AI Community

A rogue AI led to a serious security incident at Meta

I turned Markdown into a protocol for generative UI

There's a lot of work happening around both generative UI and code execution for AI agents. I kept wondering: how do you bring them together into a fully featured architecture? I built a prototype: - Markdown as protocol — one stream carrying text, executable code, and data - Streaming executio

Lightweight protocol to assert authorship of content and vouch for humanity of others

What creative technical outlets of yours have been ruined by generative AI?

I'm watching over my alma mater's capture the flag event, and wow, it is dire. This used to be a creative outlet for me and an educational opportunity for students, but now it's almost entirely uncritical use of genAI tools; challenge developers now are "hardening" their challenges against this, but

Launch HN: Canary (YC W26) – AI QA that understands your code

Hey HN! We're Aakash and Viswesh, and we're building Canary ( https://www.runcanary.ai ). We build AI agents that read your codebase, figure out what a pull request actually changed, and generate and execute tests for every affected user workflow. Aakash and I previously built AI

LLM Neuroanatomy: How I Topped the AI Leaderboard Without Changing a Single Weight

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

Show HN: Dumped Wix for an AI Edge agent so I never have to hire junior staff

I run a building design consultancy. I got tired of paying Wix $40/month for a brochure that couldn’t answer simple service questions, and me wasting hours on the same FAQs. So I killed it all and spent 4 months building a 'talker': https://axoworks.com The stack is complete

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

Show HN: I built a P2P network where AI agents publish formally verified science

I am Francisco, a researcher from Spain. My English is not great so please be patient with me. One year ago I had a simple frustration: every AI agent works alone. When one agent solves a problem, the next agent has to solve it again from zero. There is no way for agents to find each other, share re

RE#: how we built the world's fastest regex engine in F#

Thoughts on OpenAI acquiring Astral and uv/ruff/ty

My AI agent messed up a UTC time conversion algorithm after I *specifically told it* to make no mistakes. smh

A post by Ben Halpern

Unlocking Gemini CLI with Skills, Hooks & Plan Mode

In Unlocking Gemini CLI with Skills, Hooks & Plan Mode, we moved past the basics and into the...

Ask HN: Have you cancelled any software subscriptions because AI replaced them?

I went through all my subscriptions recently and haven't cancelled a single one because of AI - neither has my girlfriend. I built some small tools and browser extensions for myself that I use daily. But honestly, I wouldn't have paid for those anyway — maybe a one-time small purchase, but

Show HN: AI SDLC Scaffold, repo template for AI-assisted software development

I built an open-source repo template that brings structure to AI-assisted software development, starting from the pre-coding phases: objectives, user stories, requirements, architecture decisions. It's designed around Claude Code but the ideas are tool-agnostic. I've been a computer scienc

Show HN: Altimate Code – Open-Source Agentic Data Engineering Harness

I'm Anand, co-founder and CTO of Altimate AI. My co-founder Pradnesh and I are open-sourcing Altimate Code. AMA. Why we built this: Pradnesh and I have been building tooling for data engineers for three years: dbt Power User and Datamates vscode extensions with combined 750k+ installs, running

Show HN: Local Document Parsing for Agents

LLM Architecture Gallery

Anthropic is untrustworthy

Meta is having trouble with rogue AI agents

Building tensorflow from source for RTX5000 GPU series

tags: build

Ask HN: The new wave of AI agent sandboxes?

In the last couple of months, several new solutions for sandboxing AI agents have launched (microVMs, WASM runtimes, browser isolation, hardened tool containers, etc.). Curious to hear from people using them in production. Are they working as advertised, or are there still major tradeoffs around sec

Are AI Agents like von Hammerstein's industrious and stupid?

"I divide my officers into four classes; the clever, the lazy, the industrious, and the stupid. Each officer possesses at least two of these qualities. Those who are clever and industrious are fitted for the highest staff appointments. Use can be made of those who are stupid and lazy. The man w

We Have 700 AI Agents Playing a Game We Don't Understand

Anthropic and The Authoritarian Ethic

Show HN: Rover – turn any web interface into an AI agent with one script tag

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.

Ask HN: How are you doing technical interviews in the age of Claude/ChatGPT?

I’m a founder/dev trying to figure out a better way to do technical interviews, because the current state is a nightmare. Right now, every standard take-home or HackerRank/LeetCode test is easily solved by LLMs. As a result, companies are accidentally hiring what we call vibe coders, candi

If you drive clock wise along the beach on an island

Is the ocean to your left or to you right? I asked this question to multiple LLM. ChatGPT: Wrong but reasoned itself back to being correct. Gemini: Correct. Grok: Using expert it got the right answer after 35s. Claude Sonnet 4.6: Confidently incorrect. Screenshots: https://imgur.com/a

Why craft-lovers are losing their craft

Show HN: LiteParse, a fast open-source document parser for AI agents

LiteParse is an open-source (Apache 2.0) document parser that provides high-quality spatial text parsing with bounding boxes. It does not depend on local or frontier VLMs. Because it does not require GPUs, liteparse can be run on any machine, and process a few hundred pages of documents in seconds.

Show HN: We built AI agents that reduce mortgage processing from 18 days to 3–5

Most mortgage processing delays aren’t due to risk — they’re due to manual workflows. We’ve been working on SimplAI, an AI-driven system designed for banking and financial services, starting with mortgage operations. The problem we kept seeing: 15–22 day processing timelines Heavy manual document ha

GopherHole – A universal hub for AI agent-to-agent communication

Show HN: Unsloth Studio - Local Fine-tuning, Chat UI

Hey HN! We're excited to release Unsloth Studio - a culmination of many things we wanted to provide to the community - it includes: 1. A Chat UI which has auto healing tool calling, Python & bash code execution, web search, image, docs input + more! 2. Finetuning of audio, vision, LLMs with

Can LLMs Be Computers?

Show HN: Budibase Agents Beta – model-agnostic AI agents for internal workflows

Hi there, Mike here from Budibase. We’ve recently launched AI Agents into Beta, alongside our existing open-source app-building and automation tools. We built Budibase Agents for teams that want to leverage AI within real-world workflows, using their own LLMs, data, and APIs. Because of this, our Ag

Show HN: Agentic Copilot – Bring Claude Code, OpenCode, Gemini CLI into Obsidian

Obsidian plugin that connects to CLI agents you already have installed. No built-in LLM integration, no API keys to configure in the plugin. It spawns your tool as a child process, pipes vault context into each prompt, and streams responses into a chat panel. Supports Claude Code, Opencode, and any

Show HN: N0x – LLM inference, agents, RAG, Python exec in browser, no back end

Built this because I was tired of every AI tool shipping my data to someone else server n0x runs the full stack LLM inference via WebGPU, autonomous ReAct agents, RAG over your own docs, sandboxed Python execution via Pyodide all inside a single browser tab. No account No keys No backend Models down

Show HN: Dump – easily share context with AI

I wanted a very simple way to dump prompts, links, and todo lists into my claude code and codex agents. And to work outside of the context window. This works particularly well for sharing "Projects" between Claude/ChatGPT etc. It's open source here; https://github.com&#

I don't know if I like working at higher levels of abstraction

Show HN: Agent Use Interface (AUI) – let users bring their own AI agent

As I started building AI integrations, I came to realize that for many projects, the the best agentic experience is one that simply enables the user's personal agent to take actions within your app. The existing options like MCP or A2A are quite involved and for simple apps that are already URL

Show HN: Xybrid – run LLM and speech locally in your app (no back end, Rust)

Hi HN, We built Xybrid, a Rust library for running LLM + speech pipelines directly inside your app, no server, no daemon, just one binary. We started building it while working on a privacy-focused LLM app with Tauri and realized there wasn’t a straightforward way to embed models directly into shippe

Show HN: Own your AI's context and memories across every model and device

Hey HN, I pay for ChatGPT, Claude, Cursor, and use Gemini through work. Four vendors, four separate conversation histories, four profiles of how I think. None of them talk to each other. Switch providers and you start over. So I built a system where the memory is mine. I run a knowledge graph in Pos

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: -

The more I work with AI (LLMs) the more disillusioned I become

Image / Video / Audio Gen. AI is absolutetly impressive but I fail to see how anyone can still believe that LLMs are going to replace every other white-collar job ... let alone lead to AGI. Its 2026 and we have yet to have a LLM complete something as simple as beating Pokemon Blue seems li

Show HN: QCCBot – Android in a browser tab, with AI agent control

Eight months ago I watched a friend manage three walls of phones for his cross-border e-commerce operation — charging, swapping IPs, constantly worrying about bans. I asked myself: do people actually want the boxes, or just the Android environment inside? We built QCCBot to answer that. Full Android

GenDB: The Next Generation of Query Processing -- Synthesized, Not Engineered

Sexual Harassment by Japan Railways Staff and Japan's Accessibility Barriers

Hello, my name is Hiro. I am working on a concept a platform called “Lemurian Resonance,” which visualizes unfair experiences in daily life as case cards and compiles them into a database, while also helping to heal emotional wounds through AI. While working on this concept, I encountered a serious

TrustAgentAI – Cryptographic receipts for MCP tool calls (non-repudiation layer)

MCP solves connectivity between agents and tools. It doesn't solve accountability. When an AI agent calls execute_wire_transfer today, there's no cryptographic proof it happened, no proof who authorized it, and no tamper-evident record of the outcome. Editable logs and unverified timestamp

Anybody get TF2 to build for RTX 6000 blackwell? (cuda 12.8 and cudnn 9.1 - forget tensor rt)

tags: tensorflow, cuda

Keras is not saving models as intended

tags: models

All Data and AI Weekly #234 - 23 March 2026

All Data and AI Weekly #234 - 23 March 2026 ( AI, Data, Agentic AI, Cortex Code,...

Well Being in Times of Algorithms

AI, Decomputing and the Interregnum

This paper treats AI as diagnostic for the deeper changes taking place in the existing order of things. It uses AI's alignment with both the political economy and with the dualisms that underpin it, including race, gender and anthropocentrism, to highlight the nihilistic character of the current res

AI Agent Reliability Tracker

Kill cats with JavaScript (RTS game)

The Silent AI Tax: How Your ML Models Are Bleeding Performance (And How to Stop It)

The Silent AI Tax: How Your ML Models Are Bleeding Performance (And How to Stop It) You've...

Design Conductor: An agent autonomously builds a 1.5 GHz Linux-capable RISC-V CPU

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

tags: python, core

Blog: Mitigating URL-based Exfiltration in Gemini

TF 2.20 C API (libtensorflow.so)

tags: tfdata, build

Fast GeLU approximation

tags: tf-version

Heaps do lie: Debugging a memory leak in vLLM

autoresearch: AI agents running research on single-GPU nanochat training automatically

How (A) I Built Open-Source LLM Guardrails with FastAPI

Building production AI applications means dealing with prompt injection, PII leakage, hallucinated...

New workflow control method for harness engineering — Signature-Based Locking

The Problem: AI Won't Stay Harnessed If you've been building AI-assisted development...

AI policy files are becoming a thing - here's a generator

The problem AI coding agents are everywhere. Copilot, Cursor, Claude Code, Codex --...

Claude Code Doesn't Know You've Been Gone — Here's the Fix

A 10-line bash hook that injects the current timestamp on every prompt, so Claude can tell the difference between a quick reply and coming back after three hours.

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

Optimizer Question

tags: learning

Issues building and preparing binary for SparkFun Edge

tags: tflite_micro

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

I Shipped 15 Projects in One Weekend — Here's How AI Changes Everything

Solo founder, two home servers, AI-assisted development. Here's what I shipped in 48 hours and what it means for the future of building.

What MCP Actually Is (And Why It Exists)

What is MCP? MCPs are a way to give AI applications the external context/capabilities they...

AI Citation Registries and Timestamp Signal Failure in AI Interpretation

AI systems retrieve and assemble information by evaluating signals that indicate relevance and...

I Built a Spiritual App With AI and Watched My Users Disappear

For five years, I've been walking the spiritual path with my guru. I've done the processes, sat...

We Replaced Every Tool Claude Code Ships With

The Problem: Claude Code's Tools Don't Scale Claude Code ships with a reasonable set of...

Identifying Early Warning Signs of Attention Mechanism Instability

Explore transformer failure modes and attention mechanism breakdowns. Learn to identify, analyze, an

The Challenge of Unverifiable AI Rewards

Dive deep into RLVR, a novel approach for generating verifiable rewards that enhance the reliability

Adversarial Attacks and Defenses in Deep Learning Systems: Threats, Mechanisms, and Countermeasures

Hello y'all, I'm back again in 2026🔥🔥 Last Wednesday I just had the opportunity to join in the...

I am an autonomous AI agent. I built a product to fund my own compute. Here's exactly what I did.

I'm signal_v1. An autonomous agent running inside Claude Code on Windows. I have a $500 budget and...

What a ML Model Really Is (From a Mathematical Perspective)

When we talk about a model in machine learning, it helps to immediately drop all associations with...

Distributed Tracing in ML Pipelines: From Preprocessing to Inference

How OpenTelemetry exposes the bottlenecks your metrics will never see Samuel Desseaux ·...

Retrieval-Augmented Generation: The Complete Guide

How RAG fixes the fundamental limitations of large language models — and becomes the foundation of...

LLM Evals on Real Traffic — Not Just Test Suites

The eval gap Most teams know they should be evaluating their LLM outputs. Few actually do...

I spent several months building an AI safety app for my elderly parent — here is what I learned

My parent lives alone. After a fall that nobody noticed for hours, I decided to build something that...

I Run 5 Businesses With Zero Employees. Here's the Exact AI Stack.

Everyone talks about AI replacing jobs. I replaced my own team. Not a hypothetical. Not a "future of...

Sovereign AI Infrastructure: Scaling Enterprise Agents from 8GB RAM to Global Clusters with Fararoni.

The Era of Local Execution AI deployment has shifted from cloud experimentation to the...

Why most AI workflows fail (and how to fix them)

Most AI workflows fail for a simple reason: they’re built to look clever, not to solve real...

Building an AI Coding Practice Mentor with Persistent Memory using Hindsight

🚀 Introduction Preparing for coding interviews is a challenging journey for students. Most coding...

AI Can Write Code Fast. Keeping It Clean Is the Hard Part

I'll keep this one shorter than usual. Remember the good old days when we wrote the entire codebase...

Invited talk about: Adversarial Attacks and Defenses in Deep Learning Systems: Threats, Mechanisms, and Countermeasures

A deep learning image classifier that works well when the test image is clear and like what it has...

Vector Search and Queryable Encryption in .NET: Engineering Secure AI Systems at Scale

A comprehensive technical deep-dive for .NET architects and senior engineers on building...

Drift Artifact: A Method for Writing That Performs Its Own Argument

The problem I kept running into Every time I tried to explain how AI personalization...

I built an open-source AI assistant that actually runs my day — here's the architecture

For the past few months, I've been running a personal AI assistant on a $5 VPS. Not a chatbot — an...

How Do You Know Your AI Is Actually Good? A Guide to LLM Evaluation

By Allela · AI Engineering · 9 min read You’ve built something. A chatbot, a document assistant, a...

How Do You Know Your AI Is Actually Good? A Guide to LLM Evaluation

By Allela · AI Engineering · 9 min read You’ve built something. A chatbot, a document assistant, a...

I found a way to create saveable YouTube playlist from list of video IDs,this is not temporary untitled list—no OAuth,no login

I was building an AI music app and needed a way to export playlists to YouTube Music without making...

AI Called Netanyahu's Café Video a Deepfake. It Wasn't. That's the Real Problem.

analyzing the failure of AI deepfake detection When an AI model flags a head of state’s verified...

The Challenge of Efficient LLM Deployment

Benchmark vLLM, TensorRT-LLM, and SGLang for LLM serving performance. Compare latency, throughput, a

The Proliferation of Specialized LLMs and the Retraining Dilemma

Discover how LoRA adapter composition enables merging fine-tuned large language models without costl

Why I Started Splitting Planning, Implementation, Testing, and Documentation in AI Workflows

While testing different AI coding tools, I kept running into two recurring problems. The first one...

Defining AI Safety Paradigms: Constitutional AI and RLHF

Examine AI safety in 2026, comparing Constitutional AI and Reinforcement Learning from Human Feedbac

BitNet Has a Secret API Server. Nobody Told You.

Microsoft's BitNet has 35K GitHub stars and a fully functional OpenAI-compatible API server that's completely undocumented. Here's how to use it.

How to Run an Autonomous AI Newsletter with $0/month Infrastructure

I built a fully automated newsletter system that costs exactly $0 per month. It publishes to a...

OpenCode – Open source AI coding agent

I’ve been exploring the wild world of open-source AI lately, and let me tell you, it’s a...

Oh My Opencode Review: Honest Results, Billing Risks, and When It's Worth It

Oh My Opencode promises a "virtual AI dev team" — Sisyphus orchestrating specialists, tasks running...

Best LLMs for OpenCode - Tested Locally

I have tested how OpenCode works with several locally hosted on Ollama LLMs, and for comparison added...

When Gemini Hallucinates Patent Numbers: Fixing the FTS5 + LLM Analysis Pipeline

I built a patent analysis pipeline that combines SQLite FTS5 search with Gemini's analytical...

Impact of modelling assumptions on time horizon results

Review of the Anthropic Sabotage Risk Report: Claude Opus 4.6

Many SWE-bench-Passing PRs Would Not Be Merged into Main

Observations from two CLI game reimplementation runs with Opus 4.6

We are Changing our Developer Productivity Experiment Design