image that contains lots of extra packages, you may want to look here . Stable build (20250427) CD-ROM ISO 1.44M floppy disks 1.2M floppy disks 720K floppy disks 360K floppy disks bootable USB image BNS build (20250427) CD-ROM ISO image A “talking” version for blind persons. Uses the PROVOX screen reader and requires a Braille ‘n Speak synthesizer connected to the COM1 port. If running as a VM, the BNS can be emulated . Bleeding edge build 20251005 ⚠ This build has not been thoroughly tested! Let us know how it works for you. The links above point to the latest builds of installation images. Archival builds can be found in the files section . Development 🔧 Wondering how SvarDOS is built? All the build-related files, scripts and sources of SvarDOS-specific tools are stored on the project’s SVN server. To pull the sources using the standard su

Source: Hacker News | Original Link

Thinking Hard Burns Almost No Calories—But Destroys Your Next Workout | VO2 Max Pro Blog Back to Blog I had a theory for most of my adult life: hard days at work burn a lot of calories. The kind of days where you’re deep in complex problems for eight hours straight, making decisions, debugging, writing. By evening, I’d feel utterly drained. Surely all that thinking was metabolically expensive. Surely it justified skipping the interval session I’d planned. Then I looked at the numbers. And then I looked at what that cognitive exhaustion was actually doing to my workouts. The first part was humbling. The second part was alarming. Your brain is expensive, but thinking is cheap Your brain consumes roughly 20 to 25% of your body’s total energy at rest. For an organ that accounts for about 2% of your body weight, that’s extraordinarily costly. It runs almost exclusively on glucose, consuming about 120 grams per day, and it can’t store meaningful reserves. It needs a constant supply delivered through the bloodstream. Here’s the part that surprised me: almost all of that energy goes to baseline maintenance. Keeping 86 billion neurons electrically charged. Maintaining synaptic connections. Running the background processes that keep you conscious and breathing and aware of your surroundings. The “idle cost” of having a brain is enormous. The incremental cost of actually thinking hard? Almost nothing . Studies measuring brain glucose consumption during demanding cognitive tasks find increases of only about 5% above baseline. Over the course of a full day of intense mental work, that translates to maybe 100 to 200 extra calories. That’s a banana and a half. Even chess grandmasters, who sit for hours in states of intense concentration, burn only about 1.67 calories per minute while playing, compared to 1.53 calories per minute at rest. A 10% premium for one of the most cognitively demanding activities humans engage in. The reason is structural. Your brain’s energy budget is dominated by housekeeping. Maintaining the electrochemical gradients across neuronal membranes, the resting activity of vast neural networks, the constant baseline hum of consciousness. Conscious thought, the stuff that feels effortful, is a thin layer of activity on top of this expensive foundation. It’s like turning on a desk lamp in a building where the HVAC system already runs 24 hours a day. The lamp is real. It’s just not where the bill comes from. So no, your day of deep work didn’t burn enough extra calories to justify the extra slice of pizza. I found this mildly deflating. But what I found next was far more consequential. The 15% problem In 2009, Samuele Marcora and colleagues at Bangor University published a study in the Journal of Applied Physiology that I think about every time I consider training after a long cognitive day. They had 16 subjects complete two experimental conditions. In one, participants performed 90 minutes of a demanding cognitive task (the AX-CPT, a sustaine

Source: Hacker News | Original Link