AI 资讯

Hey everyone! 👋 I'm a complete beginner and today I'm officially kicking off my #100DaysOfCode challenge with Python. I've dabbled with the idea of learning to code for a while, but this time I want to actually commit - so I'm posting daily updates here to keep myself accountable and track my progress over the next 100 days. My plan: Post a short update here every day - what I learned, what I struggled with, and what's next Eventually move into some small real-world projects once I've got the basics down Why I'm doing this: I want to build real skills, not just "watch tutorials and forget everything." Writing it down publicly (even anonymously) keeps me honest and hopefully connects me with others on the same path. If you're also learning Python or doing a 100 days challenge, I'd love any tips, resources, or just to follow along with each other's progress! Day 1 status: Just setting up my environment and going through the basics — nothing exciting yet, but everyone starts somewhere! 100DaysOfCode #Python #Beginner #LearnToCode

The xAI account on X has announced its official rebranding to SpaceXAI, five months after it merged with SpaceX.

Imagine you are planning a surprise birthday party. You need invitations, food, decorations, and a cake. You call different places to get these things. You tell each one exactly what you need. "I need 20 red balloons." "I need a chocolate cake for 10 people." This is how many computer programs talk to each other. They use something called an API (Application Programming Interface). An API is like a menu. You pick what you want. You get exactly that. It works well for simple tasks. But what if your party plans change? What if you decide on a theme mid-conversation? Traditional APIs can feel a bit rigid then. They don't always remember your past requests. They don't understand the bigger picture. Now, imagine talking to a super-smart party planner. You start by saying, "I'm planning a party." The planner asks, "For how many people?" You say, "About 20." Then you mention, "It's for a birthday." The planner instantly suggests a cake size. It recommends decorations based on your earlier answers. This smart planner remembers everything you said. It understands your overall goal. It uses something like MCP (Model Context Protocol). MCP is a new way for computers to talk. It's like having a real conversation. It's much smarter than a simple menu order. You will soon understand why this difference is a game-changer. Traditional APIs: The Fixed Menu Approach Let's start with what you might already know. Many apps you use every day rely on APIs. An API is like a waiter in a restaurant. You look at the menu. You tell the waiter your exact order. "I want a cheeseburger with fries." The waiter takes your order to the kitchen. The kitchen prepares only that specific meal. Then the waiter brings it back to you. This is how most apps work together. One app sends a very specific request. It asks for a certain piece of information or to perform a specific action. The other app performs that task. It sends back a very specific response. Think of ordering from an online store. You click

Most AI incidents don't happen because the model gave a bad answer. They happen because nobody was...

This is part of my work with 01MVP on OpenNomos — a project that helps founders validate ideas before building. The $0 Launch I once spent three months building a product. It had everything: authentication, payments, a polished UI, dark mode. I was proud of it. Launch day: 27 visitors. Zero signups. I had spent 90 days building and precisely zero days asking anyone if they wanted what I was building. I was solving a problem that existed only in my head. The Hardest Lesson The product wasn't bad. The code was fine. The UI was clean. The problem was that I never validated the core assumption: does anyone actually have this problem, and would they pay to solve it? This is the most common failure mode in indie hacking. You build something you think is cool, polish it to perfection, and launch to silence. The code was never the bottleneck. The validation was. What I Do Differently Now Talk to 10 people before writing code. Not surveys. Not landing page analytics. Actual conversations. "Would you use this? Would you pay for it? Why or why not?" Build a mockup, not a product. A Figma prototype or even a Google Form that simulates the core workflow is enough to test willingness to engage. Charge from day one. Free users will tell you nice things. Paying users will tell you the truth. If nobody will pay, the idea isn't ready. Kill fast. Most ideas fail. The goal isn't to make every idea succeed — it's to fail the bad ones quickly so you can find the good ones. Why This Matters More in 2026 In 2016, building a product was hard. You needed to know how to code, set up servers, handle deployments. The barrier to building kept bad ideas from being built. In 2026, Cursor writes your code, v0 generates your UI, and Replit deploys it. The barrier to building has collapsed to near zero. But here's the problem: AI can help you build anything. It cannot help you figure out what's worth building. The result is a flood of well-built products that nobody wants. The bottleneck shifted from

I used to fix weak AI drafts by asking for better prose. "Make it clearer." "Make it more persuasive." "Make it sound less generic." The output improved a little. Then it failed in the same place: the article looked polished, but nobody remembered what it was trying to say. TL;DR: Before you ask AI to write, fill a five-line editorial brief: audience, takeaway, material to use, first point to place, and scope delegated to AI. The prompt gets shorter because the decision-making moved back to the human. Quick answer: what should I decide before asking AI to write? Decide these five things before the first draft: Who is the reader? What should that reader take away? Which material should be used, and which material should be cut? What should appear first so the reader can follow the argument? Which part is the AI allowed to decide, and which part stays with you? That is the difference between an AI writing prompt and an AI writing workflow. A prompt says, "write a useful article about this." A workflow says, "write for this reader, to deliver this point, using this material, in this order, while leaving these decisions untouched." Here is the copy-paste version I now use before drafting: cat > ai-writing-brief.md << ' BRIEF ' Audience: Takeaway: Material to use: First point to place: Scope delegated to AI: BRIEF Output: a five-line brief that makes the human decisions visible before the AI starts drafting. If those five lines are empty, a better prompt usually will not save the article. It will only make the generic answer prettier. Why polished AI writing still feels empty AI can satisfy the instruction you give it. If you ask for more detail, it adds detail. If you ask for simpler language, it removes jargon. If you ask for a friendly tone, it softens the edges. All of that can be correct and still useless. The missing part is not grammar. It is aim. A draft can have headings, clean paragraphs, and natural transitions while still leaving the reader with no decision,

Disclosure: I write about projects in the OpenNomos ecosystem, including Swipe Cleaner. The Problem With Photo Cleaners Most photo cleaning apps have a dirty secret: your photos leave your device. They get uploaded to some server for "AI processing," "cloud analysis," or just because the developer didn't think about it. Swipe Cleaner takes the opposite approach. Everything happens on your iPhone. Not a single pixel leaves your device. Let me break down why that matters, and how it actually works under the hood. The Architecture Swipe Cleaner is built on three principles: 1. On-device processing, always. Image analysis, duplicate detection, and similarity matching all run locally using Apple's Core ML and Vision frameworks. No cloud roundtrips, no server costs, no privacy policy loopholes. 2. Tinder-style UX for decisions. You don't manage a grid of thumbnails and checkboxes. You swipe. Right to keep, left to delete. This isn't just a UI gimmick — it's a deliberate choice to reduce decision fatigue. When you have 3,000 photos to clean, you need flow, not friction. 3. Sandboxed storage access. The app requests permission for exactly what it needs. It doesn't ask for your entire photo library if you only want to clean screenshots. This is iOS privacy-by-design done right. Why On-Device Matters Now We're in a weird moment. AI capabilities are exploding, which means the temptation to "send it to the cloud for better results" is stronger than ever. But at the same time, Apple is pushing hard in the opposite direction — Private Cloud Compute, on-device ML, differential privacy. Swipe Cleaner aligns with where the platform is going, not where the industry has been. The Technical Trade-offs Local-first isn't free. Here's what you give up: Model size constraints. You can't run a 70B parameter vision model on an iPhone. The models need to be small, optimized, and ruthlessly efficient. No cross-device sync. Your cleaning decisions stay on one device. No cloud means no sync. For

Every developer has been there: you need to format a JSON blob, decode some Base64, or convert a timestamp. You open your terminal, look for the right npm package, or — worse — write a quick script. I used to do this too. Then I discovered a better pattern. The Problem with Local CLI Tools Local tools have real drawbacks: Installation overhead : npm install -g some-tool for a one-time task Version rot : tool stops working after OS update No sharing : you format JSON but cant send the result to a colleague Environment drift : works on your machine, not on staging Online Tools as a Pattern Opennomos Json (reachable via opennomos.com/en/project/01KJ850Z7PNGXHXESBM68HE12Y) represents a shift: developer tools as a platform , not as utilities you install. What makes this different: Zero install — browser tab, done Cross-device — phone, laptop, any OS Shareable results — formatted output has a URL you can send to teammates Timestamp converter built in — ms, seconds, ISO 8601, bidirectional Base64 codec — no need for a separate site The Bigger Trend We are seeing the same pattern across the dev ecosystem: GitHub Codespaces (IDE in browser), Replit (runtime in browser), Vercel (deployment in browser). The next frontier is utility tools in browser . Why run jq locally when a well-designed online tool does it faster and gives you a share link? Try It Head to opennomos.com/en/project/01KJ850Z7PNGXHXESBM68HE12Y — the JSON tools are free, fast, and part of a broader contributor rewards system that makes open-source tooling sustainable. Built as part of the Nomos Build-in-Public series.

Quick Note In my previous article, I mentioned that Linux Troubleshooting Flow for Beginners would be the final post in this series. While preparing it, I realized there were a few practical Linux skills every beginner should learn first. These topics will make the troubleshooting guide much easier to understand and follow. Before we wrap up the series, we'll cover: Package Management Finding Files & Text Viewing Files Efficiently File Compression Then we'll bring everything together in the final Linux Troubleshooting Flow for Beginners. Introduction Installing software on Linux is very different from Windows. On Windows, you usually download an .exe installer. On Linux, software is typically installed and managed using package managers . This is one of the most practical skills every Linux beginner should learn early. What is a Package? A package is a ready-to-install bundle that contains: The main program Required libraries Configuration files Documentation Examples: nginx , git , docker , curl , vim Think of a package as a ready-to-install software box. What is a Package Manager? A package manager is a tool that installs, updates, removes, and manages software packages. Instead of downloading software manually, you simply run a command. Example: sudo apt install git The package manager automatically: Downloads packages from trusted repositories Install required dependencies automatically Upgrade installed software Removes them cleanly Instead of manual downloading, you just run one command. Why Use a Package Manager? Without package managers, you would have to: Search for software manually Download files from websites Install dependencies yourself Update each application separately Package managers automate all of this. What is a Repository? Package managers download software from repositories. A repository is a trusted online collection of software packages maintained by your Linux distribution. Instead of downloading software from random websites, Linux install

DeepSeek vs Qwen vs Kimi vs GLM: Which AI API Actually Wins in 2025? I've spent the last decade designing systems that need to stay up no matter what. 99.9% uptime isn't a marketing slogan for me — it's the difference between a happy customer and a 3am incident call. So when the Chinese model ecosystem exploded with options like DeepSeek, Qwen, Kimi, and GLM, I didn't just glance at the benchmarks. I pulled the levers, watched the dashboards, and stress-tested every endpoint I could get my hands on. Here's what I found after weeks of running these models behind load balancers, instrumenting them with p99 latency tracking, and watching how they behave when you throw production traffic at them. The Multi-Region Reality Nobody Talks About Most comparison articles treat AI APIs like they're interchangeable endpoints you curl against. That's fine for a weekend hackathon. It's dangerous for production. When I'm architecting a service that depends on an LLM, I care about three things before I care about quality: p99 latency under sustained load Failover behavior when a region gets congested Cost per million tokens at the rate I'm actually consuming I ran each of these four providers through a series of synthetic workloads — bursts of 200 concurrent requests, sustained 50 RPS for an hour, and cold-start recovery tests. The numbers told a story that the marketing pages don't. The Data at a Glance Here's the TL;DR before I dive in. DeepSeek gives you the best price-to-performance ratio, full stop. Qwen has the widest catalog of model sizes I've ever seen from a single provider. Kimi costs a premium but earns it on reasoning-heavy workloads. GLM punches above its weight on Chinese-language tasks and offers multimodal support that the others don't. Dimension DeepSeek Qwen Kimi GLM Provider DeepSeek (幻方) Alibaba (阿里) Moonshot AI (月之暗面) Zhipu AI (智谱) Output price range $0.25–$2.50/M $0.01–$3.20/M $3.00–$3.50/M $0.01–$1.92/M Budget pick V4 Flash @ $0.25/M Qwen3-8B @ $0.01/M N/A GL

Hi everyone, I'm Marc, the founder of ClaimMate AI. I've been building an AI software engineering platform that helps developers generate code, explain existing code, debug issues, create tests, review code, and build applications from simple prompts or voice. I'm still in the early stages and would really appreciate honest feedback from other developers. Why I Built It I wanted one workspace where developers could chat with AI, generate code, debug problems, and iterate on ideas without constantly switching between multiple tools. I'd Love Your Feedback If you have a few minutes, I'd appreciate any thoughts on: Is the interface easy to understand? Which feature would you use most? What would stop you from using it regularly? What feature is missing? You can try it here: https://ClaimMateAI.pro I'm not looking for praise—I genuinely want constructive feedback that will help improve the product. Thanks for your time!

If you've ever built a scraper that worked perfectly in dev and then got blocked or CAPTCHA'd the moment it hit production traffic volume, you already know why proxy choice matters. This post breaks down residential proxies from a practical, implementation-focused angle: what they are, when to use them vs. alternatives, how to wire them into common tools, and how the major providers stack up. TL;DR Residential proxies route requests through real ISP-assigned IPs, so they're harder for anti-bot systems to fingerprint than datacenter IPs. Rotating residential proxies are for scraping/data collection. Sticky sessions (or static ISP proxies) are for anything stateful — logins, checkout flows, long-lived account sessions. Nstproxy is a good default pick if you want residential, static ISP, and mobile proxies under one API/dashboard instead of juggling multiple vendors for different parts of your stack. For large-scale enterprise scraping, Oxylabs and Bright Data have the most mature tooling. For budget/prototype work, IPRoyal, DataImpulse, and Webshare are worth testing. Proxy types, quickly Type Use for Pros Watch out for Residential Scraping, SERP checks, ad verification Looks like real user traffic Usually billed per GB Static ISP Long-lived sessions, account workflows Fast + stable IP Less useful for high-volume rotation Datacenter Speed-sensitive, low-stakes tasks Cheap, fast Easiest to fingerprint/block Mobile Mobile-first platforms/apps Strongest trust signal Most expensive per GB A production-grade scraping/automation stack often uses more than one of these at once — e.g., rotating residential IPs for crawling, and static IPs pinned to specific browser profiles for anything that requires a login. Wiring a residential proxy into your code Most providers give you a host:port endpoint plus username:password auth, and let you control rotation/session stickiness through the username string. A typical setup looks like this: Python ( requests ): import requests proxy_ho

A simple bubble tea experiment completely changed how I think about AI recommendations. Last week, I asked ChatGPT a question that seemed almost impossible to get wrong. "What are the best bubble tea brands in my city?" Surprisingly，some of the recommended brands were companies I had never heard of before. A few weren't even available in the cities I had lived in. After asking the same question to Claude, Gemini, and DeepSeek, I noticed something interesting that many of the same brands which I'm not familiar kept appearing. AI Isn't Judging Your Brand Humans recommend products because they have experiences. But AI does none of those things. It doesn't know whether one brand actually tastes better than another. Instead, it tries to generate the most statistically reliable answer based on the information it can understand. So, AI doesn't recommend the best brand. It recommends the brand it understands best. The Experiment That Changed My Perspective Once I realized this, I started paying closer attention. I didn't only test bubble tea but also the restaurants, beauty brands, consumer electronics, and ravel recommendations. Again and again, I noticed a pattern. Brands that consistently appeared in AI recommendations usually had several characteristics: Clear product descriptions Well-structured websites Consistent public information Plenty of third-party coverage Easy-to-understand positioning Meanwhile, some excellent brands barely appeared at all, because AI had much less reliable information to work with. That's when I stopped thinking about AI recommendations as opinions. They're much closer to information retrieval problems than human preferences. Consumers Are Already Changing Their Habits This matters because people are beginning to use AI differently from traditional search engines. Instead of searching "Best bubble tea near me", many people (especially the youth) now ask AI to recommend a healthy milk tea brand." The AI becomes the decision maker before the c

AI is coming for your cloud budget before it comes for your job. Everyone is talking about what AI will do to engineers. Nobody is talking about what AI services are doing to cloud bills right now. Every team is spinning up AI workloads. LLM endpoints; vector databases; GPU instances; embedding pipelines. All of them expensive. Most of them ungoverned. I have seen environments where AI service costs doubled in 30 days because nobody set a budget limit. No alerts. No guardrails. Just an invoice that surprised everyone at month end. This is the new cloud sprawl. But instead of forgotten EC2 instances it is uncontrolled AI API calls and model inference costs running 24/7. The same discipline that applies to cloud infrastructure applies here. Who is allowed to spin up AI services. What models are approved for use. How are API keys being managed. Are there rate limits in place. Is anyone reviewing the cost daily. AI governance in the cloud is not an AI problem. It is a cloud security and FinOps problem wearing an AI hat. The teams that win with AI are not the ones moving fastest. They are the ones who built guardrails before the bill arrived. Govern your AI workloads the same way you govern everything else in your cloud. Before the invoice does it for you.

The terminal and Mac menu bar companion for Kickbacks.ai Discussion | Link

The capability phase is over For the past two years, the AI conversation has been about...

One of the biggest takeaways from Chapter 3 of AI Engineering was realizing that building an AI model is only part of the challenge. Figuring out 𝗵𝗼𝘄 𝘁𝗼 𝗲𝘃𝗮𝗹𝘂𝗮𝘁𝗲 𝗶𝘁 𝗳𝗮𝗶𝗿𝗹𝘆 𝗮𝗻𝗱 𝗮𝗰𝗰𝘂𝗿𝗮𝘁𝗲𝗹𝘆 can be just as difficult. With traditional software, it's usually easy to tell whether something works. If a calculation is wrong or a test fails, you know there's a bug. But AI doesn't always work that way. A model can generate multiple reasonable answers to the same question, making it much harder to determine which one is actually better. That made me think: 𝗛𝗼𝘄 𝗱𝗼 𝘄𝗲 𝗸𝗻𝗼𝘄 𝗶𝗳 𝗮𝗻 𝗔𝗜 𝗺𝗼𝗱𝗲𝗹 𝗶𝘀 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗶𝗺𝗽𝗿𝗼𝘃𝗶𝗻𝗴? 𝗕𝗲𝗻𝗰𝗵𝗺𝗮𝗿𝗸𝘀 𝗡𝗲𝗲𝗱 𝘁𝗼 𝗞𝗲𝗲𝗽 𝗘𝘃𝗼𝗹𝘃𝗶𝗻𝗴 Reading this section made me realize how difficult it is for evaluation benchmarks to keep up with the pace of AI development. The chapter explains that GLUE (General Language Understanding Evaluation) was introduced in 2018 to measure how well language models performed on common natural language tasks. But within about a year, models had already become so good at it that researchers introduced SuperGLUE in 2019 as a more difficult benchmark. GLUE evaluates tasks such as: Question answering Sentiment analysis Sentence similarity Text classification The chapter also mentions newer benchmarks like: SuperGLUE MMLU (Massive Multitask Language Understanding) MMLU-Pro Each one was introduced because the previous benchmark was no longer challenging enough. What I found interesting is that a model getting a higher benchmark score doesn't always mean it understands language better. Sometimes it simply means the model has become very good at solving that particular benchmark. 𝗨𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴 𝗘𝗻𝘁𝗿𝗼𝗽𝘆 𝗮𝗻𝗱 𝗣𝗲𝗿𝗽𝗹𝗲𝘅𝗶𝘁𝘆 Another section I really enjoyed was the explanation of entropy and perplexity. The chapter explains entropy as a measure of how much information a token carries and how difficult it is to predict the next token in a sequence. Perplexity measures uncertainty. If a model is very uncertain about what comes next, its perplexity will be higher. If

Test Isolation: A Lesson I Learned While Migrating Playwright Tests During my software engineering internship, I helped optimize our CI pipeline by identifying which E2E tests could safely run in parallel. That work quickly taught me that the biggest obstacle wasn't Playwright or Python, it was test isolation. This article is about that lesson. What is test isolation? A simple rule I now use is this: if a test can't run by itself with the same outcome, it probably isn't truly isolated. A well-isolated test should produce the same result whether it: runs by itself runs first or last runs after another test runs in parallel with hundreds of other tests To understand test isolation, it also helps to understand what state means. State isn't limited to database rows. During the migration, I found tests interacting with many different kinds of state. database records global configuration filesystem resources application caches If any of these are shared between tests, they become potential sources of hidden dependencies. How tests lose isolation As I started reading the existing test suite, I noticed a recurring pattern. Many tests assumed something about the environment instead of creating it themselves. Some expected specific data to already exist. Others modified global settings without restoring them afterward. Some searched for rows based on their position in a table instead of using a stable identifier like a name or ID. None of these looked particularly problematic when reading a single test. The problems only appeared once the entire suite started running together. One test would leave behind data another test didn't expect. A shared configuration would silently affect unrelated tests. A UI assertion would suddenly fail because another test inserted an extra row into the same table. Individually, the tests appeared independent. Together, they formed hidden dependencies. Not all shared state is equally difficult to isolate One realization that helped me reason abou