Buy A Modem

Show HN: Pit Claude, Codex, and Gemini against each other, and apply the best

Voratiq is an open-source (MIT-licensed) terminal-native CLI that lets you run multiple coding agents against the same spec, review their diffs side-by-side, and use whichever implementation is best.It's designed for experienced developers who want to make the most of agentic coding. It positions you as the architect and reviewer, and shifts implementation onto an ensemble of agents who (hopefully) in aggregate can do a good job. And in my experience, the latest generation of models do quit

Thoughts on Cursor

I believe Cursor just rolled out its two major features: Debug and Design.I had an understanding of what I wanted from the IDEs, but I could not fully articulate it before the launch. Now that it’s here, it makes complete sense.The way I see the future of programming, everything is going to be live: debugging, coding, designing, etc. Not that the idea is new, but the difference is that now it will be fully autonomous.Recently, I worked on a feature that required redesigning part of our legacy fl

Show HN: I built a circuit breaker that predicts AI failures

Hi HN — I built this and I’m explicitly asking skeptics to tear it apart.” Interlock is a safety and certification layer for AI infrastructure, not an optimizer or a vector database.The problem I am solving for is that AI systems (vector search, RAG pipelines, agent frameworks) don’t usually fail cleanly — they degrade silently, oscillate under load, or keep returning corrupted results until something crashes. Monitoring tells you after the fact; circuit breakers tend to be static and blind to c

Show HN: Axiom for Claude Code – Coding skills for iOS devs

This is a new, open-source (MIT) Claude Code plug-in for Apple platform developers. I initially built it to help me build my first "real" iOS app, and it made the process so fun and productive that I wanted to share it.I recently christened it v1.0, but I continue to evolve it quickly. It's biased toward modern frameworks and best practices, but I continue to notably broaden coverage every week. Here are the technologies and concepts Axiom currently has deep coverage for:• Swift &

Show HN: Quorum – Multi-agent CLI debates (AutoGen back end and React/Ink TUI)

I built a CLI tool to orchestrate structured debates between LLMs. The goal is to reduce hallucinations and bias by forcing models to critique each other before reaching a consensus.It supports a hybrid setup: you can have your local Ollama models debate against GPT-5.2 or Claude 4.5.Key Features: - Local-first: Auto-discovers Ollama models (Llama 3.3, Mistral, etc) with zero config. - Method Advisor: Press Tab to analyze your prompt. A meta-agent will recommend the best debate method (e.g. &quo

Show HN: Kinkora – A creative playground for experimenting with video models

Hi Indie Hackers We’re building Kinkora, a creative platform that brings multiple image and video AI models into one place for experimentation and creation.Like many builders, we found ourselves constantly switching tools just to test different models or creative directions. Each platform felt limited to a single workflow or use case. So we decided to build a space that’s more modular, exploratory, and creator-first.Kinkora focuses on:Supporting popular generative modelsMaking experimentation f

Ask HN: Best back end to run models on Google TPU?

So, I got Pixel 10 Pro, and I'd like to run parakeet (or whisper) model on it for voice to text. I'm building an ai dictation app (aidictation.com). I'm struggling to find a way to run this model on device. I have to reserve to use groq API, which is suboptimal.<p>Any recommendations?

Show HN: Bets on Post-GPU Compute

Post-GPU compute is a bet that new hardware will make different primitives cheap.<p>Extropic AI bets the primitive is sampling.<p>How far does block Gibbs sampling algorithm generalize beyond strictly bipartite models?

Dance Dancing GIF

Love You Kiss GIF by The Fuzzlets

Dial Up Mr Mackey GIF by South Park

feels social media GIF by PEEKASSO

Show HN: Isogen – Lightweight AI Coding Tool (Rust and JavaScript, <50MB, BYOK)

I built an AI Coding tool optimized for my workflow. VSCode forks use too much memory and I am over the idea of having AI Agents rewriting files and also reviewing complicated diffs with the Accept/Reject UI.I built Isogen which uses as much memory as a Chrome tab instead of a tool that can spike up to 1GB of RAM. You drag and drop or paste files into an isolated context and do fast generations file-by-file. This approach allows me to keep a strong mental model of the codebase. I also added

Show HN: Thinkmoon.ai – Build your own Alpha Arena, trade crypto with AI Agents

With ThinkMoon you can build your own Alpha Arena: plug in models from OpenRouter, OpenAI, Anthropic (or our custom ones), feed them real-time data (candles, order book, ticker) — then send them to trade real crypto (via Binance Futures).You define the rules: risk limits, which coins (BTC, ETH, SOL, XRP… whatever), even mix different LLMs and compare their performance.Every move the AI makes is logged — full prompt, reasoning, market snapshot — so you can audit or learn from it.There’s a live da

Divyam-LLM-interop:LLM responses,requests translation across APIs and models

We at divyam.ai built a library that translates LLM requests and responses across different model families and APIs, including OpenAI’s Chat Completions and the newer Responses API.It handles model-specific idiosyncrasies across popular families like GPT, Gemini, Llama, Qwen, and others. This includes dropping unsupported fields, renaming deprecated ones, normalizing structures, and generally cleaning inputs so they conform to each provider’s/model's stricter expectations.The library a

Launch HN: Mentat (YC F24) – Controlling LLMs with Runtime Intervention

Hi HN, I’m Cyril from CTGT. Today we’re launching Mentat (https://docs.ctgt.ai/api-reference/endpoint/chat-completions), an API that gives developers deterministic control over LLM behavior, steering reasoning and removing bias on the fly, without the compute of fine-tuning or the brittleness of prompt engineering. We use feature-level intervention and graph-based verification to fix hallucinations and enforce policies.This resonates in highly regulated industries or oth

Show HN: Detail, a Bug Finder

Hi HN, tl;dr we built a bug finder that's working really well, especially for app backends. Try it out and send us your thoughts!Long story below.--------------------------We originally set out to work on technical debt. We had all seen codebases with a lot of debt, so we had personal grudges about the problem, and AI seemed to be making it a lot worse.Tech debt also seemed like a great problem for AI because: 1) a small portion of the work is thinky and strategic, and then the bulk of the

Is there an hour by hour model quality tracker?

I'd love to put some actual science behind a feeling I think a lot of us share: certain hosted models seem to dip in quality at certain times of day. In my group, it’s almost treated as a “yeah, of course” kind of thing. Having something near-real-time that shows which models are performing best at any moment would be… honestly, pretty awesome.I asked Perplexity and got nothing: https://www.perplexity.ai/search/is-there-an-hour-by-hour-model-8.03Z_BlSomAY7o9F5HwmQ#0

Show HN: QonQrete – Local-first multi-agent system for sandboxed code generation

I’ve been working on an open-source project called QonQrete and would like feedback from HN.What it isQonQrete is a local-first, agent-based orchestration system for code generation. It coordinates multiple LLM “agents” to plan, write, and review code, while keeping execution inside a sandbox on your own infrastructure. Think of it as a construction yard for AI-assisted development that you run yourself.Why I built itMost multi-agent demos I saw had two issues:– Security: generated code often ru