AI 资讯

trpc has a scheduled workflow called "Lock Issues & PRs." Its own scorecard shows it failing on almost every run. It is still scheduled, still running, still red. trpc ships excellent software, which is exactly the point: if a project this careful has a workflow that has been red for ages, the rest of us almost certainly do too. It is not a one-off. drizzle-orm has one ("Unpublish release"). cal.com has one ("PR Update"). I scanned 35 popular open-source repos and the same thing kept turning up: a scheduled workflow that fails on nearly every run, quietly, for a long time. Why nobody notices GitHub does email you when a scheduled workflow fails. So how do these survive? Two reasons. First, those emails are routine. You get them for flaky reruns and transient blips too, so you filter them out. Second, a workflow that is always red stops reading as a signal. It is just how that row looks now. I did exactly this on my own project. GitHub emailed me that a workflow had failed. The next day it emailed again. I saw it, told myself I would fix it tomorrow, and promptly forgot. It was my nightly database backup, quietly broken the whole time, and I only caught it when a failure-rate number crept up where I would notice. An always-red workflow is not free It burns minutes every run to produce nothing but a red X. Worse, it trains you to ignore the failure that actually matters: the day a real one lands in the same inbox you have learned to skim past. How to find yours Open your Actions tab and look at the scheduled workflows, the cron-triggered ones nobody watches. If the last several runs are all red, you found one. From the CLI: gh run list --workflow = "Lock Issues & PRs" --status = failure What to do about it Two honest options: fix it, or if the workflow is genuinely abandoned, turn it off. Do not leave it scheduled and red. gh workflow disable "Lock Issues & PRs" Or drop the schedule trigger from the workflow file if it should not run on a timer at all. A disabled work

Two days ago GitHub emailed me to say one of my workflows had failed. The next day it emailed me again. I saw it, told myself I would fix it tomorrow, and promptly forgot. It was my nightly database backup, quietly broken the whole time, and I only caught it because a failure-rate number nudged up. A failed run at least gets you an email. A slow run gets you nothing. GitHub never pings you when CI quietly takes twice as long, runs the whole suite twice per PR, or rebuilds dependencies from scratch every time. That waste compounds where no one looks. Here are the usual culprits, each with the exact fix. When I scanned 35 popular open-source repos, not one had a fully clean config. 32 of 35 had no concurrency control, 33 of 35 had no job timeouts, and 22 of 35 ran the full suite twice on every PR. If projects this polished leave minutes on the table, the rest of us definitely do. Your suite runs twice on every PR Trigger a workflow on both push and pull_request and, for a branch in the same repo, opening a PR fires both. You just paid for two identical runs. This one is pure waste and it can roughly halve your PR-related minutes. Trigger on pull_request , and keep push for your default branch: on : push : branches : [ main ] pull_request : Old runs don't cancel when you push again Push a fix 30 seconds after the first push and, with no concurrency group, both runs go to completion. The first is dead weight, and it is holding a slot in your queue while it finishes. This hides even when you do have a group: astro has a concurrency group on one workflow but left off cancel-in-progress , which our scan estimates leaves roughly 1,850 minutes a month on the table. Add a group keyed on the branch, with cancel-in-progress , so a new push supersedes the old run: concurrency : group : ${{ github.workflow }}-${{ github.ref }} cancel-in-progress : true Every run reinstalls dependencies from scratch No cache means every run re-downloads and rebuilds your dependencies. On a typical

Hundreds of contractors working on a project for Meta pretended to be kids—and then prompted rival chatbots like Gemini and ChatGPT to discuss high-risk subjects.

Starting in IT I started in IT at a local school in my small town in December 2024. It was my first job out of college after earning my B.S. in Cybersecurity. None of the infrastructure was updated. Everything was failing, and luckily, I had one other IT person there: my director. I honestly think he knew less than I did, and he would get frustrated at almost every ticket. Even then, I knew I wanted a role where I could code. Fast forward to more recently, he had a freak out and quit. Now we have a two-person team, and everything is finally up to date and functioning. Even with things improving, I still knew I wanted to move toward a SWE-type role. Learning to Code I first decided to learn C# for Unity. I was really into game dev while I was in college, so it felt like a natural place to start. I began with the Microsoft/freeCodeCamp C# certification, and I surprisingly really enjoyed it. I made a few small games on itch.io that no one cared about, but I had fun building them. After that, I went on a bit of a language-hopping spree. I jumped from C# to C++, then into full-stack web development. I actually stuck with web dev for a while and really enjoyed it. But this cycle went on for awhile of just constant swapping. Wannabe Founder Then something switched overnight. I went from writing maybe 0-5% AI-generated code to using AI for nearly everything. I started spam-building startup ideas that did not really go anywhere. I may have made around $2-3k from them, but most of the time I was just chasing money and building whatever I thought had the quickest path to making some. I got seriously addicted to vibe coding. I tried Codex, Cursor, Claude, and basically anything with AI in it. I did like Codex the most, though. Eventually, I realized I had almost completely stopped coding by hand. I was not passionate about the startup ideas I was building. I loved coding, and I knew I had to step back. Back to Coding Now I am back to coding without AI assistance. I will eventua

Modern Property Management Systems depend on continuous data exchange between internal modules and external services. Bookings, calendar updates, guest communication, cleaning tasks, and maintenance triggers all generate operational events that must be processed quickly and reliably. Free PMS platforms such as PMS.Rent rely on robust queueing and message‑broker layers to ensure that these events never get lost and are always processed in the correct order. At the core of this architecture is the concept of distributed message‑broker orchestration, which enables the PMS to scale horizontally, maintain predictable performance, and avoid bottlenecks during peak operational periods. Why Message Brokers Matter A PMS handles thousands of small but critical operations every day. Without a message broker, these operations would compete for system resources, causing delays, blocking workflows, and creating inconsistent states. A broker solves this by: receiving events, storing them durably, routing them to the correct processors, retrying failed operations, ensuring ordered execution when required. This creates a stable foundation for automation and real‑time synchronization. Queue Types Inside a PMS A modern PMS typically uses several queue types: Operational queues for bookings, calendar updates, and guest messages Automation queues for cleaning tasks, reminders, and workflow triggers Synchronization queues for channel managers and external APIs Fallback queues for events that require manual review Each queue isolates a specific category of tasks, preventing unrelated operations from interfering with each other. Distributed Workers Workers are lightweight processes that consume events from queues. They operate in parallel, allowing the PMS to scale dynamically. If the system detects increased load — for example, during high‑season booking spikes — it simply launches more workers. Workers typically perform tasks such as: updating property calendars, generating guest notific

This blog post is created for the purposes of entering the Hack the Zero Stack with Vercel v0 and AWS Databases hackathon. #H0Hackathon Building Innward: A B2B Hospitality Operating System with Vercel and Amazon Aurora The hospitality industry is notorious for relying on "legacy" software—clunky, slow, and disconnected. For the Hack the Zero Stack hackathon, I set out to build Innward , a modern, AI-ready Property Management System (PMS) that proves you can build enterprise-grade B2B tools in record time using Vercel v0 and AWS Databases. The Vision: Moving Beyond the Spreadsheet Hotel managers don't just need a place to store "Room 101: Occupied." They need to solve the "Hidden Math" of revenue management. This means: Relational Complexity: Linking dates, room groups, and individual stays. Dynamic Pricing: Deriving rates based on occupancy and logic-based rules. Market Intelligence: Real-time benchmarking against competitors. The "Zero Stack": Vercel + Amazon Aurora To handle this complexity, I chose Amazon Aurora PostgreSQL (Serverless v2) . Why Aurora for B2B? In a B2B SaaS environment, data isolation and relational integrity are non-negotiable. Aurora provided the robust relational power needed to join complex pricing tables while scaling automatically as more hotels (tenants) join the platform. The Zero-Secret Architecture One of the most rewarding parts of this build was implementing the AWS RDS Signer . Following the "Zero Stack" philosophy, I moved away from static database passwords. Innward uses IAM-based authentication to communicate between Vercel and AWS. By utilizing the @aws-sdk/rds-signer , the application generates short-lived tokens on the fly. This means even if an environment variable were leaked, the database remains locked tight. // lib/db.ts snippet const signer = new Signer ({ credentials : awsCredentialsProvider ({ roleArn : process . env . AWS_ROLE_ARN ! , clientConfig : { region : process . env . AWS_REGION }, }), region : process . env .

AI makes it easier to build the wrong thing with confidence. That is the part I think a lot of beginner builders and freelancers miss. The obvious story is that AI makes execution faster. That is true. I can ask an AI coding tool to explain an error, compare implementation options, inspect a project, write code, refactor a screen, generate a QA checklist, or help me pick up where I left off. That is a huge change. But speed is not the whole story. When the tool gets faster, your judgment becomes more important, not less. You have to decide what the project is allowed to become. You have to decide which tradeoffs are acceptable. You have to decide whether the output actually matches the user's job. You have to decide when the AI is solving the real problem and when it is decorating the wrong one. In my freelance work, AI changed the job from searching and stitching to directing, reviewing, and verifying. That sounds cleaner than it feels. Directing means you need to know what outcome you want. Reviewing means you need to notice when the answer is plausible but wrong. Verifying means you cannot treat a green checkmark, a pretty screen, or a confident explanation as proof that the app actually works. The beginner mistake is believing AI removes the need to think clearly. The better rule is this: AI removes some friction from execution, then hands you more responsibility for scope. The Faster Tool Still Needs A Smaller Job When I started using AI heavily for software work, the old research loop changed immediately. Before modern AI tools, a lot of software work meant digging through documentation, old forum posts, Stack Overflow answers, YouTube videos, outdated examples, and half-related blog posts until something clicked. You stitched pieces together and hoped the tutorial you found still matched the version of the framework you were using. Now you can ask the tool directly. That is better. It is also dangerous if you confuse a fast answer with a good product decision

By mid-afternoon, you can know where your floor will close by end of day — accurately enough to make the remaining hours a decision, not a guess. Here's how intraday performance forecasting works and what it takes to build it. The Problem With Yesterday's Numbers Most contact centers have end-of-day metrics. Dials, connects, conversion rate against target. Those numbers are accurate, useful for trend analysis, and arrive the next morning. By the time you see them, the day is already over. The decisions that drive outcomes happen during the day — in real time, when hours remain to influence the result. Do you push harder in the final stretch? Adjust campaign priority? Pull a server that's underperforming? Those decisions get made in the afternoon with one question underneath all of them: where are we going to close? If you're answering that question with yesterday's data and experienced intuition, you're working with an information deficit that compounds every day it stays open. How Intraday Forecasting Works The system records dial conversion rates at regular intervals throughout the business day. Not a snapshot at end of day. A continuous read of how the floor is performing as it performs. Every morning, before the floor opens, the model retrains. It processes the intraday conversion patterns from previous days — how conversion tends to develop through the morning, when it typically accelerates, when it softens, how afternoon performance differs from morning — and calibrates to the current operation's historical data. As the day runs, the forecast updates on a regular schedule. Each update incorporates actual conversion data that's come in, narrowing the prediction window. By mid-afternoon, with hours remaining, the model's error range has compressed enough that the closing metric is predictable within an actionable range. Not a rough estimate. A forecast with a documented accuracy track. What this changes in practice: Before the forecasting system, the afternoon c

I recently built a multi-agent customer support system on Azure AI Foundry and NVIDIA NIM. First time doing anything like this. Made four predictions upfront about what would happen. Three of them were wrong. Here is what I actually learned. 1. "Tokens" is not a unit of cost It is a unit of work. The price per unit of work varies by 5-10x depending on which model did the work. I was tracking total token count across both the small 9B model and the large 49B model as if they cost the same. They do not. Total tokens went up in the optimized version. Cost in dollars probably went down. I was measuring the wrong thing the whole time. 2. A verbatim hash cache on natural language traffic deflects ~0% of queries I predicted 25-40% cache deflection. The actual number was 0%. Every query in my test set was a unique string, so the hash-based cache never had a single chance to fire. A verbatim cache is not a simpler version of a semantic cache. It is a different thing entirely. If your workload is natural language, build semantic similarity caching from day one, not as an upgrade later. 3. configure_azure_monitor() does not capture OpenAI SDK calls by default You need to install and initialize opentelemetry-instrumentation-httpx explicitly: pip install opentelemetry-instrumentation-httpx==0.61b0 from opentelemetry.instrumentation.httpx import HTTPXClientInstrumentor HTTPXClientInstrumentor().instrument() Without this, your App Insights Logs will show customMetric and performanceCounter entries (CPU, memory) but nothing about what your agent actually did. 4. Pin your OpenTelemetry versions or everything breaks Installing opentelemetry-instrumentation-httpx without version pinning pulled in opentelemetry-api 1.42.1. But azure-monitor-opentelemetry-exporter needs opentelemetry-api==1.40. The conflict is silent until things start misbehaving. Pin everything to the 0.61b0 / 1.40.0 line: pip install \ "opentelemetry-api==1.40.0" \ "opentelemetry-instrumentation==0.61b0" \ "opentelem

Look at almost any piece of electronics on your desk and you will find a small light staring back at you. A router with a row of blinking status lights. A power brick with a steady green dot. A development board with a tiny red point that flickers every time it does something. We barely notice these lights anymore, but each one descends from a single laboratory breakthrough in 1962, when an engineer at General Electric coaxed a sliver of semiconductor into glowing visible red for the first time. Who invented the first visible LED The engineer was Nick Holonyak Jr., a consulting scientist at General Electric's lab in Syracuse, New York, and a former student of John Bardeen, one of the inventors of the transistor. On October 9, 1962, Holonyak demonstrated the first practical visible-spectrum light-emitting diode. It emitted red light, and it worked at room temperature, which made it genuinely useful rather than a laboratory curiosity. What made his approach different was the material. Other researchers in the early 1960s were building diodes that emitted infrared light, which is invisible to the human eye. Holonyak gambled on a different alloy, gallium arsenide phosphide, and it paid off with the first light a person could actually see coming out of a semiconductor. He was so confident in the idea that he predicted LEDs would one day replace the incandescent bulb. At the time that sounded outlandish. Today it is simply how lighting works. Why a tiny red light mattered so much The incandescent bulb that Thomas Edison commercialized makes light by heating a filament until it glows. That is wildly inefficient, because most of the energy escapes as heat rather than light, and the filament eventually burns out. An LED works on a completely different principle. When current flows across a specially engineered semiconductor junction, electrons release their energy directly as photons. There is no filament to burn out, almost no wasted heat, and the device can switch on and o

By RUGERO Tesla ( @404Saint ). There is a persistent illusion in Industrial Control Systems (ICS) security research: that high-level libraries, abstraction frameworks, or protocol tooling give you a real understanding of Operational Technology (OT) behavior. They don’t. They hide the architecture. Determined to understand what actually happens when a Programmable Logic Controller (PLC) receives a control-plane command, I built an EtherNet/IP and Common Industrial Protocol (CIP) sandbox from scratch. No Scapy. No protocol wrappers. Just raw sockets, a Linux loopback interface, a cpppo simulator, and a passive monitoring tool ( enip_monitor.py ) capturing traffic in real time. It looked clean on paper. Then I reached the application layer. And things stopped behaving like theory. The Reality of the “Industrial Abstraction Layer” If you come from Modbus or traditional IT networking, you’re used to linear memory spaces—fixed registers, predictable offsets, and flat addressing. EtherNet/IP and CIP discard that model entirely. Instead, they introduce a structured object system wrapped inside multiple encapsulation layers: +-----------------------------------------------------------+ | EtherNet/IP Encapsulation Header (24 bytes) | | → Session control, commands (0x0065, 0x006F) | +-----------------------------------------------------------+ | Common Packet Format (CPF) | | → Routing, addressing, and transport segmentation | +-----------------------------------------------------------+ | CIP Application Layer | | → Service codes (0x4C, 0x4D, 0x10, etc.) | +-----------------------------------------------------------+ To communicate with a PLC at the wire level, your code must: Establish a session using RegisterSession (0x0065) Wrap all subsequent requests in SendRRData (0x006F) Encode routing information inside CPF structures Construct symbolic or logical paths for the CIP Message Router Ensure strict byte alignment across nested payload layers A single mistake in any layer b

Token burn and business outcomes are not correlated. More burn means more inefficiency, not more value. The electricity problem Imagine you walk into a dark room. Turning on a light helps you see. Turning on every light in the building does not help you see better. It's still the same room. Now every surface is equally lit, the contrast is gone, and you're paying for power you didn't use. Tokens work the same way. A focused prompt with clear scope is the single overhead light over your desk. A sprawling prompt with unlimited exploration is every light in the building — you're burning power, not producing insight. Tokens are electricity, not output. More throughput doesn't mean more value. I've had weeks where I burned through my allocation and looked back at the end to find nothing concrete. Code that worked but went unused. Exploratory branches that dead-ended. Agents that generated plausible-looking output that didn't survive first review. A lot of motion. Not much progress. The ceiling stops you from doing that indefinitely. It forces a moment of reflection: did this burn produce anything real? If the answer is no, more capacity isn't the fix. More discipline is. Three patterns I now use instead I started paying attention to what actually ships versus what just burns context. I gave the patterns names so I could catch myself faster: RTK — Read The Knowledgebase. A focused 15-minute read of the codebase, identifying the exact files and exact changes, saves 200K+ tokens of exploratory waste. The agent doesn't discover the shape of the task — it executes against a known one. Caveman — compress before you prompt. Strip greetings, filler words ("I think", "basically", "Let me know if that makes sense"), and closing courtesies. Every word in your prompt multiplies across every response token. Less fluff in means less fluff out. Ponytail — spec the minimum viable solution. "Robust", "scalable", "enterprise-grade", "comprehensive" — these words invite scope creep. Specif

South Korea targets physical AI lead and commercial humanoid robots by 2028.

While Kara Zor-El's appearance at the end of James Gunn's Superman was a very pleasant surprise, Warner Bros. Discovery's plan to fast-track a standalone Supergirl feature always felt a little dubious. It seemed odd that, after Superman, the studio wanted to flesh out its new cinematic universe with films about another Kryptonian and one of […]

VCs remain thirsty to fund AI coding startups. This one, founded by investor Chamath Palihapitiya, is no exception.

To dispose of an enemy, make use of another enemy. Use a second party to deliver the blow yourself....

Google is expanding Gemini’s personalized AI image generation to eligible free users in the U.S., allowing the chatbot to create images based on your interests and data from connected Google apps.

Google expands personalized intelligence to Gemini app image creation