news.ycombinator.com
Qwen3.8 Fetches Weird URLshttps://news.ycombinator.com/item?id=49379079I am working on opencode with the last qwen3.8-27B from unsloth (with dynamic 3.0 quants).while checking documentation for a rust project, instead of fetching the url it had listed in its own reasoning, it tried fetching urls like:https://routify-file-proxy-sg.oss-ap-southeast-1.aliyuncs.com/proxy_temp_file/production/2026-08-20/trace_0baf8c2b17874702866132490e0b56/requestId_97422c0041074091a68367881782d7b7/7c0791038d2f58b3a52c0f07714d3627.html?Expires=1818574288&OSSAccessKeyId=LTAI5t…QVZr&Signature=96y%2B…w%3Dthis happened in the same session, different batches, 16 times in total.It convinced itself there was some rewriting proxy in the webfetch tool.`Expires`, `trace` and `requestId` were consistent for a few requests, but they do change.`AccessKeyId` was always the same, "LTAI5t…QVZr"I did not add the `…` in the signature and access key.Probably training links, but that `Expires` number seems very high...in unix timestamp:1818574288 -- Aug 20271787251180 -- nowJust leaving this hereAug 20, 2026 7:32 PM
news.ycombinator.com
Show HN: We Added Notion Calendar to *Superhuman for iMessagehttps://news.ycombinator.com/item?id=49378903Another update on Attention: after launching here last week, over 100 people used Attention!We followed up on feedback and we added a Notion Calendar-style calendar!---We realized pretty quickly when working in Attention (Superhuman for iMessage) that a lot of our conversations were about scheduling things.When we talk to Righthand users, iMessage is a lot of "can we find time to debug the issue live" or "I'll get back to you by this evening with the answer". Personally it's a lot of booking travel and dinners.Since we've been spending so much time in Attention, we decided to add a calendar to fix the issue. It looks like Notion Calendar, but it also supports a few new features like "fuzzy events".The onboarding is more involved and so we'll do 1:1s for a bit (comment below if you want to be in the cohort), but we would love any and all feedback!As before:- Privacy / Security: Your data is yours, fully local in LanceDB + SQLite, never leaving your machine unless you turn on the AI agenAug 20, 2026 7:17 PM
danielvaughn.dev
Show HN: Huzzah – a novel approach to coding with AIhttps://www.danielvaughn.dev/posts/huzzah/Hello everyone. I've been working on this experimental editor called Huzzah.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.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: 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. It may not work for every use case, but in my initial playthroughs I've found it very enjoyable.Right now it's just a proof of Aug 20, 2026 7:05 PM
news.ycombinator.com
Are you good at AI, or just using it?https://news.ycombinator.com/item?id=49378057We’re working on a ladder for individual AI proficiency and would love feedback on both the levels and the definitions.L0 New: brand new to AI, or has not yet used it.L1 Chat: simple prompt-and-response use. Work is serial: ask, wait for an answer, then ask again.L2 Contextual Work: gives AI relevant documents, data, or workspace context so it can work within the actual artifact and produce a more useful result.L3 Orchestrate: coordinates multiple agents or AI roles across independent workstreams, with work that may review, challenge, compare, or build on other work. This is not just for engineering.L4 Automate: creates workflows that are triggered by business events and run without someone sitting at a laptop directing each step.L5 Loop: feeds the output of those workflows back into shared knowledge or a company brain, so future workflows improve over time.A few things I’d love your perspective on:Are these the right levels? Are any of the names unclear or overlapping? What observableAug 20, 2026 6:04 PM
radi8.dev
Show HN: We chased a weather balloon across Montana and never found ithttps://radi8.dev/blog/uplink/Since April, I have been working with Sam Flynn (https://drook.dev) to make this balloon payload, UpLink. We did a similar launch last year with Hack Club but this was our first independent launch.UpLink was a 491 gram payload testing the insulation properties of 3D printing filaments, while also transmitting 320x240 images over a radio link -- up from the 18x10 images last year!This is a writeup on our engineering process, mistakes made, and learning experiences. It covers:- Custom electronics designed in KiCad- Firmware design- Results from the data we received on the ground- Image transmission- Launch day logistics, and where things went wrongAll hardware, software, firmware, and CAD is available on GitHub: https://github.com/radeeyate/UpLink, licensed + certified as open source hardware: https://certification.oshwa.org/us002826.htmlIf you just want to see the images received, I put up a gallery here: https://uplink.gallery.radi8.dev/If you have any questions, comments, or concerns,Aug 20, 2026 5:33 PM
news.ycombinator.com
Ask HN: How to engage users in an open source project and when to stop?https://news.ycombinator.com/item?id=49377336I have been working on an OS project for years now, on and off, taking it back up for the last few months and giving it a good boost lately with AI. It has a super niche user base in quantitative finance.I really enjoy working on it, and although I get contacted every now and then by people who are using it or are interested in using it, and those are really interesting users, I always think I am wasting my time. It has been 8 years now since I started.I guess my questions are: is it worth it? And I know this question has two clear paths, the personal and the practical one, but I would like to hear from both. How can I know when to stop? If my audience is maybe in the order of two thousand potential users worldwide, how do I validate the idea? I have posted on Reddit and HN multiple times, tried LinkedIn, and that leads me to a handful of super interesting contacts, but I don't know if it is enough to keep going.Aug 20, 2026 5:09 PM
scibly.com
Show HN: Open-source tool turning company knowledge into Duolingo-like courseshttps://www.scibly.com/de/My buddy and I have been working on a side project for the past few months called Scibly.The idea is pretty simple:Companies already have an incredible amount of knowledge sitting in PDFs, documentation, internal wikis, etc. The problem is turning all of that information into something employees can actually learn from and retain.For this reason we decided to build Scibly, a tool that automatically turns boring company knowledge into interactive, Duolingo-like (but better) learning experiences.Today, we decided to open-source the whole thing.We're still pretty inexperienced when it comes to open source, so we'd genuinely love to hear your feedback, whether that's about the product, the idea, the codebase, or how we're approaching the project.Short product demo: https://www.youtube.com/watch?v=TcpLUNBRhQwGithub: https://github.com/scibly-dev/sciblyWould love to hear what you guys think!Aug 20, 2026 1:11 PM
news.ycombinator.com
Offline RAG on iOS with Spatial Integrationhttps://news.ycombinator.com/item?id=49371913I'm the developer behind CartoType. I’ve been working on bridging local language models with offline mapping, and I have just put together a demo of a completely offline Spatial RAG pipeline running natively on an iPhone. The new system is named the CartoType Field Assistant. You can find the website at https://cartotype.comDemo (1m 19s): https://www.youtube.com/shorts/a8yQPn7_jyIUse cases: Any organisation with field technicians or emergency first responders needs complex procedural knowledge tied to physical locations ('assets') where they don't have guaranteed network connectivity. Examples include offshore wind farms, power distribution networks, railway infrastructure, mountain rescue, and military uses.For this demo I query the iPhone app in Airplane mode: "A hiker near Tuolumne Meadows has a dislocated shoulder. What is the reduction protocol, and where is the Tuolumne Meadows Ranger Station?"HOW IT WORKSThe core is a portable C++ engine running vector search across an encryptedAug 20, 2026 8:28 AM
bing.com
Fish an Anime RNG codes (August 2026)http://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8956b0233d479c86423293b2d30bff&url=https%3A%2F%2Fwww.dexerto.com%2Froblox%2Ffish-an-anime-rng-codes-3400568%2F&c=12240737762254287237&mkt=en-usHere are all the working Fish an Anime RNG codes for August 2026, including Cosmic Boxes, Cash, and useful Potions.Aug 20, 2026 2:46 AM
lutefish.com
Show HN: Lutefish Stream (Play Music over Internet)https://lutefish.com/Hey all, Disclaimer: I work on embedded software for Lutefish.tldr: If you have good internet and you live within 500 miles of somebody (and ISPs are feeling benevolent that day) you can play music together over the internet (for real tho).I know you've probably seen "make music over the internet but for real this time" presented from a handful of places over the years. We designed a PCB from scratch to tightly couple the audio and networking layers as close as we possibly could, and we are able to get latency numbers as low as 28ms using custom protocol (which allows for seemingly real-time performance with bandmates over the internet).Before you ask, no, it is not an audio interface also (we're working on it). Think of it more like, "point-to-point sound cards with a lot of control over your mix".Don't really know how to do self promotion, but we have devices for sale for $250 usually (rando $200 flash sales sometimes) and there's a $10 / month subscription (2 months free w/ purchaseAug 20, 2026 1:36 AM
nytimes.com
Prince Harry and Meghan Will Move Back to U.K. After 6 Years in U.S.https://www.nytimes.com/2026/08/19/world/europe/prince-harry-meghan-return-uk.htmlThe prince and his wife, whose relations with King Charles and the rest of the royal family have long been strained, will remain non-working royals, according to a person familiar with the decision.Aug 19, 2026 11:40 PM
cbsnews.com
Prince Harry and Meghan moving back to U.K.https://www.cbsnews.com/news/prince-harry-meghan-moving-back-uk/The couple said in January 2020 that they would be stepping away from "senior" royal roles and were working to become financially independent from the royal family.Aug 19, 2026 11:27 PM
systemist.net
Show HN: Dispositio – a TUI project planner that stores plans as plain Markdownhttps://systemist.net/tools/dispositioHi all! Wanted to show you guys Dispositio, a project planning TUI that saves projects to simple markdown.I always have a ton of solo projects cooking, and in the past organized and planned them in increasingly complex `ROADMAP.md` files. After working with bubbletea on another project, and after constantly manually calculating its delivery dates for myself (as my bandwidth was quite limited at the time), I realized it would be cool to combine the two!The result is Dispositio, a tool which:- Tracks milestones, set estimated durations, rearrange and modify easily, and see the estimated timeline of the whole thing (measured in weeks for now; can add sprints or other metrics later if people want it) - Lets you set up tasks and subtasks right in it (added because I found it very handy to just keep my todos right here, one per project) - Handles overdue items etc - Helps you know how many days you get on average per milestone task in order to stay on task - Saves to plain markdown so you caAug 19, 2026 9:44 PM
sports.yahoo.com
Blue Jay runners are at a good starting pointhttps://sports.yahoo.com/articles/blue-jay-runners-good-starting-164600283.htmlAug. 19—JAMESTOWN — Times are looking good for the Jamestown High School cross-country team. "We've just got to keep building from where we are at," JHS head coach Ken Gardner said. "As long as they keep working and keep doing what we ask them to do, we would certainly predict they will get faster and stronger." Gardner and the 20 athletes who committed to run cross-country this season began ...Aug 19, 2026 4:46 PM
github.com
Launch HN: OneCLI (YC S26) – OSS sandboxed agent harness for teamshttps://github.com/onecli/onecliHi HN, Jonathan & Guy here from OneCLI, an agent harness built for teams, giving every employee a secured, sandboxed personal agent.Here’s what you can do with it:1. get a sandboxed agent, with all the OneCLI capabilities in place like connect your GitHub account, Gmail, Notion, or Dropbox simply from the chat.2. deterministic human in the loop approval in the chat itself for things that you need 100% control like sending an email or deleting the Linear ticket.3. manage team policy in one place, enforced across every agent in the workspace4. enjoy global connections at the team level, like shared LLM keys or service accountsHere’s a demo: https://www.youtube.com/watch?v=dlW-44ntpbEWe started working on this by accident, even though our careers were in the security space. We were working on a devtool called ChartDB, an open-source DB tool. When OpenClaw took off back in January, we started using it to orchestrate agents on top of ChartDB. We quickly understood there is a big issue arounAug 19, 2026 4:29 PM
variety.com
Nielsen Makes New Bid to Capture ‘Co-Viewing’ as TV Season Approacheshttps://variety.com/2026/tv/news/nielsen-co-viewing-tv-ratings-1236838209/Nielsen is really working the room. The media-measurement giant aims to measure viewers of video content more accurately and will place a focus on capturing so-called “co-viewing,” or multiple viewers watching the same piece of programming together at the same location. The effort means TV executives may get more credit for shows that bring people […]Aug 19, 2026 4:00 PM
github.com
Show HN: a javascript engine written in C3 with native TS supporthttps://github.com/ricardobeat/boomkatHi HN! For the past three months I’ve been working on boomkat, a fully ES compatible, strict-only engine. It was written from scratch using the C3 language, with Duktape and QuickJS as reference engines.The engine passes 100% of test262 (subset, excluding legacy features, sloppy mode and proposal stage features), 10x faster than Duktape, matching QuickJS in performance, and can natively run TypeScript files including module support. Supports all modern JS features.It was built with a mix of several open-weight models and harnesses, with Claude Code as reviewer. I estimate 300+ hours of work, plus agents running nearly 24/7 for 12 weeks. 2000+ commits, all read (but not fully reviewed) by me. About €300 spent in total.As the README warns, this is not production-ready code, but I think it's at a state worth sharing. Surprisingly test262 still leaves a lot of surface uncovered, there are still small issues here and there but it runs a lot of real world code successfully (e.g. Zod).Would lAug 19, 2026 3:27 PM
bloomberg.com
Target Raises Sales Outlook as Turnaround Gains Strengthhttps://www.bloomberg.com/news/videos/2026-08-19/target-raises-sales-outlook-as-turnaround-gains-strength-videoJoe Feldman, senior research analyst and assistant director of research at Telsey, says there is a “strong signal that the turnaround is working” at Target. The retailer lifted its full-year guidance after results outpaced estimates in the second-quarter. (Source: Bloomberg)Aug 19, 2026 3:21 PM
github.com
Show HN: BRep Geometric CAD Kernel and Parametric Code CADhttps://github.com/yuechen-li-dev/Aetheris/I'm pretty excited to show off something I've been working on for the past few months. A free, open-source geometric kernel that allows humans/LLMs to write code (in this case, a DSL called Firmament) and generate 3D models of CAD parts, to finally provide an alternative to OpenCascade.To get some questions out of the way. - Full STEP import/export support for AP242, AP203 and AP214 are still experimental at the time, but everything produced by Aetheris should open in any CAD app with STEP support. - Support for single edge/planar fillets/chamfers right now, should be more than sufficient for most CNC/3D printing use cases. - Fillets: https://github.com/yuechen-li-dev/Aetheris/blob/master/testd... - Chamfers: https://github.com/yuechen-li-dev/Aetheris/blob/master/testd... - Supports dimensional/GD&T annotation via STEP 242's semantic PMI. - Kernel is written entirely in C#, with bindings for Go, Rust, Python, and TypeScript available.If you just want to vibe-CAD, clone the repo, run itAug 19, 2026 3:09 PM
news.ycombinator.com
Enshitification Comes for Cursorhttps://news.ycombinator.com/item?id=49362593Sometimes I feel like the last person on HN that's still using Cursor for development work. But if anyone is curious - its getting a lot shittier, quickly.The entire application seems to exist to steer you towards using Grok and whatever new product Cursor's team has pumped out this week. You're constantly prompted to switch to using a Cloud agent, including the UI seemingly changing as you're working which caused me to accidentally select a cloud agent instead of my local system a few times. There are daily pop-ups in the app to check out this and that new thing, these are at least pretty easy to close and ignore. This morning Cursor decided to select a new Grok model by default without prompting, which I didn't notice until it had spent 10 minutes burning tokens on a simple bug investigation that Opus resolved in 30 seconds.I finally bit the bullet and setup Pi with Open Router the other day. I still prefer the UI of Cursor but this is like death by a thousand cuts.Aug 19, 2026 3:04 PM
