The best Hacker News stories from Show from the past week

from

All Show

from the past*

day week month

with these search terms:

Create follow Go back

Latest posts:

Show HN: Huzzah – a novel approach to coding with AI

Hello everyone. I've been working on this experimental editor called Huzzah.<p>I've been working almost exclusively with coding agents since January of this year, and over the past few months I began to feel utterly exhausted by them. They're great, but I'm finding it more and more tedious to write full sentences for every change I want. Not only that, but it seems there's a complexity limit for codebases - beyond a certain point the agent begins confusing itself.<p>I'd like to go back to writing code, but I don't want to go all the way back to fully manual coding. So I've come up with this interaction paradigm where you:<p><pre><code> 1. write pseudocode in whatever way makes the most sense to you 2. on save, the editor synchronizes your work to real source code 3. the pseudocode is persisted alongside the generated code, making your prompt effectively a stored record of intent. </code></pre> It may not work for every use case, but in my initial playthroughs I've found it very enjoyable.<p>Right now it's just a proof of concept - installation instructions are here in the readme: <a href="https://github.com/danielvaughn/hz" rel="nofollow">https://github.com/danielvaughn/hz</a><p>You can also watch a video of it in action here: <a href="https://x.com/danielvaughn/status/2090456808431165715" rel="nofollow">https://x.com/danielvaughn/status/2090456808431165715</a><p>Cheers!

Show HN: I trained a 125M model to autocomplete piano on-device

I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15).<p>The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device.<p>The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.

Show HN: Interactive, animated architecture of any HuggingFace models

Show HN: Automatically detect and patch walking-dead states in Sierra games

Hi HN, I've become lazier in my old age and struggle to replay my favorite Sierra games from the 80s and 90s because I keep getting into those situations where I need an item from 3 acts ago, I have no save game handy, and now I gotta make dinner.<p>So I'm building the Lucasartsifier: a static analysis tool that decompiles Sierra resource files, automatically finds those states, automatically generates code to prevent the player from getting into those states, then emits loose patch files that can be placed alongside the original game resources. There's no game-specific code involved; all the logic is generic, though of course Sierra introduces new idioms and mechanics in every game so every new supported game needs a bunch of engine work.<p>So for example in Leisure Suit Larry 2, the patched game prevents you from boarding the cruise ship until you have both the sunscreen and the Grotesque Gulp. Without them you die on the raft 3 play-hours later.<p>So far this works on Leisure Suit Larry 2 (SCI0), King's Quest 4 (SCI0), King's Quest 6 (SCI1.1), and Laura Bow 2 (SCI1.1). I'm currently working on King's Quest 5 (SCI1.0).<p>This is work done with Claude -- I do the design and playtesting and it does the rest :D<p>Any feedback, play testing, and suggestions would be great!

Show HN: Saggar, a Mac terminal that keeps sessions and your attention organized

Hi HN, like many of you talented folk, I’ve been building more, faster than ever. I built Saggar after my terminal stopped feeling like a tool and started feeling like a tab-management problem. Let me know you think!

Show HN: 1667, a terminal UI for writing fiction with language models

Hi HN. I built 1667 for my own fiction work and now use it each day. This probably has a limited audience. Maybe an audience of one...<p>Why a terminal interface for story writing? I'm a dev. I like to use terminals for a lot of stuff. Most WebUIs feel off to me. That's the only reason.<p>One thing that bothers me about writing in existing tools is that they don't fit the way I write. The mental model of my story is a tree. I try many takes usually continue with just one, but sometimes I want to try an alternate route and see where this goes. And that can branch again in many places. See what happens if I kill off this character or they don't take the job or whatever.<p>1667 is a full-screen terminal app for long-form fiction. Each story part can have several takes. All takes stay in a tree. You select one path through that tree as the story line. Export writes that line to Markdown in the project folder.<p>Some technical details: - A project stores its stories and settings in a `.1667/` directory. Exported Markdown sits beside it. - Provider secrets stay in private machine files. Requests go to the provider that the writer selects. - An optional Vault Password seals project files at rest. - An operating-system lock permits one writer process for each project. - The request viewer shows the next provider request without its credential. - Each generated take keeps a Generation Record with its model and effective settings.<p>Version 0.9.5 runs on macOS, Linux, and Windows x64. The website has Shell and PowerShell installers. An npm package is also available.<p>1667 imports Markdown, SillyTavern chats and cards, and NovelAI archives. It can use OpenAI-compatible, Anthropic, and local endpoints such as Ollama, LM Studio, llama.cpp, and KoboldCpp.<p>Current limits: the release is pre-1.0. The interface is a terminal. There is no account or cloud sync, and I don't plan to add any. No tracking.

Show HN: Desktopcolors.com – A museum for solid background colors of classic OS

Hi HN, in my last vacation, I built desktopcolors.com as a place to collect iconic solid background colors. Let me know what you think.

Show HN: Mic Drop, a real-time multiplayer karaoke game

Show HN: A public AI whose memory is shared across all users

Show HN: ThoughtDAG – An editable context graph for LLM conversations

Show HN: Eigendrum - Draw any shape and hear what it sounds like as a drum

Hi HN, I built Eigendrum, a web tool that solves the 2D wave equation for arbitrary shapes so you can hear what they sound like as drums.<p>How it works: * Solves -∇²u = λu using finite element analysis (Kφ = λMφ) on a triangle mesh. * Validated to <0.1% error against closed-form solutions for circles (Bessel zeros) and rectangles. * Sound model factors in strike location, Rayleigh damping, and mallet width. * Includes Kac drums I & II to demonstrate identical sound spectra from different geometries. * No frameworks, build steps, or dependencies.<p>Repo and tests: <a href="https://github.com/BaselAshraf81/eigendrum" rel="nofollow">https://github.com/BaselAshraf81/eigendrum</a><p>Happy to answer any questions about the FEM solver or Web Audio setup!<p><a href="https://eigendrum.com/#p=circle" rel="nofollow">https://eigendrum.com/#p=circle</a>

Show HN: C# Game Engine with its own scripting language and IDE

Show HN: Woxi - Open-source Mathematica / Wolfram Language reimplementation

Woxi is an interpreter for the Wolfram Language written in Rust.<p>It comes with Woxi Studio, a Mathematica-like GUI built with iced, but you can also use Woxi through a CLI, Jupyter kernel, Python package, npm package, or WASM module.<p>Compared with wolframscript / Mathematica, the main differences are:<p>- Free and open source - Very fast startup - Typically milliseconds rather than seconds for the Wolfram kernel, making Woxi practical for shell scripts, one-liners, and other short-lived processes - Embeddable - It can run in a browser via WASM or be embedded into another application as a scripting language<p>A more detailed comparison with Mathematica is available here: <a href="https://woxi.ad-si.com/docs/comparison/mathematica/" rel="nofollow">https://woxi.ad-si.com/docs/comparison/mathematica/</a>.<p>Conformance is ensured with ~26'000 unit tests and ~900 .wls script snapshot tests.<p>The current focus is on fixing remaining edge cases, improving performance, and growing the community.<p>If you use the Wolfram Language, I'd be particularly interested in feedback on compatibility and missing functionality. Contributions and bug reports are also very welcome: <a href="https://github.com/ad-si/Woxi" rel="nofollow">https://github.com/ad-si/Woxi</a>

Show HN: Write.md – A free, open-source, themeable Markdown editor for macOS

Show HN: Git-knife – Edit commit messages, authors, and dates like a spreadsheet

Show HN: Scroll through all 43252003274489856000 Rubik's Cube states

Show HN: iPhone app takes simultaneous images from 2 lenses, fuses into 1 photo

Show HN: Ante, a coding agent in a single binary that runs offline

Show HN: Needle2: 14MB agentic LLM for phones, wearables, smart home and robots

Hey HN,<p>Henry from Cactus here!<p>We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2.<p>The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges 300-700 on sub-$200 phones such as the Samsung A-Series.<p>On the tool call and mobile device use benchmarks, Needle 2 trades wins with closest small models like LFM2.5 230M and Apple Foundation Model, at 5x to 70x smaller, both at f16 vs Needle 2 at 2bit. Needle is based on Simple Attention Networks from our paper (<a href="https://arxiv.org/abs/2607.18363" rel="nofollow">https://arxiv.org/abs/2607.18363</a>).<p>Edge AI has lately meant Macs and PCs, but that is just 1.5 billion of over 21 billion connected IoT devices in the world today, and in emerging markets most phones ship under $200, no NPU, cheap GPUs. These include budget phones, Raspberry Pis, microcontrollers, wearables, small robots like Reachy Mini, and connected home devices.<p>A conventional transformer of Needle's width and depth spends 164 MFLOPs per token, and even one squeezed down to Needle's parameter count spends 87, Needle spends 70. Even on a high-end phone, an always-on assistant lives inside a power budget; every MFLOP is milliwatt-hours, and Needle spends 7x to 85x fewer of them per token than the smallest performant LLMs. More about the architecture in the link.<p>When we structure intelligence for consumer devices as functions with typed parameters, the only hard part is mapping a messy sentence onto them; which function, with which values. Our research found that when framed that way, the problem needs no world knowledge and no open-ended prose, which is why 45M parameters suffice.<p>Needle 2 expands to structured extraction where the schema can be passed in-place of tools and the model returns structured output. You can use Needle as a text-classification model with an enum field, as a summarization model by providing a schema that extracts key fields, everything but free-range decode.<p>Every product has its own tool vocabulary and fine-tuning needle helps it achieve frontier-level performance on custom tasks, so using the python package (<a href="https://github.com/cactus-compute/needle" rel="nofollow">https://github.com/cactus-compute/needle</a>), Needle can be fine-tuned Needle on a Mac/PC in minutes to a few hours, with automated data-generation pipeline, just pass a couple samples.<p>Nonetheless, every response carries a learned confidence score based our Cactus Hybrid technique. If above your threshold, act, below it, escalate to the cloud or bigger model. Combining Needle 2 with a private DeepSeek-v4-Flash deployment works particularly well for enterprise-level tasks at barely any cost, we can help with this setup.<p>We have put a lot of thoughts into Needle 2 but might still be missing quite a lot, please use the playground in the provided link to test Needle and share your thoughts, always appreciated!

Show HN: Voice driven murder mystery, Interview AI suspects with your voice

Hey HN!<p>I'm excited to show off this really fun project I put together. I originally built this project 2-3 years ago, AI was already booming at the time, however voice AI agents were still very early. I loved my proof of concept at the time, but wasn't quite happy with it.<p>I recently had the desire to check out the tech again, and know many of you will be interested.<p>Interviews are speech to speech with OpenAI's gpt-realtime-2.1 over WebRTC. This model is... expensive, and because of that, I have to add some amount of restrictions, conversations are tied to a authenticated Clerk user id. I have also added a 30 minute timer because well, I really don't want to go broke while I sleep tonight.<p>Each suspect has a tool they call when you make a direct accusation. It captures who you accused and a faithful list of the evidence you actually stated.<p>A separate gpt-5-mini judge then decides which of the case's required evidence facts you genuinely presented. Paraphrasing counts, vague suspicion and fishing don't.<p>The rest is Next.js, MongoDB, and Clerk.<p>Let me know whether the suspects hold up under a real interrogation.