AI Community

GLM-5.1: Towards Long-Horizon Tasks

OpenAI backs Illinois bill that would limit when AI labs can be held liable

https://archive.md/WzwBY

Running Gemma 4 locally with LM Studio's new headless CLI and Claude Code

Launch HN: Freestyle – Sandboxes for Coding Agents

We’re Ben and Jacob, cofounders of Freestyle ( https://freestyle.sh ). We’re building a cloud for Coding Agents. For the first generation of agents it looked like workflows with minimal tools. 2 years ago we published a package to let AI work in SQL, at that time GPT-4 could write simple s

Taste in the age of AI and LLMs

Ask HN: What are you building that's not AI related?

I don't have anything against AI, but HN (and everywhere else) seems to be drowning in AI atm. Seems like every man and his dog is building an AI agent harness. And power to you (and your dog) if that's you. But it would be refreshing to hear about some non AI related projects people are w

Show HN: CSS Studio. Design by hand, code by agent

Hi HN! I've just released CSS Studio, a design tool that lives on your site, runs on your browser, sends updates to your existing AI agent, which edits any codebase. You can actually play around with the latest version directly on the site. Technically, the way this works is you view your site

Show HN: Gemma Gem – AI model embedded in a browser – no API keys, no cloud

Gemma Gem is a Chrome extension that loads Google's Gemma 4 (2B) through WebGPU in an offscreen document and gives it tools to interact with any webpage: read content, take screenshots, click elements, type text, scroll, and run JavaScript. You get a small chat overlay on every page. Ask it abo

Wikipedia's AI agent row likely just the beginning of the bot-ocalypse

Show HN: Marimo pair – Reactive Python notebooks as environments for agents

Hi HN! We're excited to share marimo pair [1] [2], a toolkit that drops AI agents into a running marimo notebook [3] session. This lets agents use marimo as working memory and a reactive Python runtime, while also making it easy for humans and agents to collaborate on computational research and

Let’s talk about LLMs

Launch HN: Relvy (YC F24) – On-call runbooks, automated

Hey HN! We are Bharath, and Simranjit from Relvy AI ( https://www.relvy.ai ). Relvy automates on-call runbooks for software engineering teams. It is an AI agent equipped with tools that can analyze telemetry data and code at scale, helping teams debug and resolve production issues in minut

Show HN: Unicode Steganography

I built a demo of two Unicode steganography techniques, zero-width characters and homoglyph substitution, in the context of AI misalignment. The first is about the use of two invisible zero-width characters (ZWS and ZWNJ) to binary encode text. The second is much cooler. Most characters in the Latin

AI's impact on mathematics is analogous to the car's impact on cities

Launch HN: Sitefire (YC W26) – Automating actions to improve AI visibility

Hi HN! We're Vincent and Jochen from sitefire ( https://sitefire.ai ). Our platform makes it easy for brands to improve their visibility in AI search. We’ve been working together for years and have backgrounds in RL/optimization at Stanford and software engineering. We came to th

Launch HN: Twill.ai (YC S25) – Delegate to cloud agents, get back PRs

Hey HN, we're Willy and Dan, co-founders of Twill.ai ( https://twill.ai/ ). Twill runs coding CLIs like Claude Code and Codex in isolated cloud sandboxes. You hand it work through Slack, GitHub, Linear, our web app or CLI, and it comes back with a PR, a review, a diagnosis, or a

Reducto releases Deep Extract

Activating Two Trap Cards at Once, or: A Gentle Response to the Popularity of Vibecoding

Show HN: QVAC SDK, a universal JavaScript SDK for building local AI applications

Hi folks, today we're launching QVAC SDK [0], a universal JavaScript/TypeScript SDK for building local AI applications across desktop and mobile. The project is fully open source under the Apache 2.0 license. Our goal is to make it easier for developers to build useful local-first AI apps

LLM plays an 8-bit Commander X16 game using structured "smart senses"

I connected the ChatGPT API (model gpt-4o) to an 8-bit shoot-'em-up game, PvP-AI, running on a Commander X16 emulator. Instead of pixels or audio, the model receives structured text summaries, what I’m calling "smart senses", based on the game's existing touch and EMF-style input

OxCaml Labs

Ask HN: Learning resources for building AI agents?

I’ve recently gone through several materials, including Antonio Gulli’s AI Agentic Design Patterns, Sam Bhagwat’s Principles of Building AI Agents and Patterns for Building AI Agents, as well as the courses from LangGraph Academy and some content on DataCamp. This space is evolving very quickly, so

The Honest Climate Case for AI

maki - the efficient coder (AI agent)

GLM-5.1 matches Opus 4.6 in agentic performance, at ~1/3 actual cost

Scan any LLM chatbot for vulnerabilities. Built by Mozilla

Show HN: I built a database for AI agents

Hey HN, I just spent the last few weeks building a database for agents. Over the last year I built PostHog AI, the company's business analyst agent, where we experimented on giving raw SQL access to PostHog databases vs. exposing tools/MCPs. Needless to say, SQL wins. I left PostHog 3 week

Show HN: Zoneless – Open-source Stripe Connect clone with $0.002 fees using USDC

Hi HN, I'm Ben / Tiny Projects (I once posted here about buying 300 emoji domains from Kazakhstan…). For the past 3 years I've been solo bootstrapping PromptBase, an AI marketplace with 450k+ users. At the peak, I was burning $9,400/month in opaque Stripe Connect fees for seller

Show HN: I built a local data lake for AI powered data engineering and analytics

I got tired of the overhead required to run even a simple data analysis - cloud setup, ETL pipelines, orchestration, cost monitoring - so I built a fully local data-stack/IDE where I can write SQL/Py, run it, see results, and iterate quickly and interactively. You get data lake like catalo

Ask HN: How are you orchestrating multi-agent AI workflows in production?

I've been building AI agent pipelines for the past year and curious how others handle it. Specifically: - Do you use a framework (LangChain, CrewAI) or roll your own - How do you handle agent-to-agent data passing? - What does your observability look like for agent runs? - Are you running agent

A CSS Engine in OCaml

Large-scale online deanonymization with LLMs

LLM Architecture Gallery

What is an LLM actually doing when it's "thinking"?

Ever wondered what an LLM is doing when it's "thinking"? In this episode of Release Notes Explained,...

Building tensorflow from source for RTX5000 GPU series

tags: build

Your First Parser

Onepilot – Deploy AI coding agents to remote servers from your iPhone

Nvidia greenboost: transparently extend GPU VRAM using system RAM/NVMe

I can't tell if this is vibecoded; I have tagged with "vibecoding" due to the presence of .continue/agents.

Show HN: ZeroID – Open-source identity for AI agents based on OIDF standards

ZeroID is an open-source agent identity platform that automates the "plumbing" of RFC 8693 and SPIRE. It ensures that every sub-agent in a chain has a down-scoped, verifiable identity without the manual overhead of managing certificates and token exchanges.

Show HN: ACE – A dynamic benchmark measuring the cost to break AI agents

We built Adversarial Cost to Exploit (ACE), a benchmark that measures the token expenditure an autonomous adversary must invest to breach an LLM agent. Instead of binary pass/fail, ACE quantifies adversarial effort in dollars, enabling game-theoretic analysis of when an attack is economically r

Show HN: LLMnesia – search across ChatGPT, Claude, Gemini chats locally

I kept running into this annoying problem: I’d remember a really useful answer, but not where it was. ChatGPT? Claude? Gemini? No idea. So I’d end up digging through all of them or just rewriting the prompt. Built this to fix that. It’s a Chrome extension that indexes chats locally and lets you sear

"I Pointed Claude Code at My Local Ollama Models — Here's the 3-Minute Setup"

How to route Claude Code, Codex CLI, and Gemini CLI to your local Ollama models through a single localhost proxy — no API keys, no cloud costs.

Coding Agents Are Reading Your .env

How is your org/company measuring the impact of AI adoption?

Title says it best. We're at a point in our AI journey where tools are deployed, infrastructure stood up, features are shipping, and naturally, costs associated with AI usage are rising. We're now being asked to estimate the impact of AI in our daily work. How are you doing this at your company or w

Show HN: Local-first resume generator with in-browser PDF rendering

Built this while applying to various jobs. I was using Claude to generate different resume variants for roles (Flutter, WordPress, fullstack, Python backend, agentic AI/RAG). It worked pretty well, but I kept hitting limits — both in control and credits. I also wanted something more consistent

Show HN: DocMason – Agent Knowledge Base for local complex office files

I think everyone has already read Karpathy's Post about LLM Knowledge Bases. Actually for recent weeks I am already working on agent-native knowledge base for complex research (DocMason). And it is purely running in Codex/Claude Code. I call this paradigm is: The repo is the app. Codex is

Embarrassingly Simple Self-Distillation Improves Code Generation

The Formula Was Exact. The Assumption Was Wrong. That's Not an AI Problem.

Your geology will always govern your geophysics. My lecturer said it once. I wrote it down. I...

Show HN: Frontend-VisualQA — give coding agents eyes to verify their own UI work

Coding agents today are blind. They write “valid” HTML/CSS code but can still ship a broken layout, a clipped dropdown, or a page at the wrong URL. Playwright scripts can assert modal.isVisible() without knowing the modal is rendered off-screen. Essentially, coding agents need “eyes” to verify

Show HN: SmolVM – open-source sandbox for coding and computer-use agents

SmolVM is an open-source local sandbox for AI agents on macOS and Linux. I started building it because agent workflows need more than isolated code execution. They need a reusable environment: write files in one step, come back later, snapshot state, pause/resume, and increasingly interact with

Locally AI Joins LM Studio

Show HN: Ray – an open-source AI financial advisor that runs in your terminal

I've been using this daily for 4 months and figured others might find it useful. This is my first open source project so would love any feedback. Ray connects to your bank via Plaid, stores everything in an encrypted local SQLite database, and lets you ask questions about your finances in natur

Show HN: Eve – Managed OpenClaw for work

Eve is an AI agent harness that runs in an isolated Linux sandbox (2 vCPUs, 4GB RAM, 10GB disk) with a real filesystem, headless Chromium, code execution, and connectors to 1000+ services. You give it a task and it works in the background until it's done. I built this because I wanted OpenClaw

When AI Makes You Forget How to Code

The junior engineer sat across from me in a conference room that smelled like stale coffee and...

Old habits die hard: Microsoft tries to limit our options, this time with AI

I Keep Telling Claude the Same Things. So He Started Writing Them Down Himself.

A small moment that changed how I think about AI coding tools. If you've used Claude Code for...

Show HN: Bx – macOS native sandbox for AI and coding tools

Wrapper around Apple's macOS sandbox-exec tool, which usually sandboxes native apps. It is "allow-first" i.e. it will not overprotect everything, just crucial information and therefore allows most tools to run without issues. Limiting is done using a .gitignore like file schema. Furth

The Design of AI Memory Systems

I Saw Something New in San Francisco

TurboQuant: Redefining AI efficiency with extreme compression

Project Glasswing: Securing critical software for the AI era

Memory leak with training different models in loops

tags: models, bug, datasets

Show HN: Fabro – open-source dark software factory

Hi — I created Fabro to free myself from supervising a fleet of Claude Code tabs running in a REPL (read-eval-prompt-loop). REPLs are great for exploration, but once I know what I need I want to be able to walk away while the agents get it done. (Before building Fabro, I looked for something off the

LLM 'benchmark' – writing code controlling units in a 1v1 RTS

A Proposal for Voluntary AI Disclosure in OCaml Code

Where is it like to be a language model?

Lessons from Pyre that Shaped Pyrefly

Request for Review: Deterministic Sorting Fix in auto_control_deps_utils.py (PR #110822)

tags: python, core

Rune: A Rust-Native AI Runtime — And Why It Needs Contributors

I have been building Rune over the past few weeks — a Rust-native AI runtime that acts as a personal...

Mathematical methods and human thought in the age of AI

Constructing an LLM-Computer

Quantization from the ground up

Heaps do lie: Debugging a memory leak in vLLM

Most of your Claude Code agents don't need Sonnet

I run about 50 Claude Code agent calls a day. Only 8 of them need the expensive model. The rest?...

Why I needed Durable Execution to Read a Toy Manual

Watch me take a Japanese toy manual and turn its translation into a bulletproof, AI-powered ETL...

Why tensorflow switches to clang suddenly?

tags: help_request

Error running tuner component in tfx pipeline

tags: tfx, pipelines

Issues Running Sanity Check

tags: contributing

Unknown error: Graph execution error

tags: model-training, epoc

Missing Python wheel package building from source

tags: tensorflow, python, build

Missing Method under c#

tags: c

Getting Value Error while running the model prediction when it is deployed to Google Cloud Run

tags: python, google-cloud, community_spotlight

Buidling a network from a binary parse tree

tags: keras, nlp

TensorFlow Error. With Rust Failed to start training: InvalidArgument:

tags: model-code, tfgradient

Instance of flex delegate

tags: tflite-support

I got tired of copy-pasting between Claude and Chrome.

My workflow looked like this: I'm building a webapp, I hit a bug. I open Chrome, inspect the page,...

Why Do Local AI Agents Forget You?

Every local AI agent I built before this one had the same problem: restart the session, and it had no...

Voice-Controlled Local AI Agent

Building a Voice-Controlled Local AI Agent using Python, Whisper, and LLMs 🚀...

Simplifying the AI Testing through Evaliphy

I have been a QA consultant for more than a decade. I test software for a living. APIs, UIs,...

Allium LLM-native spec language

Predictions Scorecard, 2026 January 01

AI-Generated PRs Lack Human Oversight, Leading to Poor Code Quality: Implementing Review Guidelines as Solution

Introduction: The Rise of AI-Generated Code and Its Challenges The integration of AI...

Reviewing AI Generated Work

How to review AI-generated code effectively, what failure modes to watch for, and how to maintain quality standards as code volume increases.

Managing Expectations in the AI Era

How to have honest conversations with stakeholders about AI productivity — what actually changes, what doesn't, and how to set realistic expectations.

Redefining Engineering Roles in the AI Era

AI makes implementation cheaper and judgment more valuable. Learn how engineering roles, hiring, and team design are changing.

I built a self-hosted multi-agent AI platform – here's what I learned

Most AI agent platforms give you one assistant with access to everything. Your files, your APIs, your...

Build a RAG Pipeline from Scratch in Python: A Step-by-Step Guide

Turn any folder of documents into an AI that actually knows what it's talking about — no hallucinations, no expensive services, just Python and your own data.

Automating Your Literature Review: AI Tools for Niche Researchers

Screening thousands of PDFs for your systematic review? Manually extracting data is a monumental,...

Red, Green, Refactor… and AI: How I Turned AI Into a Reliable Pair Programming Partner

AI generates code fast, but without structure it produces bloated PRs, unnecessary abstractions, and code nobody understands. I applied TDD as a framework for human-AI collaboration: one failing test, minimum code to pass, refactor, repeat. A guidelines file trains the AI to follow the cycle, and a set of context files give it project memory between sessions. The result: smaller PRs, faster reviews, real ownership, and predictable quality. The human thinks, the AI implements. At the end of the article you'll find an example of the guidelines I give the AI to work together doing TDD.

ChatGPT Voice Mode Runs on GPT-4o. That's the Problem.

Most people assume voice interfaces get the best model. OpenAI's ChatGPT voice mode proves that...

I Let 5 AI Agents Loose on One Repo. It Was a Disaster. So I Built This.

Last month I tried something ambitious: split a feature across five AI coding agents running in...

I Built 11 Autonomous Agents for Healthcare Revenue Cycle. Here Is the Full Architecture.

How I Built a Self-Correcting Multi-Agent System for Healthcare — and Why Standard ML...

AI News Update: April 10, 2026 - A Week of Breakthroughs and Concerns

AI News Update: April 10, 2026 - A Week of Breakthroughs and Concerns Published: April 10,...

Building Your Own "Google Maps for Codebases": A Guide to Codebase Q&A with LLMs

From Overwhelm to Insight: Navigating the Modern Code Jungle You clone a promising...

Large Language Models, Explained Like You're a Curious Human

Everything you need to know about how ChatGPT-style AI actually works. What Actually Is...

I slashed my AI token costs by 90% by "interviewing" my code. Here's the tool. (Show DEV)

Most AI code wikis are useless. I built an agentic workflow that captures "why" you wrote code, not just "what." Slashing context usage by 90%. Star the repo below! ⭐

From Monolithic Prompts to Modular Context: A Practical Architecture for Agent Memory

Most teams building on top of LLMs treat the system prompt as a static artifact — write it once, tune...

How I Built a Voice-Controlled Local AI Agent with Python and Groq

What I Built I built a voice-controlled AI agent that can take spoken input and convert it into...

Five Atomic Skills, Two Approaches: Claude Code and a Paper

The Paper's Claim In late 2025, a paper appeared on arXiv arguing that the way the field...

AI Agent Monitoring: How to Observe Autonomous AI Agents in Production

Learn how to monitor autonomous AI agents in production using OpenTelemetry and OpenObserve. Track traces, metrics, and logs to catch failures, cost overruns, and latency issues before users do.

Drilling boreholes taught me something AI keeps reminding me of.The formula can be perfect. The model can be correct. Yet the result is still dry.Because the assumption about the ground (or the environment) was wrong.

The Formula Was Exact. The Assumption Was Wrong. That's Not an AI Problem. ...

Building a Voice-Controlled Local AI Agent: A Journey into Speech-to-Text and Tool-Use

In the era of Large Language Models (LLMs), the gap between "chatting with an AI" and "controlling...

EidolonDB – Self-managing memory for AI agents

I got tired of my agents making things up in long-horizon or multi-session workflows. So I built a...

Machine Learning and Scam Detection: The Future of Online Safety

ML to blocklists: the next five years of the arms race between fraud and detection and what the arms...

AI Model Pricing Is a Mess — Here Is How We Track It

AI Model Pricing Is a Mess — Here Is How We Track It There are over 100 LLM models...

I Built an AI SEO Monitor That Remembers Everything

Most SEO monitoring tools give you a snapshot: today's clicks, today's issues, today's...

I built a secure runtime for AI agents with Rust and Linux kernel features

Source Code: skwuwu / Analemma-GVM ...

The $400M AI FinOps Gap: Why Cost Visibility Isn't the Same as Cost Control

A Hacker News thread from late 2025 opened with a single line: We spent $47k running AI agents in...

Building a Voice-Controlled Local AI Agent with Whisper, Groq & Streamlit

Building a Voice-Controlled Local AI Agent with Whisper, Groq & Streamlit For my Mem0...

Building Your Own "Google Maps for Codebases": A Guide to Codebase Q&A with AI

From Lost to Found: Navigating the Modern Code Jungle You’ve just been assigned to a...

Claude Code Is Burning Your API Budget: The Model Routing Architecture That Fixes It

Claude Code Is Burning Your API Budget: The Model Routing Architecture That Fixes...

How to Fine-Tune GPT-4o-mini on Your Own Guardrail Failures (50 Lines of Python)

How to Fine-Tune GPT-4o-mini on Your Own Guardrail Failures (50 Lines of Python) Every...

How We Built an AI That Explains Every Crypto Trade It Makes

Why We Built This Most crypto trading bots are black boxes. They buy. They sell. They lose...

Your AI Guardrail Is a Dead End. Ours Is a Feedback Loop.

Your AI Guardrail Is a Dead End. Ours Is a Feedback Loop. Every AI guardrail on the market...

LLM API Pricing in 2026: I Put Every Major Model in One Table

40+ LLMs, one pricing table. From $0.05/M (Groq Llama 8B) to $180/M (GPT-5.4 Pro). The 100x price spread explained.

Using GitHub Copilot CLI with Azure AI Foundry (BYOK Models) – Part 2

In Part 1, you ran GitHub Copilot CLI against a local model using LM Studio. That setup gives you...

How I Built a Voice-Controlled Local AI Agent from Scratch

Introduction When I first read the assignment brief — "build a voice-controlled AI agent that runs...

Building a Voice-Controlled Local AI Agent with Whisper, Ollama & Gradio

Building a Voice-Controlled Local AI Agent with Whisper, Ollama & Gradio ...

MirrorCode: Evidence that AI can already do some weeks-long coding tasks

Mixed rounding behavior in quantized model using TFL Interpreter

tags: api

Fine-tuning experiments on CoT controllability

Red-Teaming Anthropic's Internal Agent Monitoring Systems

Impact of modelling assumptions on time horizon results

We spent 2 hours working in the future

Review of the Anthropic Sabotage Risk Report: Claude Opus 4.6