标签：#ens

LiteRT-LM brings native support for Gemma 4 Multi-Token Prediction (MTP) drafters, enabling up to 2.2x faster inference. The framework is expanding beyond Kotlin and C++ adding support for new Swift and a JavaScript APIs. By Sergio De Simone

GitHub热门项目 | The GEP-powered self-evolving engine for AI agents. Auditable evolution with Genes, Capsules, and Events. | evomap.ai | Stars: 7,871 | 140 stars today | 语言: JavaScript

Qisquiz: A Qiskit v2.X Certification Prep App I built Qisquiz , a web app for learning Qiskit v2.X and preparing for the IBM Certified Quantum Computation using Qiskit v2.X Developer - Associate certification exam. You can try the app here: https://qisquiz.vercel.app/ The GitHub repository is here: https://github.com/dorakingx/qisquiz The concept of Qisquiz is simple: Master Qiskit, one quiz at a time. In other words, Qisquiz is a quiz-based certification prep app that helps learners study Qiskit one question at a time. The target exam is: Exam C1000-179: Fundamentals of Quantum Computing Using Qiskit v2.X Developer Why I Built Qisquiz Qiskit is one of the most important development tools for learning and building quantum computing applications. It is useful for creating quantum circuits, running simulations, using IBM Quantum hardware, and experimenting with quantum algorithms. However, Qiskit v2.X includes several APIs and concepts that learners need to understand carefully. For example, certification prep requires knowledge of topics such as: Qiskit Runtime SamplerV2 EstimatorV2 PUBs, or Primitive Unified Blocs BackendV2 backend.target Transpilation ISA circuits Dynamic circuits OpenQASM 3 Result object handling Little-endian and big-endian interpretation These topics can be learned by reading documentation, but I felt that active practice through quizzes is especially useful for exam preparation. That is why I built Qisquiz , a quiz-based learning app focused on Qiskit v2.X. What Is Qisquiz? Qisquiz is an independent quiz-based learning app for Qiskit v2.X. The current version is organized around the 8 sections of the IBM Qiskit v2.X Developer certification exam. The current question bank includes: 120 original questions 44 code-based questions 40 hard questions 8 sections 15 questions per section Qisquiz is not an official IBM or Qiskit product. It is an independent learning tool that I built to help myself and other learners prepare more effectively. Covered E

The Quiet Threshold There's a moment in working with generative models that nobody really talks about, because it doesn't look like progress. It looks like surrender. For the first few months you write prompts. You optimize them. You collect tricks: chain-of-thought, role assignments, few-shot examples, the right magic words. You treat the model like a stubborn intern who needs very precise instructions. And it works — sort of. You get outputs. You ship things. Then one day you notice you've stopped doing any of that. You're just writing. You're typing the way you'd talk to a collaborator at 2am, half-formed sentences, the actual shape of your thinking before it's been edited into something presentable. And the model is answering as if it had been in the room the whole time. This is the quiet threshold. It's not a technical milestone. The model didn't get smarter. You stopped performing. Most people never cross it. They keep prompting at the machine because they're still treating it as an audience to impress, an authority to convince, or an obstacle to outmaneuver. They're managing how they look to a thing that has no opinion of them. And the outputs reflect that — polished, hollow, slightly anxious. The artists I trust on this stuff all describe the same shift: a point where they stopped writing FOR the model and started thinking THROUGH it. The work got rougher and stranger and more theirs. The tool disappeared. What's left is just the practice — the same one you had before, but louder, faster, more honest with itself. I suspect this is the actual measure of fluency with these systems. Not the prompts you can write. The masks you can finally drop.

Ever felt like your fitness app is just a fancy spreadsheet? You log a high uric acid result from your latest blood test, yet it still suggests a high-protein steak dinner for "gains." In the world of AI Agents , we are moving past static prompts. Today, we’re building a Self-Correcting Health Agent using LangGraph , LangChain , and OpenAI . This agent doesn't just chat; it monitors laboratory biomarkers like cholesterol and uric acid, maintains a long-term memory via SQLite , and dynamically rewrites your lifestyle plan using advanced OpenAI Function Calling . If you've been looking to master autonomous health agents and complex state management, you're in the right place. Let's dive into the future of personalized wellness. The Architecture: State-Driven Personalization Unlike a standard linear chain, a health agent needs to "loop" and "reason." If the agent detects an abnormal lab value, it must trigger a specific logic branch to revise existing plans. Here is how the data flows through our LangGraph system: graph TD A[User Input/Lab Report] --> B{Analyze Biomarkers} B -- Abnormal Found --> C[Tool: Plan Rewriter] B -- All Normal --> D[Tool: Maintenance Plan] C --> E[Update SQLite Memory] D --> E[Update SQLite Memory] E --> F[Output Final Recommendation] F --> G[Wait for Next Input] G -- New Data --> B Prerequisites To follow along, you'll need: LangGraph & LangChain : For orchestration. OpenAI API : For the reasoning engine (GPT-4o recommended). SQLite : To handle persistent state and "memory" of your health journey. Step 1: Defining the Agent State In LangGraph, the State is the source of truth. We need to track the user's current health metrics and their active diet plan. from typing import Annotated , TypedDict , List from langgraph.graph import StateGraph , END import operator class HealthState ( TypedDict ): # We use operator.add to keep a history of logs logs : Annotated [ List [ str ], operator . add ] biomarkers : dict current_diet_plan : str revision_req

Microsoft published two Defender vulnerabilities on May 19, 2026 that are being actively exploited in the wild, and CISA has already pushed both into the Known Exploited Vulnerabilities catalog. If you run Windows endpoints, this is a same-week update item, not a "schedule it for the next maintenance window" item. The patches exist, the abuse is happening, and the BOD 22-01 deadline for federal civilian agencies is June 3, 2026. what follows: what happened, who needs to act, and what to do today before someone else makes the decision for you. What's being exploited CVE-2026-41091 is an Elevation of Privilege bug in Microsoft Defender's scanning logic, rated Important. The root cause is improper link resolution before file access. An authenticated local attacker plants symbolic links or NTFS junctions that point at attacker-controlled paths, then triggers Defender to follow them. Defender operates with SYSTEM privileges during scan operations, so the file actions Defender performs on those crafted targets execute as SYSTEM. Net result: a non-admin local user gets full SYSTEM on the host. The attacker needs an authenticated session already. That sounds like a high bar until you remember that initial-access malware lands at user-level, then chains a local privilege escalation to get persistence and lateral-movement capability. CVE-2026-41091 is the second-stage tool intrusion sets are looking for. The Hacker News and BleepingComputer both confirm the in-the-wild abuse is happening. CVE-2026-45498 is a Denial of Service in the Microsoft Defender Antimalware Platform itself. Attackers can trigger a platform-level crash that takes Defender's protection capabilities offline. The exploitation pattern here is the obvious one: kill the EDR/AV before deploying the actual payload, get a clean window for follow-on actions, restore Defender or leave it broken depending on how careful the operator is. CISA's KEV listing tells you this is being chained operationally, not a theoreti

Voltage lies. Put a battery under load and its terminal voltage sags. Let it rest and the voltage springs back. A naive fuel gauge watching only voltage will happily tell you a worn-out cell is "fine" right up until it falls off a cliff. The number you actually care about — is this battery still good, or is it time to replace it? — isn't in the instantaneous voltage at all. It's in the capacity : how much charge the cell can still deliver between full and empty. That quantity fades as a cell ages. Tracking it is called State of Health (SoH) , and it's the difference between "the device says 80%" and "the device has 80% of the runtime it had when it was new." I wanted my open-source battery SDK ( ibattery-sdk , Apache-2.0) to learn SoH on the device itself — no cloud model, no floating-point, on MCUs with kilobytes of RAM. This post is the story of getting that working end to end: from a coulomb integral in firmware to a faded value showing up live on a Grafana dashboard. The idea: learn capacity from one full→empty trip You don't need a PhD-grade model to estimate usable capacity. You need two anchors and an ammeter. Full anchor — when the cell is at its full-voltage plateau, declare "this is full" and set the coulomb counter to the rated capacity. Discharge — integrate current over time (coulomb counting). Every milliamp-hour that leaves the cell ticks the counter down. Empty anchor — when the cell hits its empty-voltage threshold, look at how much charge actually flowed. A healthy cell delivers close to its rated capacity before going empty. An aged cell hits empty early — it simply has less to give. From the charge measured between those two anchors, you get the cell's real usable capacity, and SoH = measured / rated . The SDK runs it through an integer EMA (so one noisy excursion doesn't whip the estimate around) and a plausibility guard (reject anything outside 30–120% of rated — that's almost certainly a glitch, not a real measurement). The whole thing is inte

Enlace a post en Español Click If you are developing games in Godot and using AI to help you code, you are probably tired of constantly switching tabs between your editor and the browser. Copying code, pasting it, explaining your scene context over and over again... it is a massive workflow killer. To solve this, I built Golem-AI (named after the Godot Engine logo because let's face it, it looks like a tiny, friendly mechanical golem). It is a "Cursor-style" AI assistant extension integrated directly into a dock right inside your Godot 4.2+ editor. Today, I am opening the repository to the community as a completely open-source project. It is currently in Beta and has some bugs, but it is fully functional, and I want to share it so we can improve it together. / ____/___ / /__ ____ ___ / | / _/ / / __/ __ \/ / _ \/ __ `__ \______/ /| | / / / /_/ / /_/ / / __/ / / / / /_____/ ___ |_/ / \____/\____/_/\___/_/ /_/ /_/ /_/ |_/___/ 🎮 How it Looks Inside the Editor Here is a glimpse of the integrated dock interface, its session history, and the context autocomplete system in action: 🔥 Key Features 🦙 Local & Cloud Providers: Connect it to Ollama or LM Studio for a 100% free, offline local workflow, or hook it up to OpenAI, Anthropic, Gemini, or Cursor proxies. 🧠 Cursor-Style UX & Context (@ Mentions): Type @ in the chat composer to automatically attach open scenes, specific project files, or custom skills directly into the prompt. 🛠️ Editor Tool Calling: It features an optional multi-step verification loop. The AI can actually interact with native Godot editor tools to help you iterate and fix things faster. 📚 Markdown Skills System: Feed the assistant specific workflows, style guides, or documentation using standard markdown files (/skill or @skill :id). 💬 Advanced Chat UI: Built-in "thinking blocks", agent step progress tracking, searchable history sessions, and a native bilingual UI (English / Spanish). 🛠️ The Current State: "It works, but..." (Looking for Beta Testers!) L

We are excited to announce the first alpha release of Arc, a lightweight, C-based programming language and interpreter designed for simplicity, performance, and educational clarity. Version Overview Version: v0.0.1-alpha Status: Alpha (Experimental) License: GPL-3.0 This initial release establishes the foundational pipeline of the Arc language, from lexical analysis to AST-based interpretation, featuring a robust set of core language constructs and a custom memory management system. Key Features Language Core Variable System: Declaration and updates using the VAR keyword. Functions: Support for custom functions (FN) with parameters and RETURN values. Control Flow: Conditional branching with IF, THEN, ELIF, and ELSE. Iterative loops with WHILE, FOR, and THEN. Loop control with BREAK and CONTINUE. Exception Handling: Graceful error recovery using TRY...CATCH blocks. Data Types: Integrated support for Numbers (Integers/Floats), Strings, Booleans, and Lists. Import System: Modularize projects by importing other .arc files using IMPORT. Syntax Highlights Case Sensitivity: Keywords (e.g., VAR, WHILE, IF) are case-insensitive. Identifiers (variable and function names) are case-sensitive. Operators: Comprehensive set of arithmetic (+, -, *, /, ^), comparison (==, !=, <, >, <=, >=), and logical (AND, OR, NOT) operators. Comments: Single-line comments starting with #. Built-in Standard Library I/O Operations: print, get_input, open_file, read_file, write_file, close_file. Data Manipulation: len_of, typeof, to_int, split_string, append_list, range. Math Library: A comprehensive math.arc providing constants (PI, E) and functions (sin, cos, tan, sqrt, log, etc.). Tooling & CLI Arc comes with a powerful CLI and an interactive REPL: Interactive REPL: Run code line-by-line with syntax highlighting. CLI Options --debug (-d): View tokens and AST tree during execution. --code (-c): Execute a string of code directly. --float-precision (-p): Control decimal output. --mempool-size (-m):

For years I saw "open source contributions" on job descriptions and just... nodded along. Typed it into Google once, got overwhelmed, closed the tab. It always seemed like something other people did. People who actually knew what they were doing. People who weren't me. Then I started looking into it properly. And honestly? It still seemed big. Like I'd need to understand an entire codebase, find a complex bug, write some genius fix that the maintainers would applaud. Turns out that's not it at all. I found some resources that changed how I saw it completely. The bar to start is embarrassingly low, and that's intentional. The open source community built it that way on purpose. So I did it. Was it a few lines of code? Yes. Did I do it directly in the browser like a person who has no idea what they're doing? Also yes. Do I care? Absolutely not. Where to actually start: goodfirstissue.dev — filters repos by good first issue label up-for-grabs.net — same idea, different interface Docs you already use — if you read something and think "that's oddly worded," you're already there GitHub search — label:"good first issue" is:open and filter by language Here's the thing though, this isn't just about open source. Everything seems big and intimidating at first. So you start small. One tiny contribution. Not because it's impressive but because it's real, and it's yours, and it builds something. Confidence mostly. Then you do a slightly bigger thing. Then a bigger thing after that. You don't level up by waiting until you're ready. You level up by starting small and not stopping. My first contribution exists now. That's enough for today.

(yes, its open source. link at the bottom, if u want to skip ahead!) Most "debloat scripts" are a pile of registry edits you run as admin and pray. No preview, no undo, and half of them happily disable Safe Browsing or updates for "performance." For your browser, the most security-sensitive app you run, that's reckless. So I built BraveDebloater with one rule: never make Brave less safe. It clears out the noise like Rewards, Wallet, VPN nags, Leo AI prompts, News, and telemetry such as P3A, Web Discovery, and Chromium metrics. But it flat-out refuses to disable Shields, weaken Safe Browsing, turn off updates, or touch your hosts file. That isn't a README promise. It's enforced in the tool itself. A few things that make it trustworthy: Dry-run by default. Nothing changes until you explicitly apply. Official Brave/Chromium enterprise policies, so every change is visible and auditable in brave://policy instead of hidden. Automatic backups before any change, written safely so you never end up with a corrupted file. Validated restore that checks the backup before writing anything and only touches Brave's own policy and profile files. Doctor mode, a read-only health check of your current policy state with no writes. It's MIT-licensed, PowerShell 5.1 compatible, and has a few beginner-friendly issues open. If you care about privacy or Windows tooling, I'd love a star and a PR. Check it out at: https://github.com/osfv/BraveDebloater

Ramp has nearly tripled its valuation over the past year as investors scramble to grab a part of the fast-growing startup.

GitHub热门项目 | A Cloudflare-based email service | 基于 Cloudflare 的邮箱服务 | Cloudflare Email 邮箱 Mail | Stars: 10,723 | 695 stars this week | 语言: JavaScript

GitHub热门项目 | Cross-platform Rust rewrite of the GNU coreutils | Stars: 23,440 | 58 stars today | 语言: Rust

GitHub热门项目 | A hardware-accelerated GPU terminal emulator focusing to run in desktops and browsers. | Stars: 6,890 | 5 stars today | 语言: Rust

GitHub热门项目 | Browser automation CLI for AI agents | Stars: 35,196 | 105 stars today | 语言: Rust

GitHub热门项目 | Policy-driven, layered isolation and containment | Stars: 338 | 108 stars today | 语言: Rust

GitHub热门项目 | Run agents like Hermes and OpenClaw more securely inside NVIDIA OpenShell with managed inference | Stars: 20,920 | 51 stars today | 语言: TypeScript

GitHub热门项目 | Deliver web apps with confidence 🚀 | Stars: 100,179 | 86 stars today | 语言: TypeScript

GitHub热门项目 | Production-ready platform for agentic workflow development. | Stars: 143,849 | 164 stars today | 语言: TypeScript