The best Hacker News stories from Show from the past week
Latest posts:
Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash
Hey HN, Henry from Cactus here.<p>We submitted Needle 2 here a few weeks ago, and the feedback in the discussion thread was incredibly valuable, thanks! Thanks to all that feedback, we’ve been able to move quickly to release Needle 3 and I'd love to hear what you think again.<p>The key features:<p>1) Automation (tool calls & structured JSON output): Needle still doesn't chat by design, its quite challenging to pack general capacity into such small models, so we focus on tool calls and structured JSON. If no tool you declared fits the request, you get an empty list back (note for when playing with the demo).<p>2) Intelligence Laddering: Every layer (2 to 20) is a deployable subnetwork, so one set of weights, 25 to 121 million parameters at 2-bit, shipping as 8-29MB binaries. On a Raspberry Pi 5 it decodes at up to 4k tokens/sec and prefills at up to 10k.<p>3) Monarch Hadamard MLP: replaces the dense FFN with three learnable Walsh-Hadamard-initialized Kronecker (Monarch) factor pairs interleaved with per-channel diagonal scales, fixed permutations, a SiLU nonlinearity, and a rank-8 input-conditioned gate, so each token gets a fully mixed nonlinear transform of its d_model channels at O(d√d) parameters and compute instead of the O(d²) a dense 4x-expansion MLP would cost.<p>4) Performance: On Mobile Actions (phone commands, scored on the exact call) the 20-layer model gets 86.0 through the shipped 2-bit binary; LFM2.5 1.2B is at 82.4, Qwen3.5 0.8B at 76.0, Apple's on-device model at 57.6, all at f16. More results on the link, we do not win everywhere ofc.<p>5) Multilingual: Needle 3 now supports English, French, Spanish, German, Dutch, Italian, Polish, with more languages coming.<p>6) Finetuning: You can achieve DeepSeek v4 Flash grade performance on a narrow task with just 4L, stress on "narrow task", we found that production users often prefer tuning before production.<p>7) Triggers: Grounding is a common challenge for tool call, at least for Needle 2, so we added support case-insensitive regular expressions matched against each request to gate false negatives.<p>8) Confidence: Every response also carries a calibrated confidence score, the minimum of a judgement on the finished call and its decode probability. Act above your threshold, show the call and ask below it, or escalate to a bigger model.<p>9) Supported Platforms: macOS, Linux on x86-64, ARM64, ARMv7, RISC-V and MIPS32, Windows x64 and ARM, Android, iOS, watchOS, tvOS, the browser as WebAssembly, and a WASI component.<p>Thanks for reading and as always, thoughts appreciated!
Show HN: Share your AI Setup, Learn from others
I kept seeing engineers share what they were building with AI; however, I was always more curious about how they worked. Which agents did they use? What skills and tools had stuck or been thrown out the window? How did they manage longer-running tasks? So I built this with the hope we could have a dedicated space to share and be open about our setups.
Show HN: I made a flight simulator, except you're just a passenger
Buckle your seatbelt, secure your tray table, and open your window shade. Now you can simulate flying anywhere in the world as a commercial passenger, from takeoff to touchdown. Terrain, weather, and realtime sun position included. Don't worry, the latest release now features legs, so you can get up and go to the bathroom on longer flights. Enjoy!
Show HN: Redis City – Explore how Redis works in an interactive 3D model
Show HN: Hacking a $20 4G wireless hotspot into a texting device
Show HN: Capsule – Single-file web apps that save their data into SQLite
Hey HN,<p>I always had the problem that building HTML pages is really simple now, but trying to save data required hosting it somewhere, and sharing it afterwards was not easy. Over the last few months, I've been building an app called Capsule (it’s also the file extension name) written in Rust with Tauri 2.0 that allows packing an HTML app and its data into a single SQLite file.<p>The HTML file and any related assets are directly embedded in the database. User data can either be saved as a localStorage key/value store or via a MongoDB-inspired collections API as documents, saved in a table in the file. You can also save other assets, like PDF files or images, directly in the database to keep different documents together. All data can be easily exported to CSV or JSON if needed.<p>Privacy and security were a big priority for me, so documents cannot do anything out of the box. They don’t have direct access to the file system and they require permission to access the internet. The permission model is still something I’m working to improve. Capsule documents can also use local or remote AI models for document specific AI features.<p>One downside with this approach is that multiple people working on it will create different copies. To make it possible to merge different copies of the same file, each data entry has a unique UUID and timestamp.<p>I’m planning to open up the file format specification for the 1.0 version of the app so other apps can read or write Capsule files.<p>You can try it out in the web preview at <a href="https://withcapsule.app/preview" rel="nofollow">https://withcapsule.app/preview</a> with pre-built templates or use any AI provider of your choice to create a custom, Capsule-optimized app by using the following prompt:<p>"Please read the app wizard instructions at <a href="https://withcapsule.app/prompt.txt" rel="nofollow">https://withcapsule.app/prompt.txt</a> and help me design an app.“<p>I’m still working on the file format but there are migrations for each new version, so data should never be lost when using newer versions of the app in the future. Please let me know if you have any ideas or use cases where this might make sense or does not work.
Show HN: An e-ink frame that hears birds and draws them as 1800s illustrations
See also: <i>Avian Visitors</i> - <a href="https://news.ycombinator.com/item?id=48343424">https://news.ycombinator.com/item?id=48343424</a> - May 2026 (20 comments)
Show HN: An e-ink frame that hears birds and draws them as 1800s illustrations
See also: <i>Avian Visitors</i> - <a href="https://news.ycombinator.com/item?id=48343424">https://news.ycombinator.com/item?id=48343424</a> - May 2026 (20 comments)
Show HN: Neobrutalism.dev – Just added Base UI support and added new color theme
Show HN: Hacker News, Without AI
Show HN: Hacker News, without AI
Show HN: Bodily Oddities
When I was about 11 years old, my best friend and I were playing during recess at school, and I was carrying him around on my back, presumably pretending to be a multipart attack robot. All of a sudden, my heart started hurting, and I collapsed to my knees, and the robot was no more. Every time I inhaled, it would feel like a spike was being driven into my heart, and my breathing got completely shallow. I was sure I was done for, when all of a sudden, after an exhale, it disappeared.<p>It wasn't until I was in my twenties that I learned this was called Precordial catch syndrome and it happens to most everyone. Every friend I've told about it was surprised, and also most knew the feeling.<p>I collected a growing list of strange things bodies do ever since. And now finally I made it into a little page for others to see:<p><a href="https://vester.si/bodily-oddities/" rel="nofollow">https://vester.si/bodily-oddities/</a><p>It includes a form to tell me about new things, if you think they belong on the page. Feedback welcome!
Show HN: Bodily Oddities
When I was about 11 years old, my best friend and I were playing during recess at school, and I was carrying him around on my back, presumably pretending to be a multipart attack robot. All of a sudden, my heart started hurting, and I collapsed to my knees, and the robot was no more. Every time I inhaled, it would feel like a spike was being driven into my heart, and my breathing got completely shallow. I was sure I was done for, when all of a sudden, after an exhale, it disappeared.<p>It wasn't until I was in my twenties that I learned this was called Precordial catch syndrome and it happens to most everyone. Every friend I've told about it was surprised, and also most knew the feeling.<p>I collected a growing list of strange things bodies do ever since. And now finally I made it into a little page for others to see:<p><a href="https://vester.si/bodily-oddities/" rel="nofollow">https://vester.si/bodily-oddities/</a><p>It includes a form to tell me about new things, if you think they belong on the page. Feedback welcome!
The same nine streaming subscriptions cost $702/year more than in 2021
Show HN: What if the speed of light was 5 km/h?
I've always wanted to make a visualization where the speed of light was scaled down to human speeds, so that we could intuit relativistic effects with everyday objects. Here is the first version of it!
Show HN: LLM Attention Visualization
Show HN: Copperhead – Cursor for circuit boards
Show HN: Stuxnet – A reconstructed source code of the infamous cyber-weapon
Stuxnet! Here reproduced by me. Only researchs educations purposes.
Show HN: TERMy – A fast terminal assistant that does not use LLMs
I love research and development, you may have heard of me because of PJON (Padded Jittering Operative Network). It is a network protocol I started developing in 2010, which was recently implemented in silicon by the ETH Zurich university thanks to the research of Pius Sieber.<p>I am excited to share with you TERMy, a terminal assistant built on top of the NPC-Forge framework. Unlike everything else being built today, TERMy does not use embeddings, machine-learning or LLMs. It runs on the CPU (even on a Raspberry Pi Zero) both in the terminal or client-side in a browser tab and responds in milliseconds. It is a cynical but very knowledgeable Linux terminal assistant that translates your natural language into shell commands without relying on a single artificial neuron.<p>I had a chance to focus for 2 months on my personal projects since early July, during the strange times of AI price hikes and the end of subsidized tokenmaxing. I was curious to see if I could develop from scratch a terminal assistant capable of handling simple natural language requests. I have a bad memory and got used to ask to copilot "activate the virtual environment" or similar trivial operations spending a non negligible sum every month. I started thinking, maybe I can do something to make my workflow more efficient? Do I really need trillions of parameters to accomplish those tasks?<p>How it Works<p>When you type a prompt, it goes through a lightweight NLU pipeline written in ~1000 lines of Python that implement the following steps:<p>1. Strip expletives, interjections, encouraging, discouraging and thanking words (remove noise)<p>2. Sentiment analysis<p>3. Exact Match (very fast)<p>4. Template Match (slower)<p>5. Probabilistic Match (even slower)<p>Step 5 relies on:<p>1. IDF (Inverse Document Frequency) to identify rare words.<p>2. BOW (Bag Of Words) to accommodate word inversions.<p>3. IDF weighted Levenshtein to safely handle typos.<p>Permission gating is hardcoded into the dataset and enforced for all potentially destructive commands, so it's inherently safer than letting an unpredictable LLM run wild on your machine.<p>- TERMy in operation: <a href="https://www.youtube.com/watch?v=qeIp0xePLBg" rel="nofollow">https://www.youtube.com/watch?v=qeIp0xePLBg</a><p>- Variance and typo tolerance: <a href="https://www.youtube.com/watch?v=tQvGDk6fkk0" rel="nofollow">https://www.youtube.com/watch?v=tQvGDk6fkk0</a><p>- Copilot integration: <a href="https://www.youtube.com/watch?v=Wzzouhq2a8A" rel="nofollow">https://www.youtube.com/watch?v=Wzzouhq2a8A</a><p>- Advanced features: <a href="https://www.youtube.com/watch?v=qeIp0xePLBg" rel="nofollow">https://www.youtube.com/watch?v=qeIp0xePLBg</a><p>- Source Code: <a href="https://github.com/gioblu/NPC-Forge" rel="nofollow">https://github.com/gioblu/NPC-Forge</a>
Show HN: TERMy – A fast terminal assistant that does not use LLMs
I love research and development, you may have heard of me because of PJON (Padded Jittering Operative Network). It is a network protocol I started developing in 2010, which was recently implemented in silicon by the ETH Zurich university thanks to the research of Pius Sieber.<p>I am excited to share with you TERMy, a terminal assistant built on top of the NPC-Forge framework. Unlike everything else being built today, TERMy does not use embeddings, machine-learning or LLMs. It runs on the CPU (even on a Raspberry Pi Zero) both in the terminal or client-side in a browser tab and responds in milliseconds. It is a cynical but very knowledgeable Linux terminal assistant that translates your natural language into shell commands without relying on a single artificial neuron.<p>I had a chance to focus for 2 months on my personal projects since early July, during the strange times of AI price hikes and the end of subsidized tokenmaxing. I was curious to see if I could develop from scratch a terminal assistant capable of handling simple natural language requests. I have a bad memory and got used to ask to copilot "activate the virtual environment" or similar trivial operations spending a non negligible sum every month. I started thinking, maybe I can do something to make my workflow more efficient? Do I really need trillions of parameters to accomplish those tasks?<p>How it Works<p>When you type a prompt, it goes through a lightweight NLU pipeline written in ~1000 lines of Python that implement the following steps:<p>1. Strip expletives, interjections, encouraging, discouraging and thanking words (remove noise)<p>2. Sentiment analysis<p>3. Exact Match (very fast)<p>4. Template Match (slower)<p>5. Probabilistic Match (even slower)<p>Step 5 relies on:<p>1. IDF (Inverse Document Frequency) to identify rare words.<p>2. BOW (Bag Of Words) to accommodate word inversions.<p>3. IDF weighted Levenshtein to safely handle typos.<p>Permission gating is hardcoded into the dataset and enforced for all potentially destructive commands, so it's inherently safer than letting an unpredictable LLM run wild on your machine.<p>- TERMy in operation: <a href="https://www.youtube.com/watch?v=qeIp0xePLBg" rel="nofollow">https://www.youtube.com/watch?v=qeIp0xePLBg</a><p>- Variance and typo tolerance: <a href="https://www.youtube.com/watch?v=tQvGDk6fkk0" rel="nofollow">https://www.youtube.com/watch?v=tQvGDk6fkk0</a><p>- Copilot integration: <a href="https://www.youtube.com/watch?v=Wzzouhq2a8A" rel="nofollow">https://www.youtube.com/watch?v=Wzzouhq2a8A</a><p>- Advanced features: <a href="https://www.youtube.com/watch?v=qeIp0xePLBg" rel="nofollow">https://www.youtube.com/watch?v=qeIp0xePLBg</a><p>- Source Code: <a href="https://github.com/gioblu/NPC-Forge" rel="nofollow">https://github.com/gioblu/NPC-Forge</a>