frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Show HN: TikZ Editor – WYSIWYG editor for figures in LaTeX

https://tikz.dev/editor/
120•DominikPeters•1h ago•24 comments

Unlimited OCR: One-Shot Long-Horizon Parsing

https://github.com/baidu/Unlimited-OCR
273•ingve•4h ago•76 comments

Lift4D: Harmonizing Single-View 3D Estimation for 4D Reconstruction In-the-Wild

https://lift4d.github.io/
31•ilreb•1h ago•1 comments

Spying on kids to save kids from spying is stupid

https://pluralistic.net/2026/06/23/destroy-the-village/
233•hn_acker•2h ago•148 comments

Mistral OCR 4

https://mistral.ai/news/ocr-4/
157•meetpateltech•2h ago•45 comments

AI's Affordability Crisis

https://blog.dshr.org/2026/06/ais-affordability-crisis.html
74•ilreb•56m ago•64 comments

Show HN: Bun-sqlgen – Type-safe raw SQL for Bun, no ORM

https://github.com/ilbertt/bun-sqlgen
25•ilbert•1h ago•11 comments

MSG Made Dossier on Activists Who Opposed Facial Recognition

https://www.404media.co/madison-square-garden-made-dossier-on-activists-who-opposed-facial-recogn...
133•cdrnsf•2h ago•26 comments

Plotnine

https://plotnine.org/
176•tosh•4d ago•56 comments

Open Source for IBM Z and LinuxONE

https://community.ibm.com/community/user/blogs/elizabeth-k-joseph1/2026/06/18/linuxone-open-sourc...
16•ncruces•3d ago•1 comments

Show HN: Treedocs: Documentation that automatically checks for staleness

https://dandylyons.github.io/treedocs/
9•DandyLyons•1h ago•4 comments

GLM-5.2 – How to Run Locally

https://unsloth.ai/docs/models/glm-5.2
516•TechTechTech•18h ago•245 comments

Will It Mythos?

https://swelljoe.com/post/will-it-mythos/
242•mindingnever•11h ago•179 comments

Lossless GIF recompression via exhaustive search

https://blog.arusekk.pl/posts/lossless-gif-recompression/
26•ZacnyLos•3h ago•3 comments

VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO

https://arxiv.org/abs/2606.16140
306•timhigins•14h ago•162 comments

Researchers used math to crack Wordle

https://www.binghamton.edu/news/story/6327/s-m-a-r-t-these-researchers-used-math-to-crack-wordle
19•hhs•2d ago•22 comments

Five monitors on a Commodore 128 [video]

https://www.youtube.com/watch?v=ul5hC3PY1Yg
3•EvanAnderson•21h ago•0 comments

Crypto in 2026: Oh, This Is the Bad Place

https://www.stephendiehl.com/posts/bad_place_2026/
281•ibobev•6h ago•327 comments

Steam Machine launches today

https://store.steampowered.com/news/group/45479024/view/685257114654870245
1814•theschwa•22h ago•1547 comments

80386 Early Start Memory Access

https://nand2mario.github.io/posts/2026/80386_early_start/
14•nand2mario•3h ago•1 comments

Show HN: Neural Particle Automata

https://selforg-npa.github.io/
63•esychology•7h ago•14 comments

In praise of memcached

https://jchri.st/blog/in-praise-of-memcached/
232•j03b•14h ago•99 comments

The Traditional Vi

https://ex-vi.sourceforge.net/
50•exvi•7h ago•34 comments

The Low-Tech AI of Elden Ring

https://nega.tv/posts/low-tech-ai-of-elden-ring.html
16•g0xA52A2A•4h ago•3 comments

Show HN: Shumai – open-source Frame.io alternative for creative work

https://github.com/shumaiOne/shumai
37•Yiling-J•6h ago•2 comments

Giant Banana Pulled Over: Driver Says Cops Have Stopped Him 100s of Times

https://cowboystatedaily.com/2026/06/18/giant-banana-pulled-over-in-montana-driver-says-cops-have...
159•speckx•2d ago•60 comments

8086 Segmented Memory was a good idea

https://owl.billpg.com/8086-segmented-memory-was-a-good-idea-almost/
53•billpg•2d ago•96 comments

OpenAI DayBreak – GPT-5.5-Cyber

https://openai.com/index/daybreak-securing-the-world/
178•AaronO•14h ago•135 comments

Elevated error rate across multiple models

https://status.claude.com/incidents/jbhf20wjmzrf
156•rob•1h ago•182 comments

My Mathematical Regression

https://blog.dahl.dev/posts/my-mathematical-regression/
339•aleda145•4d ago•134 comments
Open in hackernews

Armbian Updates: OMV support, boot improvents, Rockchip optimizations

https://www.armbian.com/newsflash/armbian-updates-nas-support-lands-boot-systems-improve-and-rockchip-optimizations-arrive/
74•transpute•1y ago

Comments

proxysna•1y ago
Armbian is an exceptional project, even if the support might be uneven in some places, being able to roll out the same OS across almost every SBC i have is an absolute game changer. If there is support, Armbian is worth trying 100% of the time.

Edit: Also if you don't like/want Ubuntu/Debian their build documentation is pretty great.

dima55•1y ago
Their website doesn't answer the obvious question: what is it, and how is it different from vanilla debian? Do you know?
qwertox•1y ago
Vanilla Debian will not run on your nice and shiny Radxa Rocks 5B or Banana Pi whatever.
dima55•1y ago
Why not? What's missing?
qwertox•1y ago
Different boot process, U-Boot needs to be compiled for the exact board, drivers for the specialized components are needed, DTB (on ARM systems, the kernel doesn't probe hardware the same way a PC does) and other reasons.
RetroTechie•1y ago
> Different boot process, U-Boot needs to be compiled for the exact board

Why? That sounds dumb. And (assuming you're correct), how does Armbian deal with that / get around it?

ajb•1y ago
It's basically the same in the x86 world : your bios is customised to the board

The sad part is that on ARM the kernel is usually also custom compiled for the board. So what happens is that Armbian ship a different image for each board.

If you go and look in https://github.com/torvalds/linux/tree/master/arch/arm you see a zillion "mach-xxx" directories for different SoC architectures, even if they all use Arm.

Device-tree is a partial solution, but no-one seems to have an incentive to finish the job and let a single image run on any (sufficiently recent) arm board. It's difficult for the community to fix because most people have only their own board. Someone would need to pay for a CI rig with every board, and some kernel devs to do the work of building a single kernel to run across everything. (I think that's originally what Linaro was for - not sure why they didn't finish the job)

qwertox•1y ago
Right, the x86 BIOS/UEFI is baked into the motherboard firmware and handles early hardware init in a mostly standardized way. But with ARM boards, there's no universal firmware, it usually needs to be part of the image you download for that specific board.
FlyingSnake•1y ago
How does Armbian compare to DietPi?

FWIW: I’m running dietPi on my OG Pi Zero W and it doesn’t even hit 30% resource usage.

apple4ever•1y ago
Completely agree. I use it on my old PINE64 and it keeps on ticking.
chris37879•1y ago
I just stumbled across armbian recently and I must say I really like it.

I wanted to use UEFI, but my orangepi cm5 modules don't seem to have the SPI chip needed to store the UEFI there, so I'd have to load it on a partition and lose out on some features like persisting variables across boot.

The arm ecosystem really needs to settle on some sort of universal boot loader / firmware layer and stop just hacking up the linux kernel and not contributing back to it.

Nexxxeh•1y ago
I'm not an Arm dev and am just a consumer so I may be misunderstanding, but isn't Arm SystemReady pretty much the thing that's intended to solve the problem you're talking about (among others)?

https://developer.arm.com/documentation/107981/0302/SystemRe...

robotnikman•1y ago
It is, but it seems like only servers are adopting it at the moment. Or high end ARM workstations. I can't think of any consumer devices or SBC's off the top of my head that support it.
moondev•1y ago
Raspberry PI and Nvidia AGX

https://github.com/pftf/RPi4

moondev•1y ago
https://developer.arm.com/Architectures/Unified%20Extensible...
yjftsjthsd-h•1y ago
> Why? That sounds dumb.

Good, you understand the situation perfectly.

> And (assuming you're correct), how does Armbian deal with that / get around it?

You'll notice that if you try to download it from https://www.armbian.com/download/ , nearly every board has a different download image; this is because every one of those images embeds its own boot chain. There are efforts (in some projects, I'm not aware of armbian doing this) to build some amount of early bootloader per-board (often uboot), and just make the install steps something like "install this per-board thing, then install the real OS using a standard image" but that's less common and doesn't work super well when that initial bootloader has to go on the same storage device as the main OS.

dima55•1y ago
I believe that's common on ARM devices. But "vanilla debian" generally refers to userspace, and that should just work. Is this "armbian" thing quite literally "kernel + bootloader + vanilla debian"? The website doesn't say that in any obvious place
puzzlingcaptcha•1y ago
Pretty much, plus their little configuration utility for loading dtb overlays among other things.
pabs3•1y ago
The hard work of upstreaming/mainlining all the hardware support code in the userspace drivers like mesa, the Linux kernel core/drivers, bootloaders like GRUB/u-boot, boot firmware like coreboot/Tianocore/u-boot.