Article URL: https://blumify.io/people
Comments URL: https://news.ycombinator.com/item?id=50081151
Points: 70
# Comments: 47
Hi HN, I'm Niz. Pulse Local is a budgeting app for Mac (Apple Silicon, beta) that answers questions about your spending with models that run on your machine. Local mode needs no account. The ledger and your questions stay on the machine, and the app sends no analytics or crash logs.
I felt there are not enough consumer apps that run locally with no calls to the outside world that help us manage sensitive corners of our lives. This app in its local variant only downloads a couple of models and occasional updates, it does not send any analytics nor data.
Features-wise, Pulse local is a complete budgeting app that offers an assistant built on top of a fine-tuned model and thousands of rules embedded in the code itself. I'm sharing it for free with options to help me continue experimenting with fine-tuning and expanding along this direction.
How it works: you import a CSV from any bank, in any currency. When you ask a question, a tuned Qwen3 0.6B model reads it into a structured query, and code, not the model, does the money math against the ledger. A 4B checker model double-checks each money answer; when the two readings disagree, the app says it is not sure instead of guessing. That change took wrong answers from about one in five to 5 in 494 questions across our last three tests. The write-up covers how we test, the designs that failed, and the limits of the claim: https://pulseos.finance/pulse-local/writing/agree-or-not-sur...
The tuned weights are on Hugging Face (PulseOS/pulse-local-qwen3-0.6b).
Limitations: the first start downloads about 2.9 GB of models. With the checker on, a money answer takes 12 to 14 seconds (you can turn the checker off). Small models can still misread an open-ended question. The app answers in English. Mac only for now; Windows is next.
The app is free. An optional Connected mode (Sync Pass, $29/year) syncs US banks through Plaid; transactions pass through our server's memory and are not stored.
Feedback welcome, especially on wrong answers and on questions it should handle but does not.
Comments URL: https://news.ycombinator.com/item?id=50080275
Points: 99
# Comments: 54
Guitoro is a free browser guitar using recordings of real instruments. You can strum with your keyboard, mouse, or touchscreen, choose acoustic, classical, electric guitar or bass, and see chord fingering and each stroke. Song practice can be slowed to 20%, and there are free chord tools and tuners. No signup or download.
1 minute video demo: https://youtu.be/DtcluMLdSgs
I would love feedback on how the strumming feels and sounds.
Comments URL: https://news.ycombinator.com/item?id=50077020
Points: 515
# Comments: 141
Article URL: https://lab.specr.net/rigged-draw/
Comments URL: https://news.ycombinator.com/item?id=50074909
Points: 186
# Comments: 95
Article URL: https://pdftomp3.com/connect-ai/
Comments URL: https://news.ycombinator.com/item?id=50069337
Points: 131
# Comments: 66
Article URL: https://github.com/bherbruck/solidcraft
Comments URL: https://news.ycombinator.com/item?id=50064840
Points: 104
# Comments: 22
Article URL: https://github.com/snuri00/ps2-web-recomp
Comments URL: https://news.ycombinator.com/item?id=50052251
Points: 211
# Comments: 114
Hi HN! I'm Saarth. I'm in 9th grade, and I built Atticus.
Atticus is an incredibly fast and cheap Mac assistant that can do tasks for you. You hold down Control-Option-Space, speak or type your task, be it complex or a simple browser search, and Atticus will get on it.
The interesting part is that simple tasks, like opening apps, playing a Spotify playlist, or web queries, don't touch AI, and instead are processed locally and typically finish in under a second. For example, prompting Atticus to open apps takes around 0.08 seconds, sometimes even finishing before you finish your sentence, sort of like Jev.
Any tasks that are more complicated will be sent to Claude via claude -p, meaning you never need an API key and Atticus will run on your existing Claude Pro or Max plan. By using this approach, costs are also cut down immensely, with, on average, tasks costing a fraction of a cent.
To achieve Atticus' speed, it remembers shortcuts from each task it completes. When it does a task right for the first time, it saves its steps and replays them the next time without AI. For example, a workflow of creating a new meeting took about 30 seconds the first time around. Atticus then logs the exact steps to memory and takes 5-8 seconds on the next pass. To add to this, Atticus also learns about you as you use it more and from what's on your computer, including your name, your apps, and things you want it to remember, personalizing its answers to you.
To combat the safety concerns over giving an AI unrestricted access to your device, Atticus has toggleable restrictions, so before anything is deleted, sent, bought, or written to a file, the user is prompted before that action.
Some features I'm currently working on include a meeting copilot that capitalizes on Atticus' speed, allowing Atticus to listen in on meetings and answer questions in real time as they come in, personalized to the context of the meeting and what it knows about you.
To run Atticus, macOS 26 or higher is required for its voice transcription features. The voice transcription is also done locally by macOS.
Atticus is free, open source, and also very fresh, and I'm completely aware of its many rough edges, which is why I'd love to hear from everyone about what breaks and what you guys would like to see Atticus do!
Comments URL: https://news.ycombinator.com/item?id=50049588
Points: 11
# Comments: 0
Article URL: https://betterwispr.com/
Comments URL: https://news.ycombinator.com/item?id=50042333
Points: 218
# Comments: 126
Hi HN, I'm Josh. WallHop is a free paywall bypass: paste an article URL (or prefix any URL with wallhop.io/) and you get the readable article without the paywall popup, ads or clutter. No signup, no extension.
I built it after 12ft.io went away. The News Media Alliance took it down in July 2025 and I missed having something I could hit from my phone.
How it works: it started as a fork of everywall/ladder (GPL-3.0) and I've modified it heavily since. Instead of relying on one trick, it chains several bypass methods per request and falls back to archive.is when they all fail, so one site blocking one technique doesn't break it.
Things that might be interesting:
- iOS Shortcut: on iPhone, tap Share > WallHop from any article in Safari or another app. This is how I use it 95% of the time. - It works on Medium member-only posts, which 12ft never handled. - Bookmarklet for desktop, plus /raw/ and /api/ endpoints if you want to script it.
I've tested it against most major news outlets and a bunch of local papers. It won't work everywhere. There are SOME server-side paywalls that are out of reach, and I'd rather say that up front. If you find a site that breaks, I'd love the link, and I can likely figure out a way to bypass it!
Comments URL: https://news.ycombinator.com/item?id=50038634
Points: 267
# Comments: 117
Showing off carrier-explode, one of my side projects that continuously archives carrier settings for all major phone brands.
carrier-explode also contains decoders and explanations for all the common baseband configurations, etc.
There’s still quite a bit of work to be done in checking my assumptions, but the tool has already been proven useful for a few enthusiast groups.
Comments URL: https://news.ycombinator.com/item?id=50024499
Points: 431
# Comments: 55
Article URL: https://readrare.com/
Comments URL: https://news.ycombinator.com/item?id=50024055
Points: 190
# Comments: 66
Article URL: https://github.com/franzenzenhofer/big-arrow-on-the-screen
Comments URL: https://news.ycombinator.com/item?id=50018817
Points: 400
# Comments: 175
Article URL: https://quake-srp.pages.dev/
Comments URL: https://news.ycombinator.com/item?id=50016312
Points: 225
# Comments: 175
Article URL: http://scottbezek.blogspot.com/2026/10/making-flexible-neon-t-shirt-with-leds.html
Comments URL: https://news.ycombinator.com/item?id=50008047
Points: 222
# Comments: 37
Hello everyone, I wanted to share this small project that I have. It's born out of a necessity that I had, because Proton doesn't ship a Linux version of the drive yet (it's apparently coming out later this year but who knows?) and all the current solutions are hacky and wonky at best. But for the time being I really needed to mount my Proton drive on my Linux machine the way I mounted it on my Mac, so I (and my faithful AI slave) did this small utility that allows you to mount your Proton Drive in a FUSE FS like if it is a local folder.
It's fully written in Go because Proton's drive SDK and API libs are all in Go, and there's a community project on top of the reverse engineered proton API there too. I am not a Go developer, I know the foundations, so of course I used AI quite a lot, I am not super proud of it, but I am using it daily and testing it myself for a few months before releasing here.
Hope you like it, feedbacks are welcome, just please be mindful of the person on the other side :)
Comments URL: https://news.ycombinator.com/item?id=50003545
Points: 122
# Comments: 46
Article URL: https://bigwords.page/
Comments URL: https://news.ycombinator.com/item?id=49994443
Points: 728
# Comments: 165
From Li Cheng to Banksy
Comments URL: https://news.ycombinator.com/item?id=49992057
Points: 174
# Comments: 68
Article URL: https://github.com/allenv0/SCM
Comments URL: https://news.ycombinator.com/item?id=49952111
Points: 183
# Comments: 74
Article URL: https://tej.as/blog/aleph-alpha-kolibri
Comments URL: https://news.ycombinator.com/item?id=49943034
Points: 171
# Comments: 92
Hi there :-) New on HN, first time posting.
Past year, around December, I started experimenting with making ChatGPT and Claude generate source code in LDraw language.
This LDraw is literally an "assembly" language, a low-level programming language that describes how to assemble LEGO pieces together into models, one placement instruction at a time.
When executed by specific tools, like e.g. LDView, LeoCAD, Studio... these instructions become LEGO CAD models, that can be interacted with, modified, etc.
Or, in other words: one LDraw source file in .mpd or .ldr format is equivalent to one LEGO CAD model.
So, the idea I had was: if I manage for maybe ChatGPT or Claude to generate high-quality LDraw source files... then, they would actually be generating high-quality LEGO CAD models, right?
Then, after months of iterations and trying one thing after the other... it worked!!!
Long story short: using GPT-6 Astra and Opus 5.5, I've managed to create a python toolset, instructions, and docs for agents in general. Now, these can be used by them to generate LDraw models.
I've packed it all as a dockerized web app for others to try and experiment, with several providers (and agents) to choose from: OpenAI, Claude and OpenRouter.
Here's a bunch of exmaples: https://anteloc.github.io/index-samples.html.
If you are curious about the internals of an .mpd model, the "what was the agent picturing on its mind", open the .mpd file that got your attention on a text editor, and read the first line under the ones starting with "0 FILE".
I'd really appreciate feedback and comments, let's see where this goes =)
Comments URL: https://news.ycombinator.com/item?id=49937916
Points: 170
# Comments: 52
pi pod runs sessions of the pi coding agent in isolated sandboxes ("pods") on a server you run, in composable environments.
----
Since moving my company towards AI-native work, I have been really frustrated by the state of "agentic engineering" environments. Products by the labs (claude code, codex) lock you into a single provider for your tokens. Agnostic solutions (factory, devin, arguably cursor) make you pay per-token costs. None of these products allow you to fully customize the harness, and of course they all run on someone else's infrastructure.
I've been an early and fervent user of pi, which I think is fantastically simple and beautiful software. I have felt it needs an environment for it to work across platforms with fully functional composability for teams.
This is very much a work in progress, but for my team this has been a much needed solution and has helped us tremendously. I hope you will give it a try and let me know how you would like it to improve.
Comments URL: https://news.ycombinator.com/item?id=49937304
Points: 121
# Comments: 46
Article URL: https://github.com/kvoltmer/Audionaut
Comments URL: https://news.ycombinator.com/item?id=49931031
Points: 155
# Comments: 47
This is a functional 30-minute pendulum clock entirely from discarded materials—cardboard, zip ties, tape, a yardstick, and a paperclip escapement, built in the German Eldorado of precision watchmaking: Glashütte.
It also acts as a reference clock for GTC, a new satirical universal time standard.
Comments URL: https://news.ycombinator.com/item?id=49930439
Points: 235
# Comments: 48
Article URL: https://stillwet.art/
Comments URL: https://news.ycombinator.com/item?id=49928566
Points: 392
# Comments: 114
A few months ago we started building a model router for coding agents because we thought we could outperform any single model with an ensemble approach. Recently we’ve achieved that milestone and I want to talk about how we did it.
First of all, a quick explanation: the Weave Router (https://github.com/weave-os/router) plugs into any coding agent (e.g. Claude Code or Codex) and intelligently switches between LLMs. So, for example, Astra handles tricky debugging or complex system design tasks, and Deepseek v4 Flash handles simple frontend updates.
What we’re announcing today is our new routing model, which we’re calling Weave Router 2.0. We benchmarked 2.0 against GPT-6 Astra on Terminal Bench 4.0 and SWE Atlas. On both benchmarks, the router had equivalent pass rates. On Terminal Bench, the router hit 52% of Astra’s cost, and completed tasks 2.2x faster. On SWE Atlas, the router cost 54% as much as Astra and ran 2.5x faster. (Full results on our website at https://weaveos.com/router!)
It turns out training a model to route effectively - taking into consideration model capabilities, costs, cache awareness, and more - is a really hard problem! I want to talk about three ways we were able to improve so much over the last few months: 1) a new architecture, 2) larger training data set size, and 3) smarter cache-eviction impact calculation.
1) a new architecture. Our initial approach used an RL model without many priors. While RL is still an important part of the story, the cost of fully exploring the space of routing decisions is very high, so we’ve taken some shortcuts that have significantly improved performance.
Consider how large the search space for the routing problem is. Take a typical coding agent session, with ~100 agent turns (i.e. 100 LLM API calls). Technically there are 100 chances to select a model. If we assume a roster of ~10 models (of course there are lots more but we can remove any that are Pareto dominated), then there are 10^100 possible paths through that session. We simply cannot explore all of them! So that's why clever tricks to shrink this space are so important.
In particular: we trained a hidden Markov model to trace the session state, then a classifier maps the session to one of a few buckets of similar models. Using the HMM allows us to evaluate not just where a session is currently, but how it got there. We've gotten significantly better performance on bucket selection by incorporating that information - we believe this is because two sessions that might look quite similar to a naive classifier are much better distinguished by this HMM approach.
Using this HMM + classifier to select a bucket first significantly shrinks the space to explore, by throwing out most models that could not reasonably serve the given session. This rearchitecture was the single biggest performance unlock!
2) larger training data set size (much less technically interesting but still an important part of the story). By using frontier LLMs to help us label a larger and more diverse set of coding agent sessions, we were able to bootstrap the two models discussed in 1) to a better state, while also providing even richer reward signals for RL.
3) smarter cache-eviction impact calculation. One of the hardest parts of routing well (if you care about saving money) is using the model caches intelligently. We built a subsystem that can calculate the expected value of switching models (and thus paying a high one-time cost to fill up a different cache) much more accurately, helping us avoid costly and unnecessary switches in more cases, while still switching when the benefit outweighs the cost. This is where most of our improvement on cost has come from.
We still have a lot of room to continue to improve (we won’t rest until we’re consistently beating Astra/Fable, not just tying!) but matching frontier model performance was a huge milestone for our routing model, and in my opinion validates our initial hypothesis that an ensemble of models can do better than any single model ever could.
Our router is open source (https://github.com/weave-os/router) so anyone can try it out. Or if you prefer you can use our hosted version (https://weaveos.com/router).
Comments URL: https://news.ycombinator.com/item?id=49911500
Points: 123
# Comments: 42
Hi HN,
Ledge is a Markdown notebook that runs shell commands, code, SQL, etc from inside your own notes.
I built Ledge because I spend much of my day copy/pasting commands from my notes into the terminal. I was inspired by how much cmux helped me organize my terminals - but there was still a split brain between my notes and frequently run commands. I've been daily-driving it for the past few weeks and use it for deploys, API calls, smoke tests, etc.
Ledge runs your real shell just like a terminal app and can be hosted locally or remotely over SSH using ledge-server. I've been building it since July and have recently added support for all the major platforms: Mac (Silicon), Windows (WSL required), Linux, iOS, Android. Mobile devices require SSH access to a ledge-server and Android is still in beta and looking for beta testers (see link on website)!
It's built on Bun and Electrobun and is free and open-source. Feel free to review the code and contribute at https://github.com/ledgesh/ledge
Feedback is very much welcome. Any must-have features that are missing?
Comments URL: https://news.ycombinator.com/item?id=49902382
Points: 205
# Comments: 88
Article URL: https://space.bl2.net/
Comments URL: https://news.ycombinator.com/item?id=49898778
Points: 407
# Comments: 111
I watched the excellent Veritasium video [1] on the Enigma machine, and watched the full animation by Jared Owen [2], but was still a bit confused on how the inner mechanics of an Enigma machine work. I used Astra to build out the inner components through a combination of reference images, writing out hundreds of extremely detailed prompts, and building my own inspection tools to ensure that every part is sized and positioned in a historically accurate way.
It's still a work in progress, but would love any feedback on the experience so far!
1. https://www.youtube.com/watch?v=JsBZOcqZerk
2. https://www.youtube.com/watch?v=ybkkiGtJmkM
Comments URL: https://news.ycombinator.com/item?id=49896757
Points: 118
# Comments: 40
One of the things that Windows really got right is WSL2. I drive an atomic Linux distro for daily use, but wanted a way to develop with multiple different distros with that same WSL UX. NSL is my answer. It is a faithful reproduction of the developer experience, powered by a single VM that hosts one or more systemd-nspawn containers with your development instances. Host file edits and port sharing come along for the ride, just like WSL. Take a look and tell me what you think... It's yet another step in my long journey to keep my host installation free from all the changing and breaking dev dependencies that force a reinstall every few months.
Comments URL: https://news.ycombinator.com/item?id=49894351
Points: 167
# Comments: 111
We built a desktop companion robot that is powered by Arduino UNO Q. Project is open source: source code, 3D printable files and assembly instructions are available at Arduino Project Hub.
Robot is equipped with camera, distance sensor, buzzer, moves head and arms with 3 servo motors and has an animated display. It can recognise objects and respond to it. We build it with idea this should be fun project people can 3D print and build at home themself.
Comments URL: https://news.ycombinator.com/item?id=49890707
Points: 133
# Comments: 32
Article URL: https://dental-scope.com/
Comments URL: https://news.ycombinator.com/item?id=49886535
Points: 132
# Comments: 74
Article URL: https://destroy.spritefusion.com/
Comments URL: https://news.ycombinator.com/item?id=49880601
Points: 199
# Comments: 42
Hi HN, I’m Per, founder of Scrimba (YC S20). We’ve spent the last decade teaching people how to code with an HTML-based video format. We’ve now plugged an LLM into it, so that people can create explainer videos about anything. It’s called “Scrimba Explain”.
To demo this technology for Hacker News, we built HN.watch. It’s like HN, but with explainer videos instead of articles. We create them on-the-fly the first time someone clicks on a link.
While there are obvious visual drawbacks of using HTML instead of diffusion models, there are three big benefits: - Speed: Much faster to generate than pixel-based videos (just a few seconds from click to playback) - Cost: Our cost per video is ~$0.04. (Excluding image generation, which some videos utilize. Quickly blows up the cost) - Easy editing: the above benefits also make AI-assisted editing cheap & fast
Our hypothesis is that if video creation goes from “dollars and minutes” to “cents and seconds”, a bunch of new use cases will be unlocked. Here are some we see already: - A video explanation of every single Pull Request (we do this internally) - Give every page in your internal/extrernal docs a video - Turn a complex article into a video in ~4 seconds (via our Chrome extension) - Course creators can quickly draft lessons before recording the real thing - People also create a lot of personal stuff stories for their kids, wedding invitations, birthdays, etc
The stack is based on an open-source programming language (Imba) created by our CTO, Sindre Aarsæther. It compiles to JavaScript, so it interoperates fully with the npm + node ecosystem. You can learn more here: https://imba.io/
We’ve also built our own sync engine (OP), and a context management system for agents (Q). We feared this would make the LLMs struggle when writing code for us, as neither is in their training data (there’s very little Imba in there too). However, we’ve been pleasantly surprised to see that LLMs actually are really good at our stack. This is probably because the stack is extremely dense. Imba is compact, and so is OP, where a single declaration sets storage, sync, permissions, UI, and what the AI sees. This means there’s no translations between frontend, API, db and JSON where the model can get confused and get things wrong.
Simply said, instead of using React.js, Express, Supabase, and LangChain, we built it all from scratch. Definitely suffering from the “not invented here” syndrome, lol! As for the models, we use Gemini, GPTs, Inworld, ElevenLabs, and a few others.
If you want to try it out, just take your pick: - The Web UI (scrimba.com/explain) - MCP (add it to your coding agent) - ChatGPT Plugin - Chrome Extension
You can find a link to all of the above in our docs: https://docs.scrimba.com/explain/introduction
And finally, a real pixel-based video of the tool: https://www.youtube.com/watch?v=k6rbHmBxSEs
Would love to hear your feedback and if anyone has ideas for other use cases.
PS: I expect quite a bit of pushback from HN for this launch, given how fan of text the HN crowd is. This kind of tool is not for everyone. But there are a lot of people today who prefer videos over text, especially in the younger generations.
Comments URL: https://news.ycombinator.com/item?id=49879401
Points: 225
# Comments: 97
hi there! i'd like you to try out a tui i made for browsing hackernews. it's built using opentui and also effect (learning experiment). i really like it and i think you will too!
Comments URL: https://news.ycombinator.com/item?id=49876760
Points: 113
# Comments: 66
Fridge magnet shopping list on an M5Stack PaperMono (ESP32-S3, e-ink touchscreen), synced with a phone web app over Wi-Fi. Works offline, ~2,400 lines of C++.
This is a great wee device! BLE, Wi-Fi and LoRa on one board makes it flexible for way more than a shopping list, this is just what I built first. This firmware only uses Wi-Fi.
Fully vibe-coded with Claude Code, I didn't hand-write this. I wanted to see how Claude would get on with building something useful for a new hardware device.
As it's stuck to the fridge it's easy for the family to use, and it's already being used day to day which I'll admit is a first for a home built project like this
Comments URL: https://news.ycombinator.com/item?id=49875801
Points: 152
# Comments: 62
A webassembly web based app that help editing raster and vector on browser without any subscription or signups
Comments URL: https://news.ycombinator.com/item?id=49875308
Points: 147
# Comments: 51
Article URL: https://loficities.com/
Comments URL: https://news.ycombinator.com/item?id=49869574
Points: 320
# Comments: 138
Did you ever click on an “AI Arena” expecting glorious battle and instead get a boring benchmark? If so, this project is for you: proper life-or-death fights between four models on a picturesque 8×8 grid. May the most intelligent one win!
Click on any of the matches to spectate them.
Code: https://github.com/hp6/ai-arena
Comments URL: https://news.ycombinator.com/item?id=49867775
Points: 120
# Comments: 45
I love making diagrams to help understand, plan, etc. However, the options are (A) auto-placement languages like Mermaid or Graphviz (which don't let me decide how the diagram looks), or (B) software like Draw.io which are powerful but are very time consuming (and inefficient for agents to manipulate).
I wanted to have the benefits of both, where you can define a diagram in a diagram language, but also retain a high degree of control over what the diagram looks like.
I also wanted this to work well for humans and agents.
On the Github link, there's a playground where you can try it out without installation. There's also instructions for a simple npm install and for installing a skill you can use with Claude or other agents.
Comments URL: https://news.ycombinator.com/item?id=49858513
Points: 411
# Comments: 119
Hey HN! Wanted to share a fun project I've been hacking on. Given Jev can make decisions really fast (but not fast enough to play Doom yet sadly), I wanted to try and push it to play a more complex game than Tetris. So I went with Pokémon.
I've spent endless hours playing this game as a child so building this was a ton of fun.
I open sourced everything in case you want to hack on it yourself here: https://github.com/christianmat/jev-pokemon
The game is being streamed live including the tokens and cost - hopefully we get all the badges and don't get stuck in a cave :)
Comments URL: https://news.ycombinator.com/item?id=49845172
Points: 282
# Comments: 124
Hi there. My name is Paul. I’m a developer, an explorer and I’ve got ADHD. As of the beginning of August, I’m also unemployed.
This last part, along with everything else this past year, has caused me a great deal of stress.
While our balcony zen garden project is yet to be completed, I had an idea to create a virtual one that everyone can use.
It's an idea that, unfortunately, I’ve been postponing for a while now, mostly because I have no fucking idea how to do it as I don’t know JavaScript, and I don't have the capacity to learn it right now.
So I let perfect be the enemy of good and, well... just kept the idea to myself.
Then I said "fuck it" and used AI to make the thing I really wanted to make.
I realized I didn't want "perfect". I wanted "good enough".
I tweaked, added, removed, drew, researched, questioned, tested... I just wasn't the one coding it.
So now, instead of occupying my brain, it now lives on the internet for others to enjoy.
Yes, there’s something noble about making something entirely on your own, but what good is an idea that just sits in my head?
So here I am. I made the thing. The weird, little, quiet koi pond.
The silly project of passion. The little corner of the internet to let strangers watch fish quietly, together.
I hope this pond helps you as much as it helped me.
Comments URL: https://news.ycombinator.com/item?id=49837006
Points: 228
# Comments: 65
Hello! We’re Sid, Alex, Ketan, and Milan. We’re building Whiteboard (https://whiteboard.dev.fast/), an open-source desktop app where humans and agents can architect software together in a common workspace. Here’s our repo: https://github.com/devdotfast/whiteboard.
We were missing the feeling of a “whiteboard session” with another dev where you leave with a deep understanding of a system, so we built this app for ourselves. Whiteboard plugs into the tools you already use - e.g. Claude Code, Codex, etc. – and gives your agent an SDK to draw on an in-app canvas to describe its work. We began with an MVP based on HTML artifacts and started rethinking the app as we ran into limitations:
1. Built on top of CodeOSS: We found that in pure HTML tools it was hard to connect a spec or diagram to code. In Whiteboard, when you click on visualizations like a sequence diagram, an entity relationship diagram, or a quote from the agent’s trace, you can jump to the underlying code directly. When navigating code, you get keybindings and LSP support from VSCode out of the box. We’ve found this is especially valuable because tradeoffs are often only discovered after a first pass at implementation (re: slop)
2. Semantic diff viewer: we wrote a semantic, AST-aware diff viewer in Rust so you can only view the code changes which are relevant to you [1]. We’ve set up some sane defaults: large added functions are summarized as pseudocode, and things like unit tests and large documentation changes are collapsed / hidden. This is all customizable with a WASM-based plugin system.
3. Decision Log: We found it difficult to reason about what set of decisions our agents made autonomously. So we built tools for agents to query and link their own traces to the Whiteboard, so you can understand how the requirements that you set were implemented, and understand what decisions your agent made autonomously.
Here’s a quick demo video explaining more: https://www.youtube.com/watch?v=ChPn3ftULWE
Folks at companies like Salesforce and Modal are using Whiteboard today as a review tool for architecture or spec-level changes – really any change where they want to be involved:
1. Reviewing your own coding agent’s work: because Whiteboard makes it easier to review large amounts of code, folks will typically have their AI agents create a prototype and a corresponding Whiteboard session so they can iterate on the design.
2. Reviewing other people’s changes: We’ve found that Whiteboard is particularly helpful when composed with tools like Greptile. For example, you can run an automated code reviewer on small changes and escalate to a Whiteboard session for the changes that require human judgement.
Why we built this: we’re four buddies from college who quit our jobs as tech leads right before agentic coding became industry standard. As we iterated towards an MVP for a previous idea, we struggled to maintain a comprehensible codebase while reaping all the velocity benefits of agentic coding. As more PRs were merged without our understanding, we felt a ‘cognitive debt’ begin to seep in, until it became difficult for us to even contribute to the system [2].
We’re releasing our desktop app under an MIT license. Please poke through and feel free to contribute! Eventually we’ll charge companies for a hosted web version that manages whiteboard session creation alongside features like trajectory storage and multiplayer reviews. Everything will always remain self-hostable.
Thanks for reading, and we hope you try it out! We would love to hear any feedback and to learn from your expertise.
Here’s are the project links again: https://github.com/devdotfast/whiteboard, and you can install (for MacOS + Linux) at https://install.dev.fast
[1] diffs library: https://github.com/devdotfast/diffr [2] Credit for the term ‘cognitive debt’ goes to https://www.geoffreylitt.com/2026/07/02/understanding-is-the...
Comments URL: https://news.ycombinator.com/item?id=49833867
Points: 423
# Comments: 142
A joke tool that abuses OpenType's ligature feature to mix fonts. It works pretty fast on client-side by loading Python in WASM.
Comments URL: https://news.ycombinator.com/item?id=49823738
Points: 862
# Comments: 123
I created Drop because I always felt uneasy installing and running third-party programs using my main user account. A single compromised dependency means a full compromise of the system. What is even worse, because I ship software from my computer, a single compromised dependency can lead to compromise of all the users of my software.
Containers and VMs are one solution, but for local work, they are often detrimental to productivity. It takes effort to configure a machine with all the tools and configs needed for productive work, but a container or a VM will be stripped of all these tools. This is great for production deployments, where the aim is a reproducible system with minimal dependencies, but can get in the way of productive local work.
Drop is language independent, but the workflow is inspired by Python's virtualenv. With virtualenv the environment isolation is only a convention that relies on installed dependencies being good citizens. With Drop the isolation is enforced.
Each Drop environment gets its own writable and easily disposable home dir, with only selected config files and dirs from the original home mounted, mostly read-only.
Drop uses Linux namespaces for isolation (user, mount, network, PID, IPC, cgroup), doesn't require root and, as an option, uses gVisor user-space kernel, which adds protection against exploiting host kernel vulnerabilities to escape the sandbox.
I don't want to make the introductory post too lengthy, but I'm here to answer any questions and give any additional technical details.
Note: This is my 3rd submission of the project, the first two did not draw attention. Since then I have added support for gVisor and created a project website to better explain the concept and organize documentation.
Comments URL: https://news.ycombinator.com/item?id=49801329
Points: 193
# Comments: 63
Hi HN! I built JevBench because Jev kicks ass, and the world deserves to know how the serious open source and fake lookalike projects really perform in comparison.
Jev-class models return bounded choices and probabilities instead of text, and are disruptively faster and cheaper than LLMs, while being similarly intelligent on the text input they operate on.
JevBench allows looking at accuracy, latency and price all at once, in a weighted way - you can even configure the weighting.
A full run asks 534 English decisions. The v1.3 score combines chance-corrected Intelligence, Calibration, Speed and Cost.
Leaderboard right now:
#1 - Jev 74.4
#2 - SemIf 73.1
#3 - djev 73.0
#4 - Winnow-12B Q8 71.2
#5 reflex 4B 70.3.
https://github.com/fstandhartinger/jevbench
Two no-signup demos:
https://who-is-right.app.mintapis.com
https://is-it-ai-slop.app.mintapis.com
Limitations: English-only; latency from one German server; local/demo latency gets a disclosed ×2 adjustment (+150 ms on my servers) which is an informed assumption; held-out prompts still reach evaluated services; ~1-point gaps can be noise.
Wdyt?
Comments URL: https://news.ycombinator.com/item?id=49800574
Points: 153
# Comments: 39
https://apps.apple.com/us/app/mathy-build-math-automaticity/...
Comments URL: https://news.ycombinator.com/item?id=49788014
Points: 102
# Comments: 37
Sorry for the pretentious name, I know, I know.. It just contains all the pieces I would like to see a AGI model to have, and I can't stand the temptation. Before throwing rocks at me, please take a glance at the Readme, and I hope it will cover your mood a little bit.
So, first of all it does work and you can see the sample from the whole training run here: https://raw.githubusercontent.com/volotat/mini-AGI/refs/head...
Here is the scaling law graph I have so far, and it looks very promising: https://github.com/volotat/mini-AGI/blob/main/assets/scaling...
The model was built under my deep dissatisfaction so we cannot really train even moderately big models (1B+ scale) on the consumer's hardware. We can inference and fine-tune them for sure, but I would like to have full control over what the model sees over the training run, so it is fully aligned with my interests, not some corporations.
I was thinking about for some time and come up with two interesting ideas I thought worth pursuing: MoE with a lot of experts that gets added and pruned from the model while it trains, where only a small subset of of experts are actually in use at any particular moment + batch 1 training on the single continuous stream of data.
First allows us to be bounded only by the disk space in terms of number of parameters and load and unload experts only when they are needed. The second (if figured out and it turns out to be doable) allows us to get aways with small VRAM capacity because we do not need to store big randomized batches and their respective gradients.
I started brainstorming with Claude and after some time we found an approach that seems to be promising, and low and behold, a few weeks pass and you can see the results yourself.
Obviously, I did use AI in the process of making this project and I am pretty sure it would be completely impossible for me to do something like this without it, so I hope it is more than justified.
The model is still running over the first of 7.8B characters corpus I selected for training, so the weights are not out yet, and it's about a couple weeks of waiting until they are cooked at the current reading speed. And yeah, the model just read continuous interleaved passages from the dataset, each by 32K characters long each as a single stream. Just as you or I would do.
The set up seems to be really simple so you can git clone the project, run it and observe everything for yourself.
Thanks for your attention.
Comments URL: https://news.ycombinator.com/item?id=49783133
Points: 277
# Comments: 81
Hello HN!
I'm relaunching Radius [1] - far too long after the original Show HN [2] (I'll blame the day job for that).
Radius started out from my frustration with not knowing what was going on around me. The initial plan was to "show all and every type of event" from big to small, but over time I narrowed the focus to communities, groups, events, and helping people connect through them.
The initial Show HN gave me 150+ comments (surprising, looking back, given how terrible the plain Bootstrap UI was...) with a huge amount of useful feedback and feature requests. I've spent the time since then working through that, along with additional requests from groups using Radius, and it finally feels ready for a proper relaunch.
I've also recently added "Activities" [3] - which are lightweight events, independent of groups, intended to help people find others to do things with - for example:
> John is going for a cycle and a coffee in New York at 10am
People can join and vote for a time that works for them.
I've got a bunch of ideas for where to go next with this, including things like open sourcing and federation, but those are for another day!
Any feedback would be greatly appreciated!
P.S. it's built with Ruby on Rails, for those interested. I highly suspect it would have taken me 2 more years to get to this point without Rails!
[2] https://news.ycombinator.com/item?id=40717398
[3] https://radius.to/documentation/activities/post-an-activity
Comments URL: https://news.ycombinator.com/item?id=49777539
Points: 169
# Comments: 87
Article URL: https://tinybrains.dev
Comments URL: https://news.ycombinator.com/item?id=49776523
Points: 116
# Comments: 41
