Hacker News Reader

ghrlon MacBook Neo is so popular that Apple doubled production

I would say if it's only used as the build and publishing device and development happens elsewhere, this would work without problems. 8Gb for building the iOS app and testing on a real device or even an emulator would likely work. Apple's swap is also quite fast.

ghrlon Muxcard, a DIY credit card size computer

I was thinking all that too and considering commenting about being sick of those credit card size claims, but after seeing the footage I am genuinely impressed. Great work there.

ghrlon Microsoft builds MacBook Pro rival with NVIDIA-powered Surface Laptop Ultra

Wow, very interesting example.

ghrlon Meta launches Instagram, Facebook, and WhatsApp subscriptions

I believe you can get close to that with various Instagram mods for Android. They have advanced features like only show posts from followers and stuff like that. I switched to one of them after realizing the 5 second ad breaks I got were only there because I disabled personalized advertising.

ghrlon OpenRouter raises $113M Series B

I would disagree. Having all the messages locally and sending them with the request means you can switch inference providers or even models mid-conversation. It also means that the provider doesn't store the entire context, which often contains massive parts of proprietary codebases, secrets and PII and instead the agent harness manages all that. While a simple `continue thread` API field might seem more convenient, the cost is still determined by the input token count and cache rate, so it just abstracts this crucial implementation detail away.

ghrlon Show HN: Continue? Y/N: A 60-second game about AI agent permission fatigue

I am mostly using OpenCode and barely ever see a permission prompt. While they do enforce it for outside workspace read/write, with the bash tool the agent can just bypass that. I'm not quite sure why it is that way, and it certainly isn't a very good solution, but likely not worse than asking for everything which just trains the user to always accept and provides a false sense of security then.

ghrlon Indoor Wi-Fi Roaming with OpenWRT

I don't quite understand the benefit of the setup. If there are legacy IoT devices that need unique named 2.4G network, just broadcast another SSID for them. So each router broadcasts main 5G (common name, fast roam etc), main 2.4G (same as above) and legacy IoT 2.4G (with a different name for each AP, and possibly worse encryption and maybe even TKIP). That wouldn't hold back the network for legacy devices.

ghrlon Outsourcing plus local AI will soon become more economical vs. frontier labs

I would say that is highly unlikely if by SOTA models you are not just referring to coding benchmarks but more general purpose ability and domain-specific knowledge. For example Kimi 2.6, which is comparable to Opus 4.6, is roughly 500+GB large, and I don't see how that would run on consumer hardware anytime soon. Besides, this is not just about the technical feasibility, but also economically not viable whatsoever. Why should consumer laptops be capable of running such models, when they would be massively underutilized most of the time, when inference providers can produce the same results faster, cheaper and a lot more viable economically?

ghrlon SQLite Is a Library of Congress Recommended Storage Format

Something MacOS and Windows support natively would be a good start, it could grow from there.

ghrlon Claude Code's source code has been leaked via a map file in their NPM registry

Let's hope they left the having-fun part for a human to do.

ghrlon Show HN: Raincast – Describe an app, get a desktop app (open source)

Well, while Tauri is certainly nice, it's not quite what I imagine when I hear "native".

ghrlon SSH Secret Menu

I'm using that as well but had issues with tunneling where it creates the tunnel in the background and terminates and so you might not know the random port it assigned or I couldn't figure out how to un-tunnel it and tunnel again to the same port. Just bypassed the control master then.

ghrlon FontCrafter: Turn your handwriting into a real font

There used to be multiple tools like this from different websites, but they were all bought by Calligraphr to redirect to them instead, giving them an effective monopoly and letting them charge subscription fees for generating fonts over the limits of the free version. I used to create two fonts and merge them with FontForge to get a complete usable font.

Great to see some competition on the market. Completely in the browser would mean it does not depend on a server and continues working as an archived version, so that's certainly great.

ghrlon Show HN: OxiMedia – Pure Rust Reconstruction of FFmpeg and OpenCV

I see a single commit adding nearly 2M lines of code with the readme claiming to be a full production-ready project. Am I missing something here?

Also, since it mentions full WASM support, a web demo would be nice.

ghrlon MacBook Neo

Since it's just USB and not Thunderbolt, wouldn't it use DisplayPort Alt mode for the display leaving data transfer untouched?

ghrlon Making Firefox's right-click not suck with about:config

Yeah, it's one of those features where after getting used to it you just can't understand why not every browser has it. I remember trying to copy an image from OneNote and conveniently in the custom content menu there is a button to copy the image. The only thing it does however is tell you it doesn't work and to use Cmd+C instead, which doesn't work either. So Shift + Right Click saves the day again.

ghrlon Show HN: I built a sub-500ms latency voice agent from scratch

I would say that depends. When it tries to upsell Prime subscriptions into even more Amazon subscriptions I always interrupt it and say the command again so it stops, but a few times it told me "this item in your cart is on sale by some %" and that did make me buy the item.

ghrlon Show HN: Unfucked - version all changes (by any tool) - local-first/source avail

This tool seems very promising and does solve a real issue I've certainly had quite a few times already where this would have been very useful.

I really like the website design and content-wise, as well as the detailed writeup here on HN. Certainly impressive work for an individual.

I've never used Time Machine but I do have kopia set up for frequent backups of important places, like the entire Documents folder. It does use a CAS as well and last time/size/hashes for determining what to save.

Coming from that I'm curious about a few aspects. Would this work well for larger folders/files? You mentioned deduplication, but does that happen on the file level or on chunks of the file, so that small changes don't store an entire new version of the file? Additionally, are the stored files compressed somehow, probably a fast compression algorithm? I think that could make it work for more than just source code.

Great project though, so far. I could see it becoming a lot more popular given an open source code base. Maybe a pricing model like the one Mac Mouse Fix uses would work, being open source and charging a fee so small it still reaches a large audience. That would likely be fair considering the developer time/agent cost of just a single unf'd problem as ROI.

ghrlon GrapheneOS – Break Free from Google and Apple

"Break free from Google" and buy a Pixel phone from them to do so.

But unironically Pixels are currently some of the best actually open phones. They do not lock down or require shady practices for unlocking the bootloader (although they do require a network check once that happens automatically, but it will permanently allow unlocking the bootloader if successful once. Pixels are very easy to restore and almost un-brickable, allow bypassing the boot screen warning by pressing the power button twice, actually allow relocking the bootloader and don't void your warranty unlocking it, don't have a shady one-time fuse like Samsung phones do with Knox, etc.

ghrlon Show HN: Free alternative to Wispr Flow, Superwhisper, and Monologue

Yeah, that model is amazing. It even runs reasonably well on my mid-range Android phone with this quite simple but very useful application, as long as you don't speak for too long or interrupt yourself for transcribing every once in a while. I do have handy.computer on my Mac too.

https://news.ycombinator.com/item?id=46640855

I find the model works surprisingly well and in my opinion surpasses all other models I've tried. Finally a model that can mostly understand my not-so-perfect English and handle language switching mid sentence (compare that to Gemini's voice input, which is literally THE WORST, always trying to transcribe in the wrong language and even if the language is correct produces the uttermost crap imaginable).

ghrlon Suicide Linux (2009)

I did something similar while I was still working with Windows a long time ago. I had just switched to PowerShell from the basic command line and kept typing cls, which did not work. I had typed that so often it was completely in my muscle memory, and every time the ugly PowerShell error would appear. So I decided to do the proper thing and NOT alias cls to clear, but instead alias it to immediate shutdown (shutdown -f -t 0 -s iirc) and that did work eventually. Wouldn't change a thing since clear is the universal command almost anywhere so it's a lot better muscle memorizing that!

ghrlon GLM-OCR – A multimodal OCR model for complex document understanding

I remember someone building a meme search engine for millions of images using a cluster of used iPhone SE's because of Apple's very good and fast OCR capabilities. Quite an interesting read as well: https://news.ycombinator.com/item?id=34315782

ghrlon Netbird – Open Source Zero Trust Networking

https://headscale.net/stable/about/faq/#scaling-how-many-cli...

> Scaling / How many clients does Headscale support? > It depends. As often stated, Headscale is not enterprise software and our focus is homelabbers and self-hosters. Of course, we do not prevent people from using it in a commercial/professional setting and often get questions about scaling. > Please note that when Headscale is developed, performance is not part of the consideration as the main audience is considered to be users with a modest amount of devices. We focus on correctness and feature parity with Tailscale SaaS over time. [...] > Headscale calculates a map of all nodes that need to talk to each other, creating this "world map" requires a lot of CPU time. When an event that requires changes to this map happens, the whole "world" is recalculated, and a new "world map" is created for every node in the network. [...] > Headscale will start to struggle when [there are] e.g. many nodes with frequent changes will cause the resource usage to remain constantly high. In the worst case scenario, the queue of nodes waiting for their map will grow to a point where Headscale never will be able to catch up, and nodes will never learn about the current state of the world.

I find that quite interesting and it is one of the reasons I've not really considered trying out Headscale myself.

ghrlon It's hard to justify Tahoe icons

I find the snowflakes equally irritating, but believe there is a huge distinction. This is an individual's blog page, their own corner of the internet, where they can share their opinion and present them in any way they'd like, whereas Apple is designing an Operating System for premium devices used by millions of people that depend on its usability.

ghrlon Hello-World iOS App in Assembly

You could check if consumerrights.wiki already has a page on that company and if not create one. It's a great resource that will also be used to justify demands for changes to the DMCA.

ghrlon You too can run malware from NPM (I mean without consequences)

I suppose if you're using a bundler, you will ship JS bundles including the malicious packages from your own trusted domain. How could CSP prevent this or similar attacks?

ghrlon I ditched Docker for Podman

I use OrbStack too and think it's great software, both for running containers and stuff like having a quick Alpine environment. However, I don't see the point of running Docker within Alpine. Wouldn't that defeat the optimizations they have done? What benefits do you get?

ghrlon Study mode

It would be incredible if OpenAI would add a way for schools and other educational institutions to enforce the use of such a mode on a DNS level, similarly to how they can force sites like YouTube into safe mode. Many students use ChatGPT, often without permission, to do work for them instead of helping them do the work themselves. I see a lot of potential for a study mode like this, helping students individually without giving direct answers.