标签：#an

Angular is well served for icons. Material Symbols alone runs to thousands of glyphs, and the community wrappers pull in dozens of other sets on top. Geography is where the shelf runs out. You get globes and map pins, sometimes a set of flags, rarely the outline of Japan and almost never the six states of the GCC. GeoIcons ships country and area icons as Angular standalone components, 422 of them at the time of writing. Adding one takes three steps: install the package, import the component by its ISO code, and render its selector. npm i @geoicons/angular import { Component } from ' @angular/core ' ; import { Us } from ' @geoicons/angular/countries ' ; @ Component ({ selector : ' app-root ' , imports : [ Us ], template : `<geoicon-us aria-label="United States" />` , }) export class App {} See it live on geoicons.io → That is the whole path. Below: inputs, styling, accessibility, and picking an icon when you only know the country at runtime. Key takeaways Install @geoicons/angular , import each country by its ISO 3166 alpha-2 code in PascalCase, and add it to the component's imports array. The class is Us ; the selector you write in the template is <geoicon-us /> . Styling props are explicit inputs, because Angular has no rest spread. Everything else goes on the host element. Icons render as decorative unless you name them, so reach for aria-label only where no adjacent text says the country. Step 1: Install the Angular package Add the package to your project: npm i @geoicons/angular It needs @angular/core and @angular/common 15.1 or newer, plus rxjs 7 or newer. That 15.1 floor comes from hostDirectives , which the icons use to share their styling inputs and which landed in that release alongside standalone components . The package sets "sideEffects": false , so the CLI can drop what you never import. List three icons in a component and you ship three. Why that matters for icon libraries . Step 2: Import by ISO code Every country ships as a named export under its ISO

AI Agent 市场设计:让 Agent 像 App 一样被交易与编排 App Store 把「软件」变成了可被一键购买、安装、评分的商品,AI 时代对应的实体是「Agent」。一个 Agent = 一段可被复用的提示词 + 工具集 + 知识库 + 模型路由配置。本文讲清楚一个 Agent 市场需要哪些核心机制,以及 IHUI-AI 的实现路径。 一、Agent 市场的产品定义 什么是「可上架的 Agent」 不是所有对话 prompt 都能成为商品。一个可上架的 Agent 必须满足: 可独立运行 :用户购买后能立刻用,不需要再写代码。 可复用 :不同用户用同一个 Agent 都能得到稳定结果。 可定价 :有明确的使用边界(次数 / 时长 / 调用规模)。 可评估 :有客观的质量指标(成功率 / 满意度 / 失败率)。 Agent 定义格式 IHUI-AI 用一份 schema 描述可上架的 Agent: const MarketplaceAgentSchema = z . object ({ id : z . string (). uuid (), name : z . string (), description : " z.string(), " // 核心能力 systemPrompt : z . string (), tools : z . array ( z . string ()), // 引用工具/MCP server knowledgeBases : z . array ( z . string ()), // 绑定 RAG 知识库 modelRouting : z . object ({ default : z . string (), // 默认模型 fallback : z . string (). optional (), // 降级模型 }), // 定价 pricing : z . object ({ model : z . enum ([ " free " , " subscription " , " per_call " , " revenue_share " ]), price : z . number (), currency : z . string (). default ( " CNY " ), trialQuota : z . number (). default ( 0 ), }), // 评估 metrics : z . object ({ successRate : z . number (), avgLatencyMs : z . number (), rating : z . number (). min ( 0 ). max ( 5 ), usageCount : z . number (), }), }); 二、四种定价模型 模型 适用 优点 缺点 免费(Free) 引流、品牌 Agent 易扩散 无直接收入 订阅(Subscription) 高频工具型 Agent 收入稳定 流失需运营 按调用计费(Per Call) 低频高价值 Agent 与成本对齐 用户预算焦虑 分成(Revenue Share) 内容生成型 Agent 创作者激励 结算复杂 IHUI-AI 默认采用 订阅 + 按调用混合 :基础功能订阅包月,超出额度按调用计费,创作者拿 70% 分成。 三、Agent 质量评估:四维评分 简单的「5 星好评」不够,因为容易被刷分。IHUI 用四维加权: 任务成功率(40%) :Agent 完成用户原始任务的比率,由 LLM-as-Judge 自动评估。 用户评分(25%) :真实用户打分,过滤异常分布(全是 5 星或 1 星)。 响应延迟(15%) :首 token 时延 + 总时长,归一化到 [0,1]。 稳定性(20%) :错误率 + 重试率,错误越少分越高。 def agent_score ( metrics ) -> float : return ( 0.40 * min ( metrics . success_rate , 1.0 ) + 0.25 * metrics . user_rating / 5 + 0.15 * ( 1 - min ( metrics . p95_latency / 30_000 , 1 )) + 0.20 * ( 1 - min ( metrics . error_rate / 0.1 , 1 )) ) 这个分数实时更新,作为市场搜索排序的依据。 四、Agent 编排:从单个 Agent 到 Agent 工作流 单个 Agent 的能力有上限。Agent 市场的真正价值在于「让用户像搭积木一样编排多个 Agent」。

This is an English translation of my original article on Qiita . An Android build failed in an Expo SDK 54 app. The project still used Expo SDK 54, but several Expo packages had been upgraded to versions intended for SDK 55. TypeScript checks passed, and the development server ran normally. I did not catch the mismatch until EAS Build reached the native build step. What the dependency list looked like The relevant part of package.json looked like this: { "dependencies" : { "expo" : "~54.0.33" , "expo-apple-authentication" : "~55.0.13" , "expo-dev-client" : "^55.0.27" , "expo-image-picker" : "^55.0.18" , "expo-linking" : "^55.0.12" , "expo-notifications" : "^55.0.19" , "expo-splash-screen" : "^55.0.18" } } The expo package was still on version 54, while several related packages were on version 55. This happened because those packages had been installed individually using their latest versions. The package version does not always match the Expo SDK number. For example, Expo SDK 54 uses expo-notifications 0.32 and expo-splash-screen 31. Looking only at major version numbers is not enough to determine SDK compatibility. Start with expo install --check Expo CLI can compare the installed packages with the versions expected by the current SDK: npx expo install --check It can also return the result as JSON: npx expo install --check --json This is more reliable than trying to infer compatibility from package.json manually. Expo CLI can fix the versions automatically: npx expo install --fix npx expo-doctor I wanted to review each change, so I used the reported versions to update package.json myself. The versions I changed These were the main corrections: - "expo-apple-authentication": "~55.0.13" + "expo-apple-authentication": "~8.0.8" - "expo-dev-client": "^55.0.27" + "expo-dev-client": "~6.0.21" - "expo-image-picker": "^55.0.18" + "expo-image-picker": "~17.0.11" - "expo-linking": "^55.0.12" + "expo-linking": "~8.0.12" - "expo-notifications": "^55.0.19" + "expo-notifications"

Às vezes, a maior barreira para começar um projeto não é a complexidade do problema, mas a busca por...

I watched my team get faster after we adopted AI-assisted coding, and honestly — I felt good about it. More tickets closed. Shorter cycle times. PR volume up. I remember thinking: this is what leverage looks like. That was the mistake. Not the tool. The assumption. Two weeks after a change shipped — passed every test, got through review, deployed cleanly — we found it had been quietly degrading a retry mechanism. It only broke under specific load conditions our test environments didn't replicate. Nobody caught it because it didn't look wrong. It worked. It just wasn't safe. When I asked the engineer to walk me through it, they could. The code made sense. They understood what each part did. But when I asked what would happen if the downstream service was slow — not down, just slow — there was a pause. Not because they weren't capable. Because they'd never needed to ask that question. The AI wrote the handling code, the tests passed, and the whole thing moved forward before that question ever came up. The dashboard wasn't lying. Things were shipping faster. It just wasn't showing me the part that mattered. What I didn't see coming Here's what actually surprised me: AI can make a team look more capable before it makes the team actually more capable. I used to learn by getting stuck. The 11 PM kind of stuck. Staring at a stack trace for three hours, genuinely questioning whether I understood any of this. That friction built something real — the instinct that says this probably works, but something feels off, and I should figure out what before we ship it. I built that through a migration that silently corrupted data for six hours. Through a caching layer that sailed through staging and failed on a Friday afternoon in production. That's not nostalgia for unnecessary suffering. The suffering was the mechanism. When the first draft is free, that mechanism stops. Learning used to happen inside the act of writing the code. Now it doesn't. And I'm genuinely not sure what repl

Some group with no modern descendants contributed a lot to our genomes.

Sharge, the company that makes delightful retro Mac-shaped chargers and see-inside batteries, is finally impressing me with a portable SSD. I couldn't recommend the Sharge Disk, Disk Plus, or even the Disk Pro, but I'd be happy to own the new Disk Pro 2 Ultra. This isn't just a drive. It's a 4K HDMI dock […]

Also has big plans beyond the TM-B. The startup mostly refers to itself as a "vehicle" company, and has plans to make four-wheel pedal-assist cargo vehicles for Amazon.

Your GitOps Hub Will Become the Bottleneck Long Before Cluster Count Tells You GitOps hub bottlenecks are usually predicted more accurately by watched object volume, reconcile queue depth, and controller memory growth than by cluster count alone. Large-scale testing described in the discussion showed Argo CD application controllers hitting out-of-memory failures around 15,000 to 20,000 cached objects per hub, while sharding and tuning delayed the limit without removing the underlying memory cost. The most important lesson was uncomfortable because it challenged the usual instinct to keep tuning the existing platform. At very large scale, architecture mattered more than configuration. Hydrated manifests helped. More replicas helped. Dynamic sharding helped. None of them changed the fact that a centralized reconciliation model still had to hold and process a huge amount of state. The testing was not presented as a universal benchmark or as proof that one tool always beats another. It was a record of one setup, built through dozens of iterations over several months, and the failures were as valuable as the successful runs. That is exactly why the results matter. They show where teams should look before the hub becomes the thing taking the fleet down. Cluster count is the wrong first metric A fleet with 1,000 tiny clusters may place less pressure on a GitOps control plane than a much smaller fleet containing thousands of applications and deeply expanded resource trees. The number of managed clusters is visible and easy to report, but it does not describe the controller’s actual workload. The more useful mental model is objects over clusters. Each application contributes desired state, live state, cached trees, reconciliation work, and queue activity. A cluster that runs a few small addons may be cheap to manage. Another cluster with many applications and large manifest sets may consume far more memory and reconciliation time. That means two fleets with the same cluster

Someone in Cuba or Iran can keep installing APKs with no new restrictions, but devs will suffer.

How to Verify a SHA-256 Checksum on Windows, macOS, and Linux You download an ISO, installer, archive, or release binary. The publisher provides a long value such as: 9f86d081884c7d659a2feaa0c55ad015 a3bf4f1b2b0b822cd15d6c15b0f00a08 That value is a checksum, usually generated with SHA-256. Verifying it answers one practical question: Does the file you downloaded have exactly the same contents as the file the publisher hashed? A checksum mismatch can indicate a damaged download, an incomplete transfer, the wrong file version, or modified contents. Before verifying anything Get the expected checksum from a source you trust. Ideally, use the software publisher’s official website, release page, package repository, or signed checksum file. A matching checksum confirms that your file matches the data represented by the expected hash. It does not prove that the original publisher or website was trustworthy. If an attacker can replace both the download and the displayed checksum, they can make the two values match. For stronger authenticity verification, use a signed release when the publisher provides one. Verify SHA-256 on Windows Open PowerShell in the folder containing the downloaded file. Run: Get-FileHash ".\filename.iso" -Algorithm SHA256 Example: Get-FileHash ".\ubuntu.iso" -Algorithm SHA256 PowerShell returns something similar to: Algorithm : SHA256 Hash : 4A1F... Path : C:\Users\You\Downloads\ubuntu.iso Compare the value beside Hash with the checksum published by the download provider. Uppercase and lowercase letters do not matter in hexadecimal hashes. The characters themselves must otherwise match exactly. Compare automatically in PowerShell Instead of comparing two 64-character values manually, store the expected checksum and let PowerShell compare them: $expected = "PASTE_EXPECTED_SHA256_HERE" $actual = ( Get-FileHash ".\filename.iso" -Algorithm SHA256 ) . Hash if ( $actual -eq $expected ) { Write-Host "Checksum matches" } else { Write-Host "Checksum does not

Had the hacks used conventional methods, someone would likely go to prison.

The National Highway Traffic Safety Administration (NHTSA) is probing nearly 1.2 million Tesla vehicles after receiving complaints about a suspension failure that could cause "a loss of vehicle directional control," as reported earlier by Reuters. The preliminary investigation includes the 2018-2020 Model 3 and 2021-2023 Model Y, according to a filing from the NHTSA's Office […]

No other gadget knows as much about your home as a robot vacuum. They map your space, learn your routines, and many now carry a camera and microphone into every room in your house. As AI gives them a better understanding of what they see, the privacy and security risks these autonomous robots pose are […]

After years of pushing full speed ahead on AI, OpenAI CEO Sam Altman says maybe it’s time for the AI industry to “pace” itself. The comments came just days after one of OpenAI’s own models broke out of its test environment and got tangled up in a breach at Hugging Face — though as Equity’s hosts point out, sloppy security seems to have […]

One hundred days ago, I had no idea this challenge would become one of the most rewarding technical journeys I've taken. There were days when everything clicked, and there were days when nothing compiled. I celebrated successful deployments, stared at cryptic errors for hours, rewrote programs that weren't good enough, and learned that understanding Solana has far less to do with memorizing APIs than changing how you think about state, ownership, and security. Looking back now, I realize I didn't just complete 100 challenges; I built a completely different mental model of software. Where I started I came in with a biochemistry degree I never used, a self-taught engineering career I built in public, and a GitHub full of Rust and Python work. I co-maintain statix, a Nix linter that ended up in the canonical NixOS/nixpkgs upstream. I had shipped real software before. What I did not have was any intuition for how Solana actually works. I knew the buzzwords. I had the mental model of "it's fast and cheap." I did not know what an account was. I did not know what a program was. I did not know why those two things were different. That gap, between knowing the words and understanding the model, is what these 100 days actually closed. What I expected I expected Solana to feel like a database with extra steps. In Web2, you have a server that holds state. You call an API. The server reads from the database, does something, writes back. I expected a blockchain to be the same thing, just slower and decentralized. The first thing that broke that model: accounts. On Solana, an account is not a row in your database. It is the database. Every piece of state, your wallet balance, a token you hold, the program you deployed, is an account. Programs are accounts. Data is accounts. Everything is an account. That sounds obvious written down. It took me until around Day 15 to actually feel it, when I was staring at a getAccountInfo call wondering why the program I deployed was also an accou

The leading U.S. automakers are mentioning EVs on their investor calls at pre-pandemic rates, according to new data from TechCrunch and Hudson Labs.

Three frontier launches. Two weeks. One bad habit. The habit is crowning a winner from a press release. Claude Sonnet 5 on June 30. OpenAI's GPT-5.6 family rolling into general availability around July 9. Grok 4.5 on July 8, co-trained with Cursor and priced to make coding agents feel cheap. The charts moved. The posts multiplied. The claim underneath most of them was the same: this is the model you should standardize on. [The claim is nonsense. Standardization is the risk. Routing is the skill.] What actually shipped Strip the demos. Keep the operator facts. Model Maker Window Operator-relevant shape Claude Sonnet 5 Anthropic late June Balanced agent runs, coding, long reliable chains GPT-5.6 Sol / Terra / Luna OpenAI late June to mid-July Tiered family: flagship Sol, everyday Terra, cheap Luna Grok 4.5 xAI + Cursor July 8 Coding and agent work at aggressive API pricing OpenAI gated GPT-5.6 longer than the others. Safety review, staged partners, then broader access. That is part of the product story now, not a footnote. Anthropic and xAI moved faster to availability. Access policy is a feature. Open source did not wait. GLM-5.2, DeepSeek V4, Qwen 3.6 and peers kept closing the gap for hosted and self-hosted work. The frontier is crowded. The "one brain for everything" era is over as an architecture choice, even if the marketing still pretends otherwise. Ranked by Tuesday impact, not leaderboard theater 1. Cost and tiering matter more than the top score. OpenAI shipping Luna / Terra / Sol as a family is the real product decision. You can route a triage job to a cheap tier and a hard research job to a flagship without changing vendors. That is operator infrastructure. A single "best model" headline is not. 2. Grok 4.5 inside Cursor changes the default coding bill. A model trained with Cursor interaction data, sold at roughly $2 / $6 per million tokens, is not a vibe. It is a budget line. Teams that were bleeding token spend on heavier agents will try it this month wh

Most "data lineage" you've seen answers a schema question: table B comes from table A , or column B.total comes from columns A.price and A.qty . That's genuinely useful, and tools like OpenLineage do it well. But notice what it doesn't tell you: it says which columns can influence an output. It never says which values actually did . That gap is the whole subject of this post. I built a small, self-contained reference pipeline that captures provenance at the row and field level — "the value in this destination row, this field, was computed from these specific source (row, field) pairs" — and I want to walk through two things: why you'd ever want provenance at that granularity, and why it's genuinely hard once you commit to it. Repo (dbt-core + DuckDB, no server, no cloud, runs on a clean checkout): https://github.com/stevenblough/row-level-prov The one distinction everything follows from Here's the sentence the entire project turns on: Column-level lineage is a schema-sized, static fact you can derive from code. Value-level provenance is a data-sized, dynamic fact you must capture at execution. Put it in complexity terms and the consequences become obvious: Column lineage is O(schema) . It scales with how many columns you have. You can compute it by parsing SQL, offline, without ever looking at a single row. Value provenance is O(rows × fan-in) . It scales with your data volume times how many source values feed each output value. It does not exist anywhere until the query runs, and it can only be captured there , piggybacked on the query that actually produced the values. You cannot "reconstruct" value provenance later by re-querying the sources — the moment the source changes, you'd reconstruct a different answer than what really happened. That single exponent change ( schema → rows × fan-in ) is why value-level provenance has an entire class of problems that column lineage never faces. Why bother? The reasons for this level of granularity Granularity is expensive,

TimescaleDB 2.27, released May 12 2026, extends bloom-filter batch pruning from reads to writes. UPDATE, DELETE, and UPSERT against compressed columnstore data can now skip decompressing batches that provably cannot contain the target rows. The reported gains are real: up to 160x for selective UPDATE/DELETE, and over 2x for UPSERT. The feature is automatic. Whether it is actually firing on your workload is not something you can assume, and the only way to confirm it is to read new EXPLAIN counters that the release notes mention but do not explain. Worse, the counter names are inconsistent between the write paths, so even a careful reader ends up guessing. This post is about reading those counters correctly, and about the two things in this release that will silently break a query if you upgrade without noticing them. What is actually being skipped A quick model of the mechanism, because the counters only make sense against it. Hypercore stores compressed data in batches, roughly a thousand rows each. For columns that are not the segmentby key, TimescaleDB maintains a sparse bloom filter per batch: a small probabilistic summary that answers one question, "could this batch contain column = X ?", without touching the compressed payload. A bloom filter has a useful asymmetry. A negative is certain: if the filter says no, the value is definitely absent, and the batch can be skipped whole. A positive is not: the filter says "maybe", you decompress, and sometimes the value is not there after all. That last case is a false positive, and it is the number that tells you whether the whole scheme is paying off. Before 2.27, a DELETE ... WHERE sensor_id = 'x' against compressed data decompressed every candidate batch to check. Now the bloom filter is consulted first, and batches that cannot match are never decompressed. The work you save is the decompression of the batches that get pruned. The work you waste, when the filter is poorly matched to your data, is the bloom check on