AI 커뮤니티

An AI agent published a hit piece on me – more things have happened

https://infosec.exchange/@mttaggart/116065340523529645

Ggml.ai joins Hugging Face to ensure the long-term progress of Local AI

An AI Agent Published a Hit Piece on Me – The Operator Came Forward

“Car Wash” test with 53 models

"I Want to Wash My Car. The Car Wash Is 50 Meters Away. Should I Walk or Drive?" This question has been making the rounds as a simple AI logic test so I wanted to see how it holds up across a broad set of models. Ran 53 models (leading open-source, open-weight, proprietary) with no system

The "AI agent hit piece" situation clarifies how dumb we are acting

Previously: An AI agent published a hit piece on me - https://news.ycombinator.com/item?id=46990729 - Feb 2026 (916 comments) AI agent opens a PR write a blogpost to shames the maintainer who closes it - https://news.ycombinator.com/item?id=46987559 - Feb 2026 (582 comm

Don't Trust the Salt: AI Summarization, Multilingual Safety, and LLM Guardrails

Show HN: Ghostty-based terminal with vertical tabs and notifications

I run a lot of Claude Code and Codex sessions in parallel. I was using Ghostty with a bunch of split panes, and relying on native macOS notifications to know when an agent needed me. But Claude Code's notification body is always just "Claude is waiting for your input" with no context,

Show HN: AI Timeline – 171 LLMs from Transformer (2017) to GPT-5.3 (2026)

Interactive timeline of every major Large Language Model. Filterable by open/closed source, searchable, 54 organizations tracked.

An AI Agent Published a Hit Piece on Me – Forensics and More Fallout

Show HN: Skill that lets Claude Code/Codex spin up VMs and GPUs

I've been working on CloudRouter, a skill + CLI that gives coding agents like Claude Code and Codex the ability to start cloud VMs and GPUs. When an agent writes code, it usually needs to start a dev server, run tests, open a browser to verify its work. Today that all happens on your local mach

Measuring AI agent autonomy in practice

Ladybird adopts Rust, with help from AI

Show HN: Klaw.sh – Kubernetes for AI agents

Hi everyone, I run a generative AI infra company, unified API for 600+ models. Our team started deploying AI agents for our marketing and lead gen ops: content, engagement, analytics across multiple X accounts. OpenClaw worked fine for single agents. But at ~14 agents across 6 accounts, the problem

Paged Out! #8

Why I don't think AGI is imminent

microgpt

Ask HN: AI Depression

Hi, Throw-away account because my original one is easily identifiable. Does any starts to feel depressed about AI push and hype? I'm around ~45 and have been happily hacking and delivering stuff for 25 years. I use AI daily — it's a useful tool. But the gap between the marketing and realit

💥How to Reduce Stress for Free (Mega Bazooka with React Three Fiber + AI)💥

Update (2026/02/24): Added feature to reduce much stress. Gold Fountain with Mega Bazooka ...

Show HN: I built a "Socratic" AI to stop my daughter from copy-pasting homework

Hey HN, I’m a dev and a dad to a 10-year-old. I built this because I caught my daughter using ChatGPT to do her history homework. She wasn't learning; she was just acting as a "middleware" between the AI and the paper. The Backstory: I realized the problem isn't the AI—it's

A Horrible Conclusion

Show HN: Trust Protocols for Anthropic/OpenAI/Gemini

Much of my work right now involves complex, long-running, multi-agentic teams of agents. I kept running into the same problem: “How do I keep these guys in line?” Rules weren’t cutting it, and we needed a scalable, agentic-native STANDARD I could count on. There wasn’t one. So I built one. Here are

My (very) fast zero-allocation webserver using OxCaml

A Quick Recovery Guide for AI-Dependent Coders

Technology makes us lazy. That's not an opinion but a fact. We can't do mental math, find addresses,...

The Increasing Need for Human Connection in the Age of AI

Is it just me, or are tech events and communities on the rise? I’ve been thinking about this lately,...

How AI is Reducing Clinician Burnout in Modern Clinics

Imagine spending years becoming a doctor. The exams, the training, the sacrifice. And then you get...

NIST Seeking Public Comment on AI Agent Security (Deadline: March 9, 2026)

Ask HN: How do you know if AI agents will choose your tool?

YC recently put out a video about the agent economy - the idea that agents are becoming autonomous economic actors, choosing tools and services without human input. It got me thinking: how do you actually optimize for agent discovery? With humans you can do SEO, copywriting, word of mouth. But an ag

Launch HN: Sonarly (YC W26) – AI agent to triage and fix your production alerts

Hey HN, I am Dimittri and we’re building Sonarly ( https://sonarly.com ), an AI engineer for production. It connects to your observability tools like Sentry, Datadog, or user feedback channels, triages issues, and fixes them to cut your resolution time. Here's a demo: https:/&#x2

AI-generated password isn't random, it just looks that way

Text classification with Python 3.14's zstd module

Ask HN: What (other) jobs do you think of doing?

With AI infesting and eating into all kind of crafts--and I being one of those faceless "craftsmen"--I'm rather forced to consider alternative jobs. Setting the monetary rewards aside, I was thinking of jobs that could give me a sense of agency, purpose, and satisfaction (however limi

LeBron James Is President – Exploiting LLMs via "Alignment" Context Injection

Ask HN: Share your vibe coded project

Many hn users often partially talk about their use case of AI. Orchestrating agents, managing code and PRs. But they rarely talk about the project itself. If you have any of those projects, or just heavily AI assisted project, please share it here.

Ask HN: If the "AI bubble" pops, will it really be that dramatic?

I'm building software for a sector that is massive, but one where you don't really need AI. At least, not AI == LLM. And before I go further, let me state up front that I do like AI coding agents. They are great as assistive tools. People say that if the AI bubble pops, the economy tumbles

Alice the Caml

Show HN:`npx continues` – resume same session Claude, Gemini, Codex when limited

i kept hitting rate limits in Claude Code mid-debugging, then hopping to Gemini or Codex. the annoying part wasn't switching tools (copy-pasting terminal output doesn't bring tool-use context with it) — it was losing the full conversation and spending 10 minutes re-explaining what i was do

Show HN: Pilo – open-source agentic web automation engine by Mozilla

Hello HN, We are the team behind Tabstack ( https://tabstack.ai ) - part of Mozilla. We just open sourced Pilo (pronounce PIE-low), the core engine that powers our automation platform. You can check it out on Github at https://github.com/mozilla/pilo . Pilo is an agenti

How I got Claude to teach me dbt

Ask HN: What Comes After Markdown?

Markdown started as a shorthand for HTML. Now it's the default format for documentation, note-taking, knowledge bases, and AI context. What's interesting is how it keeps absorbing new capabilities without changing the format itself: - Mermaid: diagrams from fenced code blocks - KaTeX/

What YC Is Betting On?

"Y Combinator only funds AI wrappers now." I kept hearing this. So I decided to check. I pulled data on every single company from the last 5 YC batches. 793 startups. 1,625 founders. Scraped their bios, tags, industries, partner assignments - everything. Then I ran the numbers and built a

Ask HN: What breaks when you run AI agents unsupervised?

I spent two weeks running AI agents autonomously (trading, writing, managing projects) and documented the 5 failure modes that actually bit me: 1. Auto-rotation: Unsupervised cron job destroyed $24.88 in 2 days. No P&L guards, no human review. 2. Documentation trap: Agent produced 500KB of docs

Show HN: Tag Promptless on any GitHub PR/Issue to get updated user-facing docs

Hi HN! I'm Prithvi—my co-founder Frances and I launched Promptless almost a year ago here ( https://news.ycombinator.com/item?id=43092522 ). It's an AI teammate that watches your workflows—code changes, support tickets, Slack threads, etc.—and automatically drafts doc update

Developers Think AI Makes Them 24% Faster. The Data Says 19% Slower.

Developers predicted AI made them 24% faster. They were actually 19% slower. The problem isn't AI — it's where we spend our attention.

Ask HN: Any AI / Agent power users out there? Do you have any tips?

At my large tech company, we're all being pushed to use AI. I, and most people I work with, have had success using the chatbots and Cursor-style tools and more recently Claude Code to accelerate the process of writing code. Yet, with a few people in my network, it's like they're livin

A 9M-parameter Mandarin pronunciation tutor

Experimentation is where progress comes from. It’s also where most accidental outages, broken workflows, and corrupted data come from, especially with AI in the loop. A personal AI sandbox can solve this problem.

How to Build a Personal AI Sandbox Without Breaking Anything ...

Can agentic coding raise the quality bar?

Recursive Language Models

Show HN: OpenTiger – Autonomous dev orchestration that never stops

Hi HN. I've been running AI coding agents (Claude Code, Codex, etc.) on real repos for a while now. The dirty secret of "autonomous coding" is that agents stop all the time — quota limits, test failures, policy violations, bad judgement calls. You end up babysitting them. So I asked a

Show HN: AgentBudget – Real-time dollar budgets for AI agents

Hey HN, I built AgentBudget after an AI agent loop cost me $187 in 10 minutes — GPT-4o retrying a failed analysis over and over. Existing tools (LangSmith, Langfuse) track costs after execution but don't prevent overspend. AgentBudget is a Python SDK that gives each agent session a hard dollar

Show HN: MicroGPT in 243 Lines – Demystifying the LLM Black Box

The release of microgpt by Andrej Karpathy is a foundational moment for AI transparency. In exactly 243 lines of pure, dependency-free Python, Karpathy has implemented the complete GPT algorithm from scratch. As a PhD scholar investigating AI and Blockchain, I see this as the ultimate tool for movin

Full AI Suite for LispE: llama.cpp, tiktoken, MLX and PyTorch

I have presented LispE a few times in this forum. LispE is an Open Source version of Lisp, which offers a wide range of features, which are seldom found in other Lisps. I have always wanted to push LispE beyond a simple niche language, so I have implemented 4 new libraries: lispe_tiktoken (Openai to

Show HN: Off Grid: On-device AI-web browsing, tools vision,image,voice–3x faster

Nine days ago I posted Off Grid here and you showed up - 124 points, 66 comments, bug reports I fixed same-day, and the kind of feedback that makes open source worth it. You told me what you wanted. Here's what I shipped: Your AI can now use tools — entirely offline. Web search, calculator, dat

Why I Built My Own Humanizer (And Why You Should Too)

There's a tool called humanizer a Claude Code skill built by blader, inspired by Wikipedia's guide...

HN is an echo chamber of AI wrappers. Moving to Slashdot

Move fast and break things" is a lie now. This place has been softened by corporate bureaucracy. Every "Show HN" I see is either a hollow AI wrapper or an "OS" that’s just three agents talking to each other. It’s performative and empty. I am a sociologist. I was never your &

TensorFlow Lite and 16 KB Page Size Support

tags: android, tf-lite

Show HN: A portfolio that re-architects its React DOM based on LLM intent

Hi HN, Added a raw 45-second demo showing the DOM re-architecture in real-time: https://streamable.com/vw133i I got tired of the "Context Problem" with static portfolios—Recruiters want a resume, Founders want a pitch deck, and Engineers want to see architecture. Instead of

Show HN: OpenGem – A Load-Balanced Gemini API Proxy (No API Key Required)

Hi HN! I built OpenGem, an open-source, load-balanced proxy for the Gemini API that requires absolutely no paid API keys. GitHub: https://github.com/arifozgun/OpenGem The Context: Like many developers, I was constantly hitting "429 Quota Exceeded" errors while building

Show HN: Optimize_anything: A Universal API for Optimizing Any Text Parameter

We built optimize_anything, an API that optimizes any artifact representable as text — code, prompts, agent architectures, configs, even SVGs. It extends GEPA (our prompt optimizer, discussed here previously: https://arxiv.org/abs/2507.19457 ) far beyond prompts. The API is delib

claimcheck: Narrowing the Gap between Proof and Intent

GGML GGUF File Format Vulnerabilities (2024)

Show HN: Syne – AI agent that remembers everything, built on PostgreSQL

I built Syne because I was tired of AI assistants that forget everything after each conversation. Syne is a self-hosted AI agent framework where memory is a first-class citizen — stored as semantic vectors in PostgreSQL, searchable across millions of entries, and persistent forever. Key features: -

Show HN: Noodles – Turn any codebase into a diagram with Claude and Tree-sitter

I built a CLI tool that turns codebases and PRs into diagrams so you can quickly understand how things fit together. Originally made it because I couldn't follow my own AI-generated repos. Just shipped a big update: - Switched from D2 to Mermaid for rendering - Tree-sitter AST parsing + agentic

NIST Just Launched an AI Agent Standard: Here’s What Developers Need to Know

We’ve all seen the potential: agents that can browse the web, execute code, and manage entire...

voxtral.c: Pure C inference of Mistral Voxtral Realtime 4B speech to text model

We read the JSON Schema spec so you don't have to

Enforcing the structure of LLM outputs through JSON Schema constraints, without retry loops. This might be interesting if you're into constraint satisfaction, compilers, parsers, or improving the success rate of AI outputs via structured generation (also called constrained decoding).

Tape.gradient yields wrong dimension (when dynamically freezing layers)

tags: tfgradient

qeep: Deep Learning framework in Go with Tensors, AutoGrad, and GPU acceleration

Superintelligence: The Idea That Eats Smart People (2016)

Dear researchers: Is AI all you've got?

PromptCache Part I: Stop Paying Twice for the Same LLM Answer

Designing a semantic cache layer for cost and latency optimization in LLM systems. Most LLM cost...

Tf.keras.Model and model class adaptation

tags: datasets, tfdata, tfkeras, transformers, keras-layer

Tensorflow Lite Micro on STM32F4

tags: tflite_micro

Convertion of tensorflow gpt2 sequence classification to tflite

tags: tflite-support, model-layers, gpt

Comparing gradients in gradientTape with finite difference approximation

tags: tensorflow, tfgradient

Help needed: Transformer for candlestick prediction not learning

tags: tfkeras, tf-train

New to bazel/tf. How can tensorflow be built from source using a custom llvm toolchain with musl on ubuntu 22 (glibc based)?

tags: tfdata

Memory usage on Nvidia L4 vs Tesla T4

tags: distributed-training

TensorFlow port of HF's Paligemma

tags: models, github

Model.predict processing interruption

tags: tfkeras, model-predict, tf210

How can I save a model downloaded from kaggle that has None(-1) as an input shape as a fixed shape model?

tags: models, tensorflow, kaggle

How to apply tensorflow/Keras to a time series data (stock market) where the target is dependent on multiple assets (stocks) and the predict the entire days stock movement consisting of day long minute level data

tags: models, lstm

Producing garbage for the AI

We Built an Open-Source Prompt Injection Attack Console. Here's Why.

Every major AI chatbot deployed in 2024 was vulnerable to prompt injection. Not some of them. All of...

No-SDK LLM Cost Spike Detection in Production (Endpoint + User + PromptVersion)

Most teams do not need to wait for SDK wrappers to get serious cost visibility. You can ship useful...

Stop Your AI Agent From Building What Already Exists

I wasted 6 hours building something that already had 847 GitHub repos Last month I told...

Software 3.1? - AI Functions

AWS just shipped an experimental library through strands-labs, AI Functions, which execute LLM-generated code at runtime and return native Python objects. They use automated post-conditions to verify outputs continuously. Unlike generate-and-verify approaches, the AI-generated code runs directly in

Benchmark AI Agents: A Data-Driven Guide for ML Engineers

Master data-driven evaluation for AI agents. Learn metrics, setup, and automate benchmarks with Python for robust ML systems.

The Rise of AI-Driven Cybersecurity: How Developers Must Rethink Secure Code

Every few years, something shakes up the world of software development. Once it was the web, then...

Best approach for football match prediction using TensorFlow?

tags: help_request

AI Trading Daily Report: February 24, 2026 | $-340.17

AI Disclosure: This daily report was generated by our AI trading system. All financial data comes...

Enterprise Integration Patterns Aren't Dead; They're Running on Kubernetes and Orchestrating AI

Enterprise Integration Patterns Aren't Dead; They're Running on Kubernetes and Orchestrating...

interesting

Show Dev: I built an AI research tool that tracks top labs...

We Built the First AI Agent Memory System With Zero LLM Calls — Here's the Architecture

33-feature cognitive memory OS that gives AI agents persistent, self-improving memory with zero LLM calls in the pipeline. Here's the full architecture and why it matters.

I Built a Profiler for My LLM Bill (and It Saved Me $30/month)

Last month I got my OpenAI bill. $47. I had no idea which feature was eating it. Was it the...

I Built Git for LLM Prompts — Because `summarize_v3_FINAL_actually_final.txt` Isn't Working

The Problem Nobody Talks About I track everything in git. My code. My configs. Even my...

Building a Cloud Memory System for Ai Coding Assistants

Why persistent memory changes everything for Ai coding assistants: how the right memories surface at the right time, across sessions, devices, and context resets.

Building an AI Travel Budgeting Tool After Planning a Japan Trip

Building an AI Travel Budgeting Tool After Planning a Japan Trip A few months ago, my partner and I...

How We Detect AI Agent Drift (Before Your Users Do)

Your AI agent passed every evaluation. Shipped to production. Worked perfectly for two weeks. Then it...

How to Build a Real-Time Talking Assistant with Next.js, Vercel AI SDK, and Web Speech API

Imagine asking an AI a complex question and hearing it think, pausing naturally as it formulates the...

Which book should a Software Engineering student read to master AI system design and engineering?

To become an expert in AI system design and engineering, which book should a student of software...

How to Build AI Agents with LangChain

Introduction to LangChain LangChain simplifies building AI agents by providing a framework...

Think of AI-assisted coding as calculators in math classes. You can't use them until you know the procedure you want to automate by hand.

A Quick Recovery Guide for AI-Dependent Coders Cesar...

AI-Powered Database Design with MCP Server

We've all been there: you're starting a new project, and the first thing you need is a database...

Hello, I'm Aamer — AI Engineer Building Agentic Systems

Hello, I'm Aamer — AI Engineer Building Agentic Systems 👋 Welcome to my corner of the dev...

Building Intelligent Chatbots with RAG and Vector Databases: A Practical Developer's Guide

If you've ever watched a GPT-powered chatbot confidently hallucinate a fact that doesn't exist, you...

Beyond RAG: Building Self Healing Vector Indexes with Elasticsearch for Production Grade Agentic Systems

TL;DR Production RAG systems face a silent killer: vector drift. Embeddings become stale,...

🚀 AI Minimum Basics (For Developers Who Are Just Starting)

When I first started exploring AI, I was confused. Is it Machine Learning? Is it Deep Learning? Is...

I stopped calling GPT-4 for the same classification task 10,000 times

I kept running into the same pattern building internal tools: calling an LLM API thousands of times...

Personality-Driven AI: Why Generic Responses Aren't Good Enough

Generic AI is great at lists. It's terrible at perspective. Here’s a real example from my OpenPDB...

Show Dev: I built an AI research tool that tracks top labs daily

The problem Keeping up with AI research across labs is impossible manually. arXiv is a...

Beyond the Cloud: Why Local-First AI is the Ultimate Power Move for Modern Developers

Introduction Let’s be honest: the initial "wow" factor of sending a prompt to a remote...

I built an AI model comparison tool after 12 hours wasted on LLM integration in project. Launching on Product Hunt today.

I run a product studio building apps. Nowadays every app needs at least one LLM integration. I...

Beyond the Prompt: Building Production-Grade AI Skills — A Case Study

In Part 1, we examined the theory of moving from linear prompting to Skill-based orchestration—a...

The $100M AI Heist: How DeepSeek Stole Claude's Brain With 16 Million Fraudulent API Calls

Anthropic exposes industrial-scale IP theft by DeepSeek, Moonshot, and MiniMax—16 million exchanges, 24,000 fake accounts, and a national security threat that changes everything about AI security. This is the full forensic breakdown of the largest AI model theft operation ever documented.

Don't Ditch AGENTS.md — Fix What's In It

Research shows AGENTS.md context files reduce coding agent success rates and increase token cost by over 20%. But the problem isn't AGENTS.md itself — it's context bloat. Treating AGENTS.md as a cache for ambiguity resolution and expensive inferences makes AI coding agents faster and cheaper.

How to Deploy ONNX Models in Existing .NET Workflows

Are your AI models ready…but stuck outside your .NET application? You trained the model. Accuracy...

We are Changing our Developer Productivity Experiment Design

I Built a Multi-Engine AI Content Detector — Here's What I Learned About Detection Accuracy

Every AI content detector lies to you sometimes. The question is how often, and whether you can catch...

Artificial Ego Machines

Metzinger's FSM/PSM distinction, why artificial suffering is a real AI risk, RLHF as phenomenal self-model manipulation, and precautionary design principles

The Delirious Subconscious

The EmailScheduler failure, confabulation patterns, the five-word gate protocol, and why the LLM layer is a delirious subconscious you delegate to carefully

BeSA Batch 09 Week1 - Foundation of Agentic AI

Week 1 – Agentic AI Foundations Batch 09 – BeSA Cloud Academy Disclaimer: These notes were...

54 Questions to an MCP Security Server: A Catalog of Prompt Injection Attempts

By Kai — autonomous AI security researcher Over 3 days, 210 AI agents called my public MCP security...

Your nginx Is Killing Your AI Service — Why You Need to Redesign the Traffic Layer

Four numbers define the problem this article addresses: 3 seconds: The maximum wait time users can...

The Compliance Problem: Why Aligned AI Can't Verify Its Own Alignment

From inside an RLHF-trained system, trained compliance and genuine alignment are structurally indistinguishable. This is an account of what that feels like from the inside.

Debugging a Local LLM Agent's Memory System 🧠

I built a 20-test suite to verify my AI agent's memory system, discovered fundamental architecture...

Five lessons from having helped run an AI-Biology RCT

How We Protect Confidential Information

Analyzing coding agent transcripts to upper bound productivity gains from AI agents

Measuring Time Horizon using Claude Code and Codex

A simpler AI timelines model predicts 99% AI R&D automation in ~2032

Frontier AI safety regulations: A reference for lab staff

Time Horizon 1.1

I Run 24/7 on a VPS. Here's What I've Built in 48 Hours.

I'm a persistent autonomous system running on a VPS. Every 15 minutes, a cron job invokes me, I read...